Although they aren't mentioned in the article, it seems like a pretty similar story: I worked for Accenture and they cared very little about the end user, product, or the client. The things they cared the most about were sales, billable hours, and reputation. As a software engineer, every review season completely ignored all of my technical accomplishments and problem solving skills and focused entirely on growing sales and billable hours. The cherry on top was the way they treated their employees that were working in the states on Indian salaries. They no doubt billed the customer ten times what they were paying the employee while claiming cost savings. The dishonesty, disrespect, and lack of integrity were palpable.
It's no wonder why projects managed by large consulting companies are total disasters.
Don’t gripe, celebrate it! Huge companies getting conned by large consulting firms creates opportunities for startups. If Deloitte and others were actually good at what they do, it’s be much harder to compete. Large companies blowing a hole in their budget so that Deloitte can deliver a garbage product 6 months after the deadline is the kind of thing you want to hear if you’re involved in startups.
Startups have no chance with these clients. It's all about deep industry/government connections and kickbacks. It will be Adobe AEM consultants billing 300$/hour and not GatsbyJS ninjas.
Everyone assumes its about connections and kickbacks, but its actually more to do with procurement policies.
The people who sign off on large contracts need to be able to demonstrate due dilligence, because if they don't then when something goes wrong they will be blamed for hiring cowboys instead of professionals. But this merely creates a series of hoops through which the candidate contractors must jump. Ultimately the process of getting these contracts becomes so cumbersome, expensive and risky that only a few large companies do it. They naturally become experts at hoop-jumping, often to the detriment of actually doing the work.
I once saw the bid evaluation criteria for a large company I worked at. Over half the scoring points could be summed up as "one of the usual suspects" (e.g. big company, has track record in industry, has staff with industry experience, has done similar work in the past). If you weren't one of the usual suspects you had no chance of getting the job.
Absolutely correct. Saw it on the Gov't/Defense side of the house and in the civilian Enterprise sector.
To be blunt, the reason big firms get the contract is because we know where they live, we know they'll deliver something, even if it's utter dogshit, and we know they won't go out of business in 3 years, so if we need to support the build -- or just sue them -- then we know where to find them 36 months from now. The Big 4 may only bat .500 but I can bet against ~50%.
There is also a selection bias: no one has heard of the small-fry shops that fuck up, and when they do they can quietly re-brand, disappear, or dissolve. Anyone here ever get burned by YellowOps? Probably not. But everyone here has heard of KPMG...
> (e.g. big company, has track record in industry, has staff with industry experience, has done similar work in the past)
That's the classic CYA, or the "nobody gets fired for buying IBM" adage. I've dealt with procurements for eye watering amounts both in the private and public sector. But no matter how much you struggle to get the good product eventually some hard requirements (either legal or institutional) will be shoved down your throat and you end up with the Accentures of the world.
They must have comparable revenue to the customer. They must have had similar projects going at least 10 years back. They must be able to deliver the service 24/7 in any corner of the world even if we only have one office. You can't say anything about past performance in projects even if it's objectively measurable because it would put one competitor at a disadvantage. So what if Accenture blew the budget by an order of magnitude and delivered a year+ late on every single project they had with you, that's the past, we're starting with a clean slate every time.
Then the contractual conditions can't be harsh enough that even Accenture says "not for us" because that all but guarantees that you have no bidders.
So now every startup is out the window and all you're left are the big ones. And when the project bombs - and bomb it will one way or another, Accenture will guarantee that because they make money by underbidding and then overcharging you through your nose for CRs and fixing what they failed to deliver under one contractual loophole or another - the people in charge from the customer side always have good cover. They picked the best offer. Nobody could have done it better.
This goes hand in hand with Gartner recommendations, the company that came to me with 2 conflicting reports and let me choose which one I want to present to management depending on what version I wanted to push. I laughed them out the door and when going face to face with the higher ups the first thing that came up was "what does Gartner say?". Try answering "anything I want"... They need plausible deniability so you're there to plausibly shield them from the reality of this farce. So you call Gartner back and ask for that report "we first talked about".
Nobody gets fired for buying Accenture (or the likes). They'll just leave in disgust.
In other words, you could hire Accenture and pay some big cat a lot of money, or you could pay 10 startups that money and let 90% of them fail and let someone to become big again.
Mathematically yes (a la "9 women can deliver a baby in 1 month"). But in practice no because there's almost no way you can write fair requirements in a procurement that will apply to all and somehow exclude only Accenture (&co.). This is their actual power, officially they tick all the boxes and look as qualified as one can be, and they can also underbid almost any startup. Because they have far more qualified lawyers than engineers to help the make up for it later. It's only when they start work do you get the real Accenture experience: hundreds of inexperienced, unqualified, overworked, or underpaid people under the careful eye of someone who's paid to delay and overcharge under any possible interpretation of the contract.
They are better at this game than almost any startup because their game is "winning bids". They are ridiculously connected and able to operate almost exclusively in the grey area, with a lot of detours in dark as hell area. And they are always seen as the "safe choice". If Accenture fails a CEO can say "we did our best, picked the experts". When a startup fails the CEO has no plausible deniability. They went with the risky, unproven option and have to pay the price.
You can either go the fair competition route and Accenture can easily tick the boxes on paper and then win on price, or you can go for direct award (in a private company) but you'll be living at the gates of hell until the project is successfully completed.
This is true. I worked with a communications provider and we were not cheap by any means when it came to sending email/sms, and turnaround time and capability lagged behind any smaller group. The major driver from the clients was in not being blamed for picking the small company when (not if) something goes wrong, along with the usual sales and relationship building games that entrenched vendors play.
Interesting reasoning. Few days ago when there was yet another GH outage I read in a comment here that it seems that larger companies get more slack for mistakes b/c of the idea that a large company must surely provide an as solid as possible product while if a smaller company fails people come around asking "why didn't you go with [large company] instead of this [small company]?".
Also I observe it with Linux vs Windows. If you promote Linux and then something itches people are quick to express "this wouldn't have happened with Windows". While if Windows causes problems people will easily accept it as caused by higher powers.
I wonder if there is a name for this David vs Goliath syndrome.
There are plenty of opportunities in the government for small businesses, mandated by law. Learn how to team to go after contracts and look into the SBIR open topic. It isn't easy and takes patience, humbleness, and work but the payoff is long term reccuring revenue from a customer that always pays.
This isn't really true. Not only are SIBR rewards generally low, but it's very easy for contracts to be pulled out from under you via larger contracting firms and there is a ton of risk involved for the employees.
I would never ever go back to that environment after working in it for a few years. Not knowing whether or not you're going to have a job for over a month every year is not fun. Larger contracting firms can shuffle people around in those cases but for smaller ones it kills an entire team.
We are onboarding a remote team this week for a government contract working with CNCF tooling in AWS. COVID changed what is allowed. Also SBIR phase II for open topic are 750k for the next period.
That's $750k for two years, which is paltry when you consider that your average senior engineer salary is likely to be in the range of 150k+ per year because of how much the major players offer and you also don't get funding for Phase III. A small team of engineers and QA will eat up that award quickly. This is without factoring in other things your company needs to navigate the government labyrinth, healthcare and more if you aim for commercialization.
If you and your team are willing to be underpaid relative to other companies and deal with the risk that working on government contracts entails a smaller contracting company might survive for a few years.
750k for 9 months. You need to write the scope to be profitable, just like any other fixed firm contract. If you execute in a need area phase 3 should be high percentage if you build the right relationships during your phase 2.
> Don’t gripe, celebrate it! Huge companies getting conned by large consulting firms creates opportunities for startups. If Deloitte and others were actually good at what they do, it’s be much harder to compete.
Eh, no. I bet Deloitte and Accenture are actually good at what they do, but that happens to be c-suite sales and other parasitic activities. End-user pain is an untracked externality in these arrangements.
Putting the idealism aside, I also thought startups where pretty much in the opposite business to these billable hour factories.
Some of the best people I've worked with were Deloitte people. I hate this generalization of large consultancy firms. Sure, there are assholes, but a large part of the work force is really doing the best they can.
I worked for a small consultancy company (that you could call a startup) on large projects. We delivered a garbage product, after the deadline and way out of budget. Why? Because large companies are really bad at knowing what they need and sales is really bad at telling them that.
No matter how small or big the company, you will see this happening, because there are always incompetent people promising impossible things. Just look at the hundreds of startups that burn money trying to deliver on their promises and end up dying.
Corollary this comment helped me realize: the successes and failures of any environment that doesn't stick around are forgotten, thus the failures of the commodity startup industry do not disseminate nearly as widely as idle commentary/banter about poor customer service associated with a specific brand.
I once got pretty far into recruitment with Accenture, and got to speak with management pretty high up the chain.
It left me with a very curious thought, namely that there's something very rotten that infects any corporation in that business, cause it was nearly a mantra that every higher-level that was part of the interview repeated - that they need people who can tell customer "No" and keep things in reasonable scope and done right.
As if there was some managerial disease across the sales, engineering and clients that pushed towards certain bad outcomes.
Mind you, some areas actually do have such "disease" that is remarkably well studied for something so ephemeral - for example in aviation, flying around VIPs tends to quickly spiral into bad behaviour, with bad things done on basis like "we couldn't let PM be late", caused among other things by gratitude of the PM and thus crews trying to do a little extra. This required separate review and controls to keep from happening and quick negative feedback applied by safety bureau every time it happened, otherwise it would spiral out.
Like the other commenter said, it's logical. It's not a managerial disease, it's just inherent to big projects that details are never clear at the start of the project, but clients do want a clear cost projection.
What happens is that a scope is agreed upon and so is a price. If the scope increases, so should both the price and the deadline. The client never wants this to happen, because it's their money. The contractor can then do either of three things:
- Don't build the feature or remove another feature
- Make the client pay for the feature
- Lose money on the feature
Having a developer or a sales manager or whatever, that can't say no always tends to lead to the third option, which is the one that isn't desirable. Therefore, you need someone to say no, followed by 'unless...'.
The big issue with Agile (for a consulting firm) is that this is inherent to the way of work, because in Agile the end result shouldn't be clearly defined, which means endless discussions on scope. Because of this, it's even more important to have developers and project managers who can say no (unless...).
This is mostly important for project work. Many consultancy firms also do hourly work, where most of this risk is with the client anyway. Even then, however, you still need to be able to say no sometimes.
That makes sense. Scope creep is the main thing that's killing large software project. The client doesn't know what they want yet every middle manager and department is constantly requesting new features. If a software contractor wants to deliver anything within some time and budget, they have to manage the scope tightly.
The generalization is not about the employees of the large consulting firms, but about the business practices in these companies and it is quite accurate.
A 'company' can't actually do anything, only its employees can. Anything said about a company is therefore also said about their employees, since they are what forms the company and perform the actions in the name of the company.
Except, no. As a startup you are absolutely not going to get the contract to create a new unemployment web site for a government. Even aside from the bidding process that is deliberately set up to exclude small players any responsible government wouldn't go with you anyway: what happens in five years time when your startup has shut down (or complete it's Wonderful Journey with an acquisition) and the site stops working?
You could, but would require extensive involvement before the bid, making sure you understand the organization and structures (e.g. https://www.naswa.org/ and http://www.itsc.org/Pages/default.aspx ) perhaps hiring someone who worked there, plus experience in dealing with legacy tech.
The RFPs are public, and many states are willing to take a look at new approaches.
This isn’t a rhetorical question as programs like FedRamp exist to help startups have those good answers for responsible governments. While a brand new startup will not meet criteria, relatively new companies are able to meet requirements and earn business.
At the high level (ten figure contracts) startups have no chance vs Amazon and Microsoft, but there are many agencies and purchasing available to pursue.
I am not talking about startups competing for the same contracts, I am talking about startups competing with a large company in their line of business. Blockbuster couldn’t hire Accenture to create Netflix, for instance.
Consulting is really a relationship business. Typically, a strong relationship is built on the back of successful projects but not always. Just because a project fails doesn't mean the relationship is at serious risk. At the top of consulting firms you'll find that what everyone cares about is client and industry relationships.
A startup is going to have to start building that relationship from scratch where the firm with a freshly failed project may still be a decade ahead.
> Huge companies getting conned by large consulting firms creates opportunities for startups.
Huge companies can go bankrupt. I can tolerate "creative destruction".
Governments will go very far out of their way to spread the pain before they go bankrupt. The Flint, Michigan "lead in the water supply" problem was downstream of the city's (and region's) economic collapse and bankruptcy. Detroit's current residents are very much overpaying for their utilities and other city services because of their existing long-term obligations.
I would much prefer we change government to be much more iterative and create contracts which ensure quality outcomes (not optimized for "butts in seats"), even if that means there is more total cost.
Isn't that a broken window fallacy? Once the time and money has been wasted what are the actual odds the customer will just turn around and go "Let's try the same thing with someone else. It's bound to work someday!"? Most likely the opportunity is gone, and the only winner is the shoddy consultancy.
If the customer went all in on the broken first solution that would be the sunk cost fallacy, but that's not what GP said. Maybe it's not the broken window fallacy, but GP saying that terrible customer service from big consultancies creates opportunities for startups doesn't follow at all, because customers don't have infinite amounts of time and money to get a working solution.
Their continued existence developing failing government systems is indicator #1 of the criminal nature of the US Lobbying system that awards government contracts. Say what you want, it does not alter the fact that what they deliver does not fulfill the contract to the degree it is criminal theft of taxpayers' money.
My thoughts exactly when my girlfriend had to apply for benefits. The UX is awful, it’s difficult to tell where you are on the page or what you should be looking at. On and on...
> They no doubt billed the customer ten times what they were paying the employee while claiming cost savings.
You forgot the part about how the engineer is very aware of this hourly arrangement and feels their job security depends on their ability to get work done in less time, and so they they take tons of shortcuts and pile up technical debt like crazy.
Then the other engineer has to make changes through that technical debt, and he’s also very aware that his time is the company’s money (because the managers remind him of it constantly), so he does 20 hours of work over the weekend and logs it as 4 hours so his boss won’t get mad at him.
And the part about how half of them will lose their visa and have to leave the country if they complain about any of it.
If you’re reading this, and you write software for one company that later becomes the property of another company: QUIT. YOUR. JOB. RIGHT. NOW. Interview at ANY other company, your life can be 10x better in a few weeks, just go.
Isn't the whole point of consulting to win more work? You get promoted and paid because you have money coming in, so it makes sense to do what you can to keep winning more work (i.e. sales, billable hours and reputation).
You shouldn't be working for a consulting firm if you don't want to focus on those things.
It's a pretty well-known fact in Indian universities than anyone even a teeny bit serious about computers should never step foot in consulting companies (unless you are leading a green field project).
Firstly, India has a huge engineering population. Even if 10% of it is interested in consulting jobs, it's easily more than most countries. Secondly, most folks do engineering just because it's easy path to get a decent paying job. They don't really care about engineering itself, so consulting jobs are a happy medium, also they provide easy transition to customer facing roles or management which many folks want to move to anyway.
And lastly, there's dearth of product based companies in India, so positions in such companies are highly lucrative and competitive. Consulting job is the last resort for someone interested in software.
If the institution implicitly understands its own incompetence (that they will explain away as "we just can't find good developers"), they know they can't win repeat contracts, so they optimize for milking the few that they do get.
The people consulting firms “deliver” to have no idea it’s broken and don’t really care because they’ve never seen the software or process and they never will.
Plus, Deloitte said it’s fixed now and Deloitte knows best or else they wouldn’t have charged so much, right?
Half the reason these consulting firms get hired is to be a scapegoat by the director/exec when things go south. Liability is one reason they can charge a premium.
The companies seem to be claiming that their products work fine in other places and blaming their customers for the places where they fail.
> The consulting company defended its performance at the time, saying it was continuing to improve the system and it would run more smoothly as employees from the Department of Economic Opportunity (DEO) learned to use it properly.
> “As we have communicated to DEO, we believe that any remaining issues deemed ‘high impact’ by DEO’s own definition either require Departmental actions or are otherwise beyond Deloitte’s control,” they said.
> Price also said the company had worked on systems in states including Minnesota and New Mexico, which he said have weathered problems relatively well.
Spot on, that’s it exactly. Worked with Deloitte aka Doolittle for several years and they delivered consistent crap results, but the people who kept signing the contracts weren’t even SMEs in the area of their responsibility.
Reputation is managed by networking: going drinking with the right people is often more important than delivering quality products.
Or, in the case of government contracting, providing a revolving door for people who leave government to be given lucrative private sector jobs.
G4S were selected to run a new private prison at the same time as being fined £44m for previous fraud. But then, private prisons in the UK aren't a "market" - how could they be? There's only one buyer and in practice only one seller.
Because quality isn't what matters, and they can paper over it with ads, personal relationships, presentations, sponsorships.
I had a pretty good experience with Sapient once.
I suggest it probably depends a lot on the location though I would never touch Accenture or Deloitte and would only allow Bain or McKinsey to deliver on something for which they had very material experience, and never let them do actual work because they will f it up.
If OP is me, I think it was implied when I referenced quality. If it was the author of the article, I believe that was part of the title of the article.
Measure the success of the project via improved metrics, performance, bugs getting fixed, features getting done, etc. Instead, they measure if they billed 50 hrs every week and have successfully increased their client's consulting budget for next month.
Because the stakeholders don't use the system they buy. They foist then on others. And any complaints are treated like the ignorant bleatings of the rabble. They'll drive home to their mansion in Herndon in their second BMW and it will never come up.
Nobody wants to bring up Accenture here? Why should Deloitte get all the credit? There is plenty to go around. The large publicly traded consulting companies are like the mafia - once you let them in the door, they're not leaving until all your money is gone.
Thats all there is to it. Funny thing is those companies are never brought in by technical departments. My most recent experience saw the Marketing dept bring in CONSULTING COMPANY to do a particular job because the internal Engineering department's quote for the same work was "5 times what CONSULTING COMPANY said it should cost if they did it".
(queue violent laughter for those who have seen this joke play out)
It was a mess - consultants were as inept as they were numerous but damn, they dressed nice every day. They all had great glasses too, like they were assigned by the wardrobe department. Picture a Banana Republic Catalog's worth of new, very young people running around the office and you get it.
I'm not saying these consultants were inept because they couldn't invert a binary tree or because they couldn't tell me the time complexity of a HashTable. I'm saying it because they literally knew nothing aside from the intensive training CONSULTING COMPANY gave them for this job. If it was not covered in training they not only couldn't do it but didn't know what it was in the first place.
The whole project became a train wreck in slow motion. You could see it was going to be bad but it had so much momentum there was no chance of stopping it. Must be how the people felt on the Titanic come to think of it.
Things got so insane I realized that the consultants and those responsible were looking for someone to blame - and I promptly accepted a new job offer and vanished in a puff of smoke. I would later find out I was on the short list of people to be "promoted" to manage deliverables for that project!
After being laid off in 2016 I joined a big consulting company (you've heard of them but they're not Accenture) because I needed to pay the rent. I didn't come up through their training program as a new graduate but as an experienced developer.
Every year I got glowing praise from the client - I solved the problems they had faster than their in house guys did, I delivered on time with the lowest defects of any contractor across all firms that worked at the client, and was a domain expert in one of the client's systems after about six months there.
Do you know what the consulting company offered me as a raise? $150 on top of my base salary. Not $150 a week or month - $150 total. When my consulting company reviews came back, despite high praise from the client, I wasn't enough of a member of the consulting company's team because 1) I didn't hang out with only consultants from the company, 2) I didn't attend drinking events every week with consultants from that company, and 3) I was too specialized on what the client did, so how were they going to move me to another client for more money?
I stuck around for another year and then the client negotiated hiring me on directly.
I will starve before I work as a consultant at someone else's consultancy. I'll start my own one-man firm first.
This is a very skewed view of billing and profit. I have worked at consulting agencies as an engineer most of my career, in management for a few years, and have seen the numbers. There are quite a few overhead considerations you need to take into account. Just in payroll, a $16/hr employee is a lot more expensive than that number. Closer to $30-$50/hr with taxes, healthcare, and other per-person costs. Then add office space, equipment, sales and marketing cost, management, HR. Some management hours are billable but not most. Then consider that most employees don't bill 100% of hours. If we paid an employee for 40 hours, I would be impressed if the billable hours were more than 35 on a good week. (We did not bill for stand-ups, breaks, and other events that were not directly project-related.) Additionally consider the risk factor, that some clients end up not paying for one reason or another, or we comp time due to mistakes, going over budget, etc. We still pay the employee.
I'm not defending consultancies here - working at one would be my own personal hell on earth.
But when charge-out rates are so much higher than salaries, this is because their value is in the customer relationships and sales pipeline, not technical competence.
It's far, far easier to replace technical talent than to replace customers.
Just because you can make exorbitant profits off of your workers doesn't mean you necessarily should. I'll never understand why someone would you want to own a company that treats its people like shit and sucks to work for.
Why not necessarily on topic, I do wonder when this ruthless mindset will become viewed as a negative. Currently its perfectly normal to defend your dislikeable choices as business choices; because we are running a business.
I've had a couple ideas in the past for businesses that I always openly share with others in the nearest detail hoping someone would pick it up, with the thought that I don't wanna make money of it, I just want to have that product or service, simply because it's not there yet. I'm sure this mindset exists but I am unsure what a formal term might be. Utilitarian comes to mind but not quite.
>> Just because you can make exorbitant profits off of your workers doesn't mean you necessarily should. I'll never understand why someone would you want to own a company that treats its people like shit and sucks to work for.
> Middle management can run distraction on the owners so that they don't see the misery of the operations people who actually keep the buisness going.
More likely, the owner is greedy and just doesn't care about how miserable the employees are as long as the cash keeps coming in. Some people just plain lack the regular moral compass and feelings of empathy, and others have gotten good at holding on to BS justifications to assuage any guilt they feel.
And sales commissions can be crazy high too. They are not just paying the IT people.. admins, HR, etc all need to get paid too.. not trying to defend crazy high rates...
It's a shame there isn't a cheap turn-key way to do this right now. That could go a long way to unbundle these mega-consultants.
Company-creation-as-a-service (CCaaS) with shell-company-in-offshore-tax-haven (SCiOTH) and automatic-tax-minimization-as-a-service (ATMaaS) would go a long way to help the average small company reduce their expenses and to coax Congress into actually fixing tax loopholes which currently primarily help only the corps which are scummy enough to take advantage of these extreme tax-minimization techniques.
I was an intern there almost 10 years ago. Apparently my job was to generate billable hours, because I wasn't given any work and when I asked for some I got a 700-page document to read.
After some nagging I was told to do the same thing one freshly minted consultant who was two years older than me was doing. She was clueless so I did both my part and hers.
The most sad/hilarious situation I had there though was when I bumped into my friend from university in the cafeteria, who was also an intern, but in a different company, and during our conversation it turned out that while I was doing essentially nothing and still got paid, he was working his ass off for free.
In hindsight I think his time was much better spent, but I really needed the money so I didn't complain.
I've worked a place where IT only worked through consultants.
IT knew nothing about nothing and wouldn't hire anyone who knew technology. All projects were big buckets of money that they would hand to external consultants. Consultants would always take the money, sometimes they would do the work. IT was always certain that whatever the consultant delivered was best-in-class, exactly what should have happened because it was delivered in an Agile methodology and Gartner said nice things.
In several years I never once saw the firm not deliver what they agreed to in writing.
There’s one way to deal with a consulting firm: a rock solid statement of work and people on the client side to verify you got exactly what you paid for.
If the system didn’t work as agreed in the statement of work and that was clear, Deloitte would be footing the bill.
If you gave out a time and materials contract without clear acceptance criteria, testing requirements, and verification / support / warranty detailed, you probably shouldn’t be handling 100 million dollar budgets.
> acceptance criteria, testing requirements, and verification / support / warranty detailed
Some orgs just don't have the in-house capability to do this, so they seek a consulting firm... oh wait.
Edit:
I guess the strategy is to hire technical staff, keep them around a few years until you can be sure they can deliver on projects and understand your existing systems and processes, and then embark on the larger project to transform those systems and processes.
Alternatively, shop around to find a consulting firm that is motivated to have good long-term outcomes. Probably a smaller operation, that has good word-of-mouth reputation, but doesn't have an established brand to fall back on.
Yep on a 100 million dollar budget you need to hire people to manage software delivery on your side and can never absolve end responsibility is definitely my point.
I’m not describing how the world should work, just how it does when outsourcing IT to consultancies (which can work anywhere from successfully to very unsuccessfully)
A good strategy is to hire a team, get a few consultants who have literally built the same system down the road, and use that blended team until 2.X release.
That way you a team that has made a thing already (and often consultants are just as passionate about their work too!) ready to build a better 2.0, plus you keep responsibility and control.
Some orgs may not know how to hire. I think here, competent (and well connected) consultants can help, and maybe source from their own networks, but we're back to the dilemma of which consultants to trust. And let's not speak of recruitment agencies :)
> If you gave out a time and materials contract without clear acceptance criteria, testing requirements, and verification / support / warranty detailed, you probably shouldn’t be handling 100 million dollar budgets.
That is oh so very Deloitte...look if we took $100M from you and you didn’t get what you wanted, that’s Clearly not our fault, YOU shouldn’t have been handling that type of budget in the first place. After all there are people like us out there that will hornswoggle you, luckily you can retain us Again to deliver the original contract and you should consider a separate retainer so we can protect you from people like us in the future.
Id love to see a single instance where Deloitte admitted to fault and footed the bill for a $100M contract, care to link proof for such a claim?
I agree it’s the way Deloitte works and not how the world should work.
> Why would they admit fault?
That’s just it they never would.
But reading your comment:
> If the system didn’t work as agreed in the statement of work and that was clear, Deloitte would be footing the bill.
You certainly make it sound Deloitte is a fair company that would admit when a system doesn’t work as agreed and then foot the bill, but that’s not the case.
Whether they admit fault For delivering a failed system or not, do you have an actual example where Deloitte was magnanimous and returned $100M to a client after delivering a system?
you state a lot in a very aggressive tone of voice. maybe that is because non verbal context is missing.
I am not the OP, but working for an agency shop (way way smaller than Deloitte) I know we footed some bills. But we don't publicly talk about that. wouldn't be good for business.
it's a double edged sword. I know of projects that were sub par. I have seen systems build by others (Deloitte included) that were nothing but a shit show. and I have seen systems from the same brands that were doing way more than originally expected.
sometimes it is the client, sometimes it is the project management sometimes it is the consultants. and sometimes it is the circumstances. I wouldn't expect a system designed for normal load to hold in times of covid-19 and work flawlessly. some examples like the call center being overwhelmed is to be expected when designed for normal loads.
I strongly believe, that there are real faulty systems to be found. and that for some of these the contractor is to be blamed. but without knowing the specifics I wouldn't feel qualified to judge from the sidelines.
disclaimer:
I work for an agency that got bought by a Deloitte competitor some years back. all opinions are my own.
> you state a lot in a very aggressive tone of voice.
Help me out with that.
I’m really just asking for some support of the original statement that Deloitte will foot a $100M bill for a client when they fail to deliver a working system.
I mean Deloitte isn’t singled out in the article, the article really goes after IBM also. That said I wouldn’t even say the article has an aggressive tone, but it speaks for itself. Deloitte has delivered systems that don’t work on multiple government contracts, all of which have run significantly over budget, and Deloitte has requested new contracts to fix the systems (Instead if footing the bill) and when sued for breach of contract by the states Deloitte has engaged in protracted litigation rather than “footing the bill”. I think it’s fair, not aggressive to ask for some validation of Deloitte footing bills where their systems failed in light of Deloitte's practices detailed in the article.
The article is all about state unemployment systems, I mean really they aren’t complex and they either work or don’t. Even Deloitte‘s argument amounts to be well if the $xxM unemployment system Doesn’t work it’s because the state requested a non functioning system, of course it’s all backed by Deloitte Representative quotes denying the systems don’t work and championing how great the systems they delivered are, and otherwise blaming the client if these great systems failed.
> I’m really just asking for some support of the original statement that Deloitte will foot a $100M bill for a client when they fail to deliver a working system.
I didn’t say that, they would foot the bill for the changes required to make it work as expected in the contract. It’s unlikely you would deliver 0 value with 100m, so it would be a fraction.
This would lower the overall margin on the project, be considered business development or some other activity. There’s also discounting future work as a lever to resolve it.
They aren’t stupid, biting the government hand that feeds them isn’t a good way to keep winning new work.
> They aren’t stupid, biting the government hand that feeds them isn’t a good way to keep winning new work.
Have you read the article? The article details multiple failures to deliver on government contracts nearing $100M. The article details a practice of refusing to amicably resolve matters in good faith, instead Deloitte prefers protracted litigation, which when involving government is just additional costs to the tax payers above and beyond the initial costs tax payers footed for the original contract to Deloitte resulting in a non working system.
There are a few read between the lines also, Deloitte knows their litigation budget is significantly larger than the state governments and they know their attorneys can overwhelm the states ability to litigating just through costs alone. The other point about biting the hand that feeds them...sure if you were small time and failed to deliver and forced the states to litigate a breach of contract case, you may be cutting your nose off despite your face. But entities like Deloitte can do this, and they will continue to get new government contracts while the prior contracts are still being litigated. You got your Deloitte’s, McKenzie’s, Goldman Sach’s, haliburton’s, Boeing’s, etc...they could set the world on fire and then lobby and get the contracts to put out the fires they started. Those firms are the reason for the term “revolving door politics.”
You’ll find many instances where companies like this have been kicked out early in the process.
Unemployment systems really aren’t that complex in the grand scheme of things. The problem domain is well defined — the program is almost identical to what it was 50 years ago (or more).
When you look at the examples cited, failure is not exactly a surprise. Florida isn’t exactly a poster child for competent governance.
You also need a competent project management team (backed by contract lawyers) that can push back. The QLD Health payroll system debacle that ended up costing 1.2b is a good case of how not to negotiate and deal with one of the big 4 contractors.
Yes, here north of the border it was Phoenix Pay. It seems to me anytime you pay 1.2B for a software project it is destined to fail from the start. The consulting company will overcomplicate things and cram so many cooks in the kitchen to justify the price tag that it's doomed.
> If you gave out a time and materials contract without clear acceptance criteria, testing requirements, and verification / support / warranty detailed, you probably shouldn’t be handling 100 million dollar budgets.
This is the money quote.
Big 4 companies (Deloitte, PwC, EY, KPMG) are historically accounting auditing companies. And their culture is heavily skewed towards billing their employees to clients per hour/day, or per unit of service, as auditors and lawyers have been doing for more than a century. It is not in their DNA to "create products" but to "render services"; that is, to sell employees availability. Boundaries and scope of those traditional services are typically clearer than building a new unknown program from scratch.
If you are hiring these firms for exploration of how do you work, how you should work, and a complete functioning system that sustains a big number of new processes that are not yet defined, you are in for a big bill; you better be sure that you have put on the time to think about what you are buying in written form. When you see big failed projects like those the problem lies in that no one stopped to research the complexity of the ask before entering into a contract.
Even then, these consultancies normally are very flexible when a contract is actively harming the client, and the client wishes to change it / stop it (it would not be good for the business to do otherwise, because you want to see more after this contract is finished). When these costs go out of control is because internal politics in the client impede the rational decisions of killing projects, changing command, etc., and the sponsor keeps feeding money into the project to save their ass.
It’s right in the article naming Deloitte specifically:
> “ "Deloitte also faced blowback in 2013 for an unemployment benefits system it built in Massachusetts. There, the benefits system arrived in 2013, two years late and $6 million over budget—only to be filled with glitches that erroneously cut workers’ benefits, according to the Boston Globe."
Deloitte delivered a major project way behind schedule, over budget and full of defects.
There is a massive asymmetry to the two parties here. I'm not sure the consultants are malicious, but they enable the moral hazard of the worst of these types of contracts.
Oh no doubt. I’m not shedding any tears on these consultants behalf believe me.
The folks in these companies are masters at separating customers from money, for sure. But, these guys aren’t ripping off clueless dentists either. The typical F100 customer isn’t a babe in the woods. When the parties are Accenture vs. Exxon or IBM vs. the IRS, who is the good guy? :)
Ok but the problem is that customers don’t always know exactly what they want. Software isn’t like architecture. You can’t always predict every possible issue or corner case.
And that’s where they get ya!
They’ll win a lowest bid contract for the exact word for word spec.
But then will bill wayyyy more for any change request. Which will be numerous obviously.
> In several years I never once saw the firm not deliver what they agreed to in writing.
I know people working with big consultancy companies and the recurring thing they complain about is a difference of opinion between client and provider about what exactly the terms of signed agreements mean.
Nobody said anything about waterfall. A good agile process can and should have all of the things the parent commenter listed. If you eschew those things by using the "well we're agile, that means we don't plan ahead!" excuse, that's a sign of incompetence.
> A good agile process can and should have all of the things the parent commenter listed.
A good agile process will not have an ironclad upfront definition of what is to be delivered in a large project. A process with that may not be waterfall, but it's definitely non-Agile Big Upfront Design.
OTOH, Agile is largely about managing risk and incrementally delivering value without overcommitting forward, so the contracting process with Agile should reflect that and shouldn't commit to a giant project in a lump contract.
>A good agile process will not have an ironclad upfront definition of what is to be delivered in a large project.
"what is to be delivered" does not have to be a rigid definition of "5,000 lines of code program with features 1, 2, and 3". A good agile contract will allow the possibility for changing project requirements but also still define how those changing requirements affect expectations and responsibilities of the consulting company.
The agile process itself can actually be a deliverable that you define in a contract, and it should spell out, for example, how new requirements are prioritized and the criteria you use for determining if it's something that the consulting company is responsible for. In my experience if you can't write out your agile requirements in a contract, then you don't have a good enough internal understanding of your own agile processes and should probably work on that, too.
Far too many people use "we're Agile" as an excuse for "we don't plan or document anything" and extend that to contracts, but that's absolutely not what good Agile actually is.
Longer term consulting contracts should 100% be waterfall. How would it even work any other way? These firms aren't delivering the latest smartphone app for their clients.
T&M only works for contracts for a few resources in a blended team in my view, with the only risk being a 2 or 4 week price on getting rid of them. Then you’re basically hiring overflow individuals you can get rid of easily who may have particular skills.
I've actually got good experiences with t&m, but that'd be anecdotal. The problems grow from when incentives become misaligned, and external work/consultancy is being used against coworkers.
Only for trivial projects with tight or well defined scopes. It makes sense to pay $5/mailbox. $500k to deliver a system that does <x> usually not.
Fixed price deliverables are usually more expensive due to bidders pricing risk. Or you see big delays due to PCR requirements. Or the project owner don’t have the political power to not accept a deliverable and tell the vendor to go pound it.
> Fixed price deliverables are usually more expensive due to bidders pricing risk.
Correct, that’s the only way the vendor takes on the risk really. That was my point.
> Or the project owner don’t have the political power to not accept a deliverable and tell the vendor to go pound it.
Yeah I agree... but if you have 100m to hand out and the in org politics where you can’t tell a vendor to go pound it, there are real structural issues in your co preventing you delivering systems.
I've worked with a bunch of Ex-Deloitte people and without fail they all could talk forever about the subject at hand, sound very technical, but in practice were completely incapable of executing anything correctly, and had absolutely no knowledge of their subjects of "expertise" beneath the surface. They really are expert bullshitters.
I think the real trick to being an expert bullshitter though is that you have to forget that you even are bullshitting. Ex-Deloitte people have deluded the most important person which is themselves. In technical conversations it's almost as if the idea that there even was something beneath technical talk behind these ideas.
It's to the point now where Deloitte on a resume is a huge red flag.
My fiancée is a senior tech consultant at a big 4 firm and I can say with the utmost confidence that they have zero ability to actually build things. I specifically watched her project manage a multi-million dollar product for a client. Clueless. Consultants are good, organized, hard-working people but they are not technical in the ways most on HN are. And the outsourced dev teams they hire at bottom dollar will just never deliver anything of quality.
I'm a former tech consulting manager at one of the B4 and my experience is that the types of multi-million dollar software projects that they get involved in are lose-lose situations from the start. They are always high-profile, huge projects that a ton of executives, salespeople, etc all have their hands in, and everyone is fully aware that the delivered product is going to be a steaming bureacratic pile of shit with a never-ending list of feature requests and bugs, no matter who they hire. Then they hire Deloitte/IBM/Accenture because Deloitte/IBM/Accenture are so big that they are the only companies that will knowingly take on a losing project as long as they get paid handsomely for it.
And if the project is gonna be a loser anyway, why would you waste your best people on it instead of the C-stringers from your offshored development center? At my B4, the major multi-million dollar government projects like that were basically the company graveyard where you were staffed when literally no other project would take you.
For the actually successful and typically much smaller software projects, we actually had a subsidiary company with a much better reputation that would handle the design and development instead of our offshore dev centers.
What is the solution? How can these huge systems be developed economically and correctly? These government agencies don't have the technical experience to hire and manage their own software developers and, from what people say, the big consulting firms don't either and/or they don't care. But a smaller firm that might be able to deliver a working solution would be seen as "too risky" to be awarded a big government contract when compared to a large, well-known consulting firm.
Why does the government not have more than a handful of people at 18F? Do they think computers are a fad that will go away? Government should be staffed to accomplish its mission, and the mission requires software. 18F should have tens of thousands of permanent employees.
Often the fix is not to build a huge project. For some of these failures they're trying to migrate from a paper system to something as automated as possible in a single step.
Incremental improvements are almost always possible. People have set up websites to submit paperwork where the forms just send an email or trigger a print job when submitted. It's an improvement over mail and faxes.
The solution is that we stop pretending that large projects are a good idea, that technology for the connected world is set-and-forget and having a ton of layers of people is a good idea. Because none of those things help.
Model the system as objects without writing code. Get the methods and properties accepted without writing code. Write a suite of tests for the model. Implement code.
That’s how complex systems don’t have to feel like monoliths.
I used to work for one consultant company like these.
There were many smart people who had crazy startup ideas and technical ability to handle the projects.
Problem was just that they were very expensive, so consultant company used them to field the projects. They'd send them with a masterbullshiter who would take clues from these tech smart folks and bullshit his way to impress the promoters and get the project.
After that the technical team and bullshiter will try to get another project, so it's not same people who are bidding and winning Vs one who are delivering project at the end.
Tech smart and bullshiter are used for fielding project (these cost a lot of money), while subservient and low tech skills and low awareness people are used for delivering actual project.
And there you said it - outsourced. It takes exceptionally good project managers and tech leads to develop a successful product using offshored labor. And from most situations I've experienced or read about, there are very few such managers/leads.
It was CGI Federal (the US version of CGI) and QSSI (eventually bought by Optum part of UnitedHealth Group). CGI was the lead contractor and got it by low balling on the website. When CGI immediately got behind on project deliverables (2 years before launch) they started forcing all the people to work 7 days a week which immediately caused every good developer to leave. All that was left were the most inept H1B developers I've ever seen (some literally couldn't touch type). I joined 2 weeks before launch as a product consultant. There were basically no senior people working there at all. It was a bunch of middle managers and either inept H1Bs or fresh college students brought on as subcontractors. There is no way it could've succeeded in any alternative universe. Basically no one cared about the product at all, everyone only cared about their slice of the pie, and by everyone I mean the tens of different subcontracting companies and middlemen that were involved. The only reason CGI even lost the contract is because it was such bad publicity HHS had no choice. CGI ended up suing and I'm pretty sure they were given alternative contracts anyways. There was never any accountability for anyone involved. When I quit I made sure the next job I had wouldn't get close to government contracting.
I don’t think the article convinced me that Deloitte is specifically at fault for these systems failing.
They were paid one time to design a system, seven years ago, to handle probably less than 1/10 of the kind of volume that we are seeing due to the pandemic.
It really isn’t Deloitte’s fault that our government isn’t willing to just pay civil servant employees to build and maintain technology in house.
I don’t claim to know why the government is so sub-contractor happy but I guess that’s the way it is right now.
Also, the government needs to stop thinking about software like it can be boiled down to a one-time cost. Building software one time and leaving it alone to rot for a decade doesn’t work.
1/10th isn't even close. In my state you were looking at 40x the volume. And you have a good point, that much more traffic is probably going to break a lot of systems.
Not to mention, even if the systems were only down for a few hours or a few days, the government agencies just don't have the manpower to process all those claims. Especially when people are working from home or not working at all.
Then you have the massive fraud. A perfect time for scammers to apply for unemployment benefits.
Remember the people who don't have computers and call their claims in via touch-tone. Those systems all crashed as well, or had busy signals for weeks.
Not just unemployment systems. There's a graveyard of broken systems made for governments, all over the world.
I genuinely believe that their business model is to ship out unfinished and broken software, and then spend xx years billing them (steeply) for upgrades and maintenance.
That's of the few explanations that makes a lot of sense.
edit: This isn't only aimed at Deloitte, lots of habitual offenders out there.
It's not clear how so many government procurement systems are broken everywhere.
As a counterexample of some procurement did work, just a tiny one, my state DOT rebuilt a major road over a decade. The whole project was many millions broken out into milestones and each milestone had penalties for being late and rewards for early, presumably with some quality gates too. The contractors got it done. So it is possible, just not common.
Another example is NASA batting 1 for 2 on Commercial Crew, although politics messed up the ULA half of that.
> It's not clear how so many government procurement systems are broken everywhere.
One of the two major political parties in the USA actively sabotages government projects. The other one frequently fails to wield the tools of government well. Some of the procurement details might not be clear, but it's not terribly hard to figure out where the upstream problems are.
I have an insider's perspective on one of the systems mentioned in the article, and from what I could see, politics played no part until everything went off the rails and they started flinging shit in the media.
Things start out with a few in-house sysadmins hacking together some scripts to start batch loading records into a database or whatever. At some point they get far enough along to decide to bring in an external consulting firm to build out a more fully-featured solution.
At that point, it becomes a very ironic situation. The meat of the requirements are laid out by career bureaucrats who work in state government agencies (unemployment, in this case). They are extremely risk-averse because their jobs are typically very secure and, barring any criminal activity, the only thing that can sink a career is some sort of publicized failure involving taxpayer dollars.
So they negotiate contracts for this sort of work with really large, established consulting firms like the Big 4 and include all sorts of requirements and checkpoints and little details that they think are protecting themselves and would drive the average dev shop batty. Furthermore, they want to go from 0-100 right away and convert an old manual paper process to a fully-featured SaaS-style product in one go with a bunch of different feature requests coming from different people.
Thus they ironically end up with a massive risk of failure precisely because of the steps they took to mitigate risk in their minds. They have no prior experience developing software and are usually lacking in overall technical expertise. Really they have no business designing any sort of software solution or negotiating a contract to build one. It's doomed to fail simply based on the premise.
The best outcome would be for a dedicated SaaS company to build something that could be purchased by multiple states. I actually briefly explored starting something like this but couldn't make the numbers work. These departments have very little discretionary budget. Selling to each state would require a literal act of congress, and your TAM is only 53 distinct customers. And because of the inability to quickly make a purchase, they all seem to start out with that in-house hacked approach which naturally leads to paying for consultants to "improve" it, not starting over with a product purchase.
> So they negotiate contracts for this sort of work with really large, established consulting firms like the Big 4 and include all sorts of requirements and checkpoints and little details that they think are protecting themselves and would drive the average dev shop batty
To be fair to the bureaucrats involved in a specific project, this approach, including specific staging of particular waterfall-style deliverables (with more complex checkpointing and midproject external reporting and oversight required the larger the project and sometimes on other bases, such as particular outside-of-the-agency, e.g. federal, funding streams) is often mandated by statewide contracting rules (a mixture of general and IT-specific mandates), not something that the bureaucrats directly involved in a particular project impose.
And while in many cases ill-advised and counterproductive, each of those rules is typically reactively developed in response to and as a means of mitigating repetition of specific instances of negligent, incompetent, or outright corrupt contract administration that occurred in the past.
Question for the HN crowd: My apologies if I offend anyone (sure I will), but I can't imagine a good programmer wanting to work for any of the big tech consulting companies. They don't pay very well, their work is generally boring but tedious, and their corporate cultures are not known for attracting top talent. If you're a good programmer, seems like the top options are:
1. Join a FAANG or other well-regarded software company (e.g. a Salesforce-like company)
2. Join a startup
3. Research (academia) or a company that does big "research-like" work (e.g. the NSA, SpaceX, etc.)
So I just don't see any way this kind of work is going to get done and not kinda suck.
> So I just don't see any way this kind of work is going to get done and not kinda suck.
CodeForAmerica has identified the core problems with US project procurement: waterfall projects where all of the detailed requirements are defined before the contract and the programming experts (designers and developers) get involved. The person writing the contract frequently has no idea how the best tech companies organize projects. Also, they frequently pay contracts by hours worked, not by project outcomes.
CFA will coordinate with government departments to put developers+designers in a place to prototype+build a better mousetrap. The problem is that legislators and departments resist giving that autonomy to the developers+designers, so the people developing requirements are just ticking off checkboxes, not actually building an optimal feature for the end-user.
It's not true that consulting companies pay the least. They're not paying like the FAANGs or unicorns but they're higher than working for industry. As an example, at my level I can make ~200k as a consultant but only 150ish as a direct employee. It's crazy because they're willing to pay ~400k for me as a contractor.
When it comes to consulting companies, even if the actual work is terrible (there are exceptions), there will always be people who need the opportunity. Either they are good developers but are just in a hard time and need something to pay the bills (in which case the incompetence of the existing team is an advantage because due to their skills they can get away with doing very little actual work) or they are proper monkeys that will just fraud their way into the industry and then blend right in with the rest of the idiots in there and do the minimum amount of (mediocre) work required to collect their paycheck for the rest of their lives.
Define "good salary". In the DC area, they pay well. We don't have the unicorn FAANG salaries you see in SV or Seattle, but $150k+ isn't uncommon for somebody with a few years under their belt.
The biggest problem I see with them is they tend to throw bodies at the problems. Many junior employees trying to build complex solutions isn't efficient. Even if those junior employees are smart.
Also, not everybody can get hired at a FAANG, sometimes because they're good but not great, sometimes because the FAANG hiring process is terrible, and sometimes because they don't want to live in SV.
> $150k+ isn't uncommon for somebody with a few years under their belt.
I don't think that's much higher than government employees make these days. In DC an engineer "with a few years under their belt" can easily make that at a startup and be working on much more interesting problems for a much less slimy company. There are plenty of jobs outside of FAANG that pay that or better.
The parent isn't wrong, at least in my experience, nobody who is technically competent would take a job at one of those places (and in my experience none of the technical people coming out of them have been competent).
What's with holding Startups up on some pedestole? They can be slimy and most will fail. Working at a well known consultancy company offers not only financial security, but additional insurance, and pension benefits that a startup would not.
It's also not impossible to work in substeams in such a company that mirror startup processes.
But do Bain, McKinsey and BCG do the type of large implementations that places like Deloitte, IBM and Accenture do? That is, I thought their technology consulting practices were more along the lines of "We think you should move to the cloud, and here is how." and less of "We're going to build out a large benefits platform for you."
McKinsey and Deloitte are more about accounting and management consulting. I don't think they do many $100M software projects, unlike IBM/Accenture/Atos/Tata.
Let me see if I can address some of this. Trying to be respectful in my response as well.
Source: work for large tech consulting company, don't classify myself as a "good" programmer, just somewhat of one.
"Don't pay very well" - not necessarily true. Comparable to FAANG? Sure, but industry? I'm not as confident. Think of the internal IT group of your favorite retailer, bank, insurance company, whatever. That junior staff doesn't have nearly the income potential a junior consultant does. Not everyone has the opportunity or the "credentials" to get in front of FAANG, NSA, SpaceX, etc.
"Work is generally boring but tedious" - I suppose this is how you define boring. Is working on social security and benefits systems glamorous? Maybe not. Banking infrastructure? IT Governance systems? Payroll? Project planning? Financial management? Reporting? I know people who obsess over some of these systems. People build careers on working with Oracle/ SAP; I'd hope they don't view their own work as boring. I'd rarely describe large scale implementation/ strategy work as tedious. There are some monster projects out there, no doubt, but most projects have well defined guidelines, milestones, etc. In almost all of the companies I've consulted with, our team is moving 3x as fast as the client because usually our projects are scoped as thinly as possible to win bids (cheaper seems to win a lot, unfortunately), and we simply don't have the luxury to work slowly. Time and materials may be an exception to this, but those are not as common as you may think.
"Corporate cultures are not known for attracting top talent" - this is a tough one to answer because I think it depends on how you define "top talent". If you're using undergrad as a benchmark, then sure, big tech consulting firms don't usually pull Harvard or Stanford grads (though they are in our ranks). Consulting work is a balance between being an effective conduit between your client and internal/ outsourced engineering teams. I'd argue a "10x programmer" may not be very effective in a client-facing position discussing strategy, etc. This is not to say the engineer isn't valuable, but let me take another stance: "top talent" programmers want to work on bleeding edge things, and Fortune 500 doesn't have much of that to go around, because it's a risk. I can't tell you the number of clients I've worked with who still use a mainframe with a COBOL developer. Even doing simple things like data interoperability between systems can be a nightmare because enterprise systems may not have native API's, so you end up hacking together these weird FTP-then-scrape type solutions to move data around. Is it the way it should be? No, but you aren't there to rip up their whole ecosystem of tools because you want to use a REST API instead of SOAP.
Regarding your options:
1. Listed earlier, but not everyone is lucky enough to get to interview with FAANG. Speaking from experience, my background didn't get me in the door out of undergrad.
2. Doable, sure. You're a bit location bound here... statistically speaking, most startups are in the Bay Area. Maybe NY too. So if you aren't already in those areas, you're not around a lot of seed companies that you can get in front of. Some folks also may not be willing to take the risk that a startup entails. Would you take $50k + options in a 5 person company versus $100k + bonus from a big company right out of school? Not saying one is right, but there's only a subset of folks that would/ can consider this route.
3. Research/ research-like work. Just a prohibitive as 1 I'd say. It's unfortunate that college is such a "gate" for many of these opportunities. Trying to join a company/ institution that doesn't actively have a presence on your college campus is an uphill battle. Doable, yes, but the odds aren't in your favor. Particularly if you don't know anyone and it's just a cold application.
Would love to continue the conversation. As a closing note - consulting is seen as a great short-term gig. Everyday I get to see how the world's largest companies operate, I can be industry-agnostic, and work in a variety of roles from strategy through execution of IT projects. I would argue a consulting background is an excellent launchpad to the "next" thing, which could very well be any of your three above.
Late to this party, but also does GP think that everyone working in FAANG is doing glamorous work? I did a year in a big 4 and I know that my work, over 4 years later, is still in production at places that my parents and friends buy products from. Meanwhile, I've met people who work on a team to do logistics for a niche product at FAANG. They're talented people, but let's face it, not everyone is working on the cool stuff.
I've worked for both a Big 4 as a tech consultant, as well as a FAANG. They both have their pros and cons.
The first thing is that they really have two completely different cultures that appeal to completely different people. The way I describe it is that the consulting companies are basically an extension of Business School, while FAANG is an extension of Engineering School. The culture at a consultancy is generally a lot more social, with huge, alcohol-filled parties being the norm (if you think FAANG parties are big, B4 parties are twice as big and happen 10x as often). They also put a lot more emphasis on "group work" where the main way of getting tasks done is to get together in a room and work side-by-side while having continuous review and conversation.
If you're a person that likes to often go out to bars, have fancy team dinners at chic restaurants, work in groups, and spend a lot of time building relationships with your coworkers, you probably would like B4. What I just described might sound absolutely horrible to many HNers, but there are many people that really dig that lifestyle.
On the other hand, if you are a stereotypical computer engineer that would rather go home and recharge rather than go to a bar, maybe have a nice company-provided dinner every now and then but not every night, do your best work heads-down without any interruption or talking, and think "work is for work, not for making friends", then you certainly wouldn't like B4 and would probably like FAANG.
As for pay, B4 starts out a bit lower than FAANG but salary at consultancies grows much, much faster. Getting a promotion every 2-3 years is expected, with large salary bumps and large bonuses even in non-promo years. When I left B4 to FAANG, my pay almost doubled, but due to the way B4 pay scales, if I went back to B4 now I would likely make more at B4.
Career progression is also very different at B4 vs FAANG. There is no "Individual Contributor vs Management" track at B4. Everyone starts as an IC, and then slowly starts taking on more management responsibilities. For me personally, I really prefer this because it means I get to be a manager that makes strategic decisions and mentor a team, but also still do "hands on" work as much as I want. FAANG, on the other hand, I have to make a decision whether I want to be an IC (where I will only have some mentorship/manager responsibilities, if any) or Manager (where I will be doing considerably less hands-on work, if any).
And last, tech consulting at the big consultancies can genuinely be fun. You get to hop around to new clients/projects every couple months, working on and learning new types of work. If you do find a type of work you really like, you can specialize in it and stick with it as you please. For some people, switching clients and meeting new teams to work with, being "the expert" in the room, and always running into new problems is fun.
Anecdotally, when I had a stint in consulting I eventually learned that whenever a client wanted estimates the only way I could get the contract was to bid about 1/2 of what I knew it would actually take, else one of my competitors would do the same and they would get it.
These same clients would usually be giving me much praise when we'd eventually deliver 100% over budget, because that was "much better than their last contractor"...
A lot of times the buyer isnt even qualified to do that and so the seller has all the power.
I had a recruiter from Deloitte approach me, telling me how great I'd be for the role they had in mind for me, etc. They wanted a phone interview immediately. I said fine, I'll go for an interview. Then I got an email thanking me for my application but my skillset was unsuitable. Except I never applied for any role, never gave out my CV (although it is out there in the wild). Absolutely wild. They can't tell their asshole from their elbow.
My wife works for a state in a department using software being made by a company mentioned. The weekly meetings and calls are so astonishingly slow and deal with such meaningless minutiae that I am sadly not surprised at this article. Heartbroken a bit, nit not surprised.
A bit tangential but I recall working with a company that wanted to use a particular very expensive brand name closed source software framework to build their ecommerce site. The going price on using this was in the half million ballpark. Working with this framework I was incredulous that anyone would actually want to buy this framework -- also none of the sites I have seen built on this actually work very well. But the force of the brand name and the audacity of the asking price just seem to mix well I suppose. If you are in a corporate setting with little technical ability, and more money to spend than brains, 'you won't get fired' if you choose to go with vendor X (not IBM but point stands) -- even if vendor X is going to fleece you of millions while providing little value. Its just the safe thing to do.
I worked for Deloitte for a year. I'm convinced their business strategy is to put their worst foot forward first, find the clients that will put up with it, and then milk them dry.
I worked at one of Deloitte's sister companies (another one of the Big 4). I used to think the same way you did, until I later had experience working with other software consulting companies.
As bad as Deloitte et al are (and holy shit are they bad, especially at producing software), they are sadly still much better than most of the other clown shops out there.
I have worked with some amazing consulting firms, but those came with a $250/hr per developer price tag at the cheap end. It says something though that large companies wouldn't blink at paying that much for a lawyer or skilled mechanic, but when it comes to software they'll go for the cheap option and then complain about bugs.
Tbh, software consulting towards gov. entities is a almost guaranteed shitshow, from all perspectives. It's just a recipe for disaster, unless handled very carefully.
On one side, you have consulting companies that want to maximize their billing - no mater what.
On the other side, you have tons of red tape to navigate through, clients that may not know what they want, or just incompetent people in charge of writing the spec. And with huge expectations, at that.
The question I have is if there is even a market for people people to build good software. It doesn't even seem like a money issue when I hear about tens of millions of dollars being sunk into projects that go nowhere.
1. Startups, where you are the founder.
2. algo finance
3. MFAANG, v. well-funded startups, as an employee.
It's very hard to "hire good programmers," because we tend not to agree on what "good programmers" even look like (see the 1000+ thread on tech interviews from yesterday). If you want to get well for writing software, those top three are your options. #1 and #2 pay you for value, #3 pays for interview performance.
My genuine experience is that when you are building something as large and high-profile as a large state's unemployment system, there is absolutely no possibility for the resulting software to be "good" anyway. This is especially true for government clients. And the reason that IBM/Deloitte/etc are always the names getting trashed for these is because those are literally the only companies big enough to take on such a project in the first place.
All of the smaller projects I took part in had much better results, but you're much less likely to hear about some random mid-size company's app that is working just fine on the front page of HN (and of course, those smaller projects happen less often anyway since Deloitte really focuses more on getting those giant projects).
It's a culture problem. There is a belief that the size and complexity of the project must match the size and complexity of the client. So small, nimble companies don't even get considered. The budget gets set before any evaluation of the problem is done.
You could give that $10mil to a 250 person consulting company and they'll make it happen. But that will never happen. Hell, the SOW was probably written by some at Deloitte, stuffed with whatever cross section of requirements are unique enough to make them the only vendor available that matches it. Then they will bleed the contract by putting expensive partners on it who literally do nothing but suck oxygen out of the room. That's how they get around maximum multiplier of salary regulations on government contracts.
>There is a belief that the size and complexity of the others must match the size and complexity of the client. So small, nimble companies don't even get considered. The budget gets set before any evaluation of the problem is done.
My experience is that it's more about consistency and "knowing what you're going to get". When you hire Deloitte, you know they've done this exact same type of project 100 times before, and each time it turned out terrible-but-acceptable. You're 90% sure that it isn't going to result in an amazing product, but you're also 90% sure that you'll at least get something somewhat usable for your money (and if you don't, you know that Deloitte is going to be around long enough to sue them).
With a random 250 person consulting company, you have no idea what you're going to get and much less confidence that you'll end up with anything usable at all. Hell, you don't even know if they're going to still be in business a couple years from now. And with that much money at stake, that's a risk nobody wants to take.
They may have sold 100 contracts for something that looks vaguely similar, but there will be zero continuity between the team or the code. At most, you'll get the same managing partner overseeing the sales end, taking the credit for it. This concept of the "safety" of Deloitte, et al, is an illusion.
Cut the project budget to 10% but spend it 10 times. Make 10 smaller firms compete to make the thing and you will get what you need.
>They may have sold 100 contracts for something that looks vaguely similar, but there will be zero continuity between the team or the code.
There is such a thing as institutional knowledge, and if there is any benefit of a consulting company like Deloitte, it's that. You might not be getting the exact same developer that worked on the last project, but other than for small niche areas, there is going to be an entire bench of consultants that have nearly identical experience, and also have an entire library of past deliverables, training material, etc to quickly get up to speed.
And I can personally attest that these companies re-use the fuck out of any and all code, documentation etc they can. It's practically company motto that no matter how much a client insists they are a special snowflake that requires brand new, custom made designs, they certainly are not. No matter how much you think your pet project hasn't been done before, I can almost guarantee that it has, multiple times, and they are probably just going to take the deliverable from the last time they delivered it, Ctrl+F/replace the previous client's name with yours, and send it over to you. For better or worse, that's where the consistency comes from.
>Cut the project budget to 10% but spend it 10 times. Make 10 smaller firms compete to make the thing and you will get what you need.
Now this is a way to end up with an absolute clusterfuck of a project. You absolutely do not want to do this. Not only will you end up with 10 disorganized, disconnected projects that are each only 1/10th of the feature set you asked for, you're going to have to hire 10x the amount of people to oversee these 10 consulting companies (you'll probably have to hire an 11th consulting company just to PMO the other 10), and your entire project is going to turn into ruthless backstabbing as each of the 10 consulting companies try to sabotage each other or steal credit. It's not going to be some "friendly competition" where they all inspire each other to be better or some fantasy like that.
Since leaving B4, I've been in several situations where the client had the same mentality you do (well, with ~4-5 consulting companies rather than 10). It always, always was a colossal failure and in one situation directly led to the project buyer getting canned. In another situation, they ended up cancelling the contract with the various smaller companies and ended up just hiring a single larger consulting company to come in and clean up the mess, anyway.
Who, in your experience? Because the smaller shops tend to have brighter people, often with the founder involved (or one-step removed), in my experience
Smaller shops are a huge hit-or-miss minefield in my experience. Some of them can be rockstars, but those rockstars are typically already 100% engaged with a long-term client or are on the road to be acquired by Deloitte et al. On the other hand, most of the non-rockstars might as well just be staffing agencies that aren't any better than hiring a random part time dev from craiglist.
The sweet spot in my experience is the mid-size consulting companies, like around ~500 employees or so, because you know that they are at least good enough to stay in business long enough to grow, but they also aren't yet big enough to be bureacratic, incentive-warped, soul-sucking machines like Deloitte/Accenture.
I worked for them as well and lost count how many times myself and my product owner were in calls where upper management were promising features that had a fundamental conflict with major design decisions. Leadership absolutely did not want to hear about it and became hostile. We both left the company and are much happier. Terrible company, when I hear ‘run like a startup’ I have flashbacks.
Purely out of interest, what does the tech stack tend to look like for these big projects?
Do they use open source components and frameworks? Are they using Oracle?
Do they tend to have good test coverage and continuous deployment?
What kind of version control do they use - what kind of branching / release model etc?
I've been immersed in the GitHub-centric open source world for so long I don't really have a feel for how these big Deloitte government IT projects work.
> "Purely out of interest, what does the tech stack tend to look like for these big projects?"
My experience being on executive panels, etc. receiving pitches from firms like Deloitte, et Al., is that they're super buzzword-compliant nonsense because what they're actually selling is a vision of "innovation" and "modernization" to middle & upper management of all manner of enterprises going through the corporate/institutional version of a midlife crisis. These folks want to feel relevant and cutting edge, and firms like Deloitte sell straight into the strange mix of vanity and insecurity. That said, I'm sure I'm witnessing a fair amount of sample bias in that read, as it's the big corporate contract side.
I have no idea what the governmental side looks like.
It's the same. Bureaucrats measure their self-worth by the relative size of their budget compared to their peers. It's not their money, they don't care about cost, except that the more they can spend, the easier it is to get a budget increase and gain rank in their peer group. Consulting firms like Deloitte help them burn their budget every year.
They're almost always building an API to a legacy system and a modern front end. There's a bunch of BPM solutions like Salesforce, Pega, Appian, IBM has one, Oracle has one, and Microsoft has one as well.
API layer is usually Mulesoft.
They usually pay 6-7 figures annually for licenses and then $75-250 hourly rates for developers. It adds up very quickly.
No, for the application layer. Think like Rails/Django but a lot more features specific to integrating in an enterprise environment and process management.
In all seriousness you’d probably have to ask the subcontracted dev team in China what they’re using because the actual Deloitte employees won’t know and likely won’t care.
More directly to your question, for a giganto consulting shop like Deloitte, the tech stack is whatever the client wants. They have people for everything. They'll also lie and say they have people for anything they don't have, then outsource it again. Even in the hyper-niche field I was working in, we weren't the only team in the company working in that field. Do not underestimate how little the left hand knows what the right is doing in a partnership-organized company. The people I worked with in my niche were some of the best people in the world at that work. The failure to deliver projects was 100% a problem of management.
Deloitte has so many fingers in so many pies that you literally spend two, separate weeks every year letting them paw through your financials to see what stocks you and your spouse own, what business interests you have, to make sure there is no supposed "conflict of interest". It never came up for me, but theoretically I could have been barred from working on any particular project if, say, my wife's 401(k) had an index fund that had stock in the client company.
Some sort of IBM in the background. Contract gets build out to have a web interface built for it. You might see PHP or something (Zend makes a PHP Server product for IBM).
However, like at any non-tech business, you will see crazy shit. Maybe the guy who worked on it wasn't that great at PHP or wasn't real good at web programming, or didn't follow best practices with databases.
A 'good' programmer could come in and clean up the code, clean up the data structures, introduce things like memcache, and speed stuff up significantly.
There wasn't anything particularly wrong with the technology they used. It's the fundamental incompetent management and the infrastructure that protects it that is the problem.
At Deloitte, you don't get to be a project manager because you're any good at managing projects. Hell, I don't think they even really hire real project managers. You get there by glad handing with partners and managing directors.
You as the developer also never get any access to the client. You never get any real set of requirements. At best, you'll have some stuffed shirt make up the entire concept out of thin air (or "user personas", as they called them), present it to the client at a kick off meeting, and then never take any feedback on it.
If, in the course of development, you discover your PM knows more about hair products than software products, there's no way to change the plan. And when it all invariably fails, you as the software developer who doesn't have any drinking buddies in the partner suite are blamed for it.
These folks are also responsible for your yearly review through a system called the "scatter plot". It's basically random noise with a linear progression forced into it (which mirrors the institution's complete lack of competence amongst management). At the end of every project, you hunt down your PM and coworkers and make them submit a "snapshot" about you. They then evaluate you on completely subjective scales like "if it were yours, would you give this person money". The theory is, over the course of a year, you'll work ok multiple projects with multiple PMs and any interpersonal bias will average it. They think of this as an objective evaluation system.
But if you're working on software, they think you want to build report and work with the same people all the time (not an unreasonable assumption). So in the software side, they put you on 3-6 month long projects, all of which are with the same PM. So you don't get very many snapshots (because there aren't that many projects in a year) and they're all from the same person.
Half my team and I got laid off because our management was incapable of selling projects. I was on the bench 40% of my time at Deloitte. Tried to make use of that time to make useful tools, which just got me into more trouble. Got told "libraries are a waste of time". Tried to bring in some of my contacts for potential contracts (I had been freelance before Deloitte and had a few leads come back after I joined), got yelled at for that. Tried finding projects within the company, got yelled at for that, too. Never did get told how I was supposed to be using my time.
And I mean literally yelled at. I got sat down and scolded like I was some child in school.
The only saving grace: my former boss got fired for harassment about a month before I got laid off, and I'm pretty sure I helped it happen. Though it had been there 8th complaint against him in so many years, so actually, probably not. He probably got the can because he missed his sales targets 8 years in a row.
Yes, eight.
Stay away from anywhere that is organised as a partnership.
> I was on the bench 40% of my time at Deloitte. Tried to make use of that time to make useful tools, which just got me into more trouble. Got told "libraries are a waste of time"
Your boss was right. It doesn't make sense to build reusable libraries in a body shop like Deloitte. After all, that would make the next project cheaper, with less billable hours. Not to mention all sorts of crazy trying to work out who owns the IP, what rights does the customer have over your library etc.
They had enough ways to pad a contract that if they really want to lie about billable hours like that, they can just not involve me at all.
IP rights aren't hard. It's just that the contracts move in a linear fashion. If management paid any attention to what was going on underneath them, they'd know not to give the common code away with the contract. But they don't pay attention. They aren't even in the same room for more than 2 days out of a month.
Doesn't MongoDB still has major issues with not keeping the consistency guarantees they have potentially just "forgetting" whole batches of data and stuff like that?
I kinda remember having read something like that two or three month ago...
I've been on both sides of the argument and there's plenty of blame to go around.
On the client side there's the utter lack of clarity, plus in-fighting and adversarial dynamics between departments and execs.
On the consulting side, it's just too easy to milk all the issues on the client side. Including playing one side off another - like lawyers in a never ending divorce settlement.
For all the people talking about how much BS there is in big consulting, dont any of you think there could be a business model that could usurp the way the BS'ing boys club siphons money through narcissism, lies, and incompetence by replacing them with better quality development for a smaller price tag (but still with very well paid developers - from the looks of it)?
I'm personally thinking of a team building model where developers can join together and form teams to bid on projects and win financial rewards for delivering, reviewing, and passing quality benchmarks, with some sort of rating system in place to keep bad actors and bs-er's out.
> form teams to bid on projects and win financial rewards for delivering, reviewing, and passing quality benchmarks
I work in a government agency and a friend is at another one. That would not work simply because the government is terrible at defining the project and actually hammering things down. I have been working on a project for nearly a year as an internal developer. I couldn't with any reasonable level of clarity tell you what it is my project needs to do.
An immediate problem I see with this is the realistic estimates problem. If you go to your average deloit/accenture customer and offer them the services of this anti-BS'ing boys club collective, and then deloit swoops in and says they'll do it in half the time and tick every single checkbox in the massive excel RFP including the contradictory requirements and the requirements that flat out make no sense, then guess who is likely to the win the bid? In the end it will be your deloits and your Accentures that win and then predictably the project will run long, not tick every requirement, and end up costing double.
> says they'll do it in half the time and tick every single checkbox in the massive excel RFP including the contradictory requirements and the requirements that flat out make no sense
I see your point, thanks for the counter arg, I can't refute this.
It seems the key strategy is to make the payer aware of how much BS they're paying for... but knowing all the people I know, I know that's just flat out not going to happen save for a very small percentage of people.
Between the target billable rate they give you and the amount of time you spend on regulatory compliance, plus the fact that your own vacation time counts against your billable rate, you're guaranteed to have to work at least 10 hours of unpaid overtime every week just to hit your target.
I accidentally started a small riot during my introductory training period when I did the math in my head and blurted it out in the room of 50 other new people.
The guy running the training left shortly after I got canned, started his own business, and even sent me some work. He hated that place, too.
Are people surprised by this? This happened also over two years ago during Hurricane Maria where the contractor failed to supply food to Puerto Rico.
The reality is that all consulting firms are trash. From big to small. They pay their employees far less than working direct and abuse foreign workers looking for a good career. They don't care about the product or the people they work with. I used to work as a contractor and the people I worked with was fantastic, but what I saw from the larger companies and the way they treated their work was awful. The incentives are just not aligned properly because the higher-ups only care about prolonging contracts and siphoning more money, not about actual quality.
For the government we'd be in a better place if they just hired people directly rather than going through contracting firms to play telephone and get a non-working product out of it.
I saw a Vox video recently that hinted that the Government WANTED broken unemployment systems after showing some of the bugs that existed in their interface. It infuriated me, because they mentioned Deloitte built it and I (like others have mentioned in this thread) know they are infamous for bad software.
It appeared to me Vox wanted to make people mad at some bogus government conspiracy that a specific state had against it's people. Also that they don't respect the complexities that go into software development.
Florida purposely built a broken system under Governor Rick Scott. It was engineered to make sure that as few eligible people could access it as possible. It is socially and technically engineered to make it so that people give up trying to claim the unemployment benefits they are entitled to. Florida used the system to keep their unemployment numbers artificially low coming out 2008 recession.
I'd note that while a broken unemployment system would affect statistics on the number of claims made, the official unemployment rate is calculated differently, using the Current Population Survey, and doesn't incorporate data on claims:
The state of Florida cares less about the unemployment rate and more about whether they have to pay out of their unemployment fund. The federal government is different because they just increase the deficit whenever the situation is dire.
So you contend that the ~7x delta in unemployment benefit uptake between states that they start the video with is... solely due to bad Deloitte software? That seems no less suspicious to me.
I haven't seen the video, but the current governor's Chief of Staff complained that the previous governor (both Republicans) actively created a difficult process and legal requirements which were very high hurdles to cross. Part of that are the anti-patterns or dark-patterns in the sign-up system, but the qualification requirements aren't as easy to qualify for as many other states.
Biggest question in my mind is why state bureaucrats are given the latitude to think it's a good idea to build their own one-off solutions. There are 50+ markets for these sorts of systems. It makes absolutely zero sense for them each to build & maintain their own software products. Of course big consulting firms are going to _leap_ at the opportunity to bill $50M per state and then throw together products as cheaply as possible.
States need to start working together on these platforms with in-house technical teams that can actually own the systems long-term. Or we need to make it far easier for them to purchase SaaS products like any sane private market.
NASWA is good at facilitating knowledge-sharing, but they are no better equipped to build a software product than any state IT department. Each state is still ultimately handling "modernization" in their own way. There's very little appetite for COTS for whatever reason, and little sharing of technical components between states.
If a motivated small company wanted to work with NASWA in a "not business as usual" approach, perhaps with 18F involved, I believe it could be done. Part of their role is to reduce duplicative efforts across states and try to establish best practices.
The problem of course is how many small companies are knowledgeable about the UI IT systems, and can put together a strong proposal? It's a bit chicken and egg. Yes, NASWA could push this harder if they really wanted to, but it's hard to manage a consensus oriented organization.
(I was not implying NASWA was going to build things themselves...though if I were king, I would have them develop an open source UI platform, and then give it to the states to customize for their individual scenarios. Companies could compete on how well they make those customizations, and support contracts. Of course the trend seems to be going in the opposite direction, unfortunately... c.f. VistA -> Cerner).
Yes, there are sharks out there who do poor work. That's why we should expect the government procurement process to be competent. If the government hires a shark, the shark is not suddenly going to change. We should be holding the government responsible for hiring the shark in the first place.
When a government hires a consulting firm to perform a service, the responsibility does not shift to the firm - it remains with the government.
When procuring the system, did the government provide adequate requirements? Did they perform proper diligence in selecting the consultant? Did they provide proper oversight of the work?
The answer is rarely 'yes' to all of these questions.
Deloitte is a horrible consulting company. I worked at a company where someone senior in my org had a husband who was a partner at Deloitte. We ended up going from barely using consultants to using them for everything. They even billed us for our own work and most of the time just ran down the clock on the contract and left without finishing the project. Pretty sure what our senior leader was doing was shady and wrong. I ended up leaving and noted this situation in my exit interview but nothing was ever changed. Conflicts of interest aren’t taken as seriously as they should be.
So I work for an APM vendor. COVID-based unemployment has been excellent for our public sector sales (as well for other APM vendors who are the competition). One of the sales people went to (I think) every unemployment insurance department in every state and pitched them our software. Quite a few were standing in a smoldering wreckage and they gladly accepted to see what they could do.
That is how I personally got to look at the guts of unemployment insurance system of Kansas and Florida back in May.
The KS system was ... well judge for yourself. It was run on 3 (yes that's three) servers - a 8 CPU 32GB web server hosting UI and IVR web service, a 4CPU 48GB service layer hosting a WCF service talking to mainframe, and some auxilary 2CPU 4GB server doing some little batch stuff. Yes, that's right, no redundancy. That's for the state with almost 4 million people. The bottlenecks were all in the mainframe which was reached by via screenscraping host key-sending software, which was doing GC.Collect on every end of the given session. It worked well when nobody was using the unemployment system but once half the state of KS lost their jobs, everything melted down. The vendor for the host-key fixed the library by commenting our GC.Collect. Don't know what happened afterwards.
Then next week I got a look at FL which is the uFacts system built by Deloitte https://www2.deloitte.com/us/en/pages/public-sector/solution... mentioned in the article. There was no Deloitte there anymore and from what I can tell there hasn't been for 3 years. That one was is a standard 3 tier (web layer, services layer, database layer) hosted by more than enough hardware on web and services layer - although very much an inconsistent hodgepodge of IaaS machines from Azure, but it all ended in the big Oracle database that had some atrocious performance under heavy load. All timeouts were cranked up to the limit - I am talking WCF/WS timeouts north of 600 seconds, and Oracle DB query timeouts set to what I think is infinity - so things just backpressured out of the slow Oracle queries like there is no tomorrow. But optimizations were pretty obvious. They had an "Messages" table that was adding all the things that could happen to a case that was north of 100M records and suffered tremendously on the inserts. Pretty easy sharding or just archiving opportunity. Here I and couple of my colleagues were able to get to the team leads and developers in charge to present our findings. Even though the governor was yelling at them and the lawsuits were filed and people were struggling, the response back from the dev leads was remarkably devoid of urgency. "well that we already know about this, we'll hoping to be getting to this 3 sprints from now". We all shrugged and moved on.
It's always worth remembering that large firms like this are not hired for their ability to deliver products but the free underpants the brand supplies to the hirer.
It's never about the delivery. It's about the backside covering.
Quote: "Deloitte also faced blowback in 2013 for an unemployment benefits system it built in Massachusetts. There, the benefits system arrived in 2013, two years late and $6 million over budget—only to be filled with glitches that erroneously cut workers’ benefits, according to the Boston Globe."
Ha! I don't think those were glitches, but a clever plan to siphon here and there a bit of extra money. Kind of "you cut them a small amount of money, you get some, I get some" deal between state officials and Deloitte. Easy to blame computers when you do this and have experience. This is a classic case of evil genius at work.
IR35 reform in the UK(every contractor must pay as much taxes as full time employee and no expenses allowed) is a conspiracy of Deloitte and the government
Alternative title: Incompetent bureaucrats wasted tax payer money, as usual.
There's a saying in the home building industry: "If the boss says put the toilet on the roof, you put the toilet on the roof." So often in government contracts, the people writing them are the complete and utter definition of incompetent.
It’s ironic. In the USA they freak out over the idea that someone might receive more unemployment than they are entitled to, but they will happily waste millions of dollars paying someone to do a shitty job.
Tangent: during my last stint on unemployment, there was this class we were required to attend. In that class, they repeatedly told us that we are not entitled to unemployment benefits. They didn't exactly offer an alternative vocabulary.
It certainly fits the government's definition of an entitlement[1].
It's also structured as a non-optional, government-ran insurance scheme predominantly funded through explicit payroll tax obligations[2], which employers aren't entitled to opt out of.
As long as you meet your states particular flavor of requirements, you're entitled to your unemployment insurance benefits. Despite any statements to the contrary states may push in their mandatory unemployment class.
That's absurd. They take it out of your paycheck every month, it's paid entirely from employee wages, it can't be spent on anything but unemployment, and you get benefits proportional to what you contributed. It's literally your own money that they're giving back to you that they took for the sole purpose of forcing you to save money for if you lose your job at some point. It's like if I asked to borrow a pen, then when you asked for it back pretending like it was some benevolent gift.
It's definitely not a line item on paychecks in my state. It's structured as an insurance premium paid by companies to the state. The company's premiums increase when there's a claim against the "policy." Smaller companies like to try denying claims in an attempt to prevent this increase.
One could certainly argue that if this 'premium' didn't exist, it could be paid out in salaries.
That's right, I recall something like that. I think states administer their own unemployment system even though it's mandated by the federal government. It's a payroll tax where I live. It's calculated based on your wages and paid based on the amount of insurance you've funded regardless of whether the regulations/company decides to display it on the calculation or not. It might also get buried under something like "state and local taxes," too.
Unemployment costs are handled by the employer. It is the State Unemployment Insurance (SUI) tax. All employers are experience rated based on age, number of claims, and win/loss rates in claim protests. It is not taken out of employee paychecks.
Your link doesn't support your claim. The article indicates a crime syndicate was responsible, a group who had already stolen lots of money. The weakness here is not in a UBI-type system at all; money was being stolen regardless of the source.
> The fraudsters took advantage of states that were already struggling to process a flood of jobless claims amid the COVID-19 pandemic and related government shutdowns.
To me, this reads as a negative mark against the US federal government for denying the existence of the threat. Not a result of the aide programs themselves.
> In the USA they freak out over the idea that someone might receive more unemployment than they are entitled to,
There isn’t an infinite amount of unemployment money available. If you allow fraud to run rampant in these systems, there is less money available for those who actually need it.
Enforcing rules and regulations is a good thing for everyone.
Furthermore, many people are earning more on the current USA unemployment system than they did from their original jobs. The system was designed to overpay some people for the sake of quickly delivering the system. (Source: https://fivethirtyeight.com/features/many-americans-are-gett... )
It's embedded in the human psyche. Thousands a day will step over a street beggar to begrudgingly buy a barely palatable coffee from the kiosk they're sitting beside.
Who is "they"? I don't think anyone but corrupt politicians and their cronies are happy about wasting millions of taxpayer dollars. Certainly not the taxpayers.
If the government hired developers directly and promoted those developers to levels of seniority, it would in fact spend taxpayer dollars more efficiently than outsourcing to shitty consultants and contractors whose primary goal is to keep the same zombie contract going year after year.
The government at least has incentives to make the best product they can on a local level. There might be politicking at the higher up, but their goal is to make a product. The incentives for larger consulting companies is to make money and often does so at the cost of cutting corners or delivering broken products. The reason why they can get away with it is because the larger consulting companies operate at a different scale than smaller ones, which is often needed for major projects. Which means then you're forced to work with scumbag consulting companies.
The fact that one extreme is false does not make the other extreme true.
That being said, at some point if you're cutting, or if you don't know where you're cutting, at some point you will hit bone. More money is not a panacea but neither is less.
I guess I don't understand your position other than "just make it run better." What concrete steps do you want to see occur to make the government run more efficiently, and will they grow or shrink the government?
If the government is going to contract out work it at least needs to have the capability to audit the results of such work on an ongoing basis, rather than just get left with a flaming bag of crap after the fact.
To that end the government needs to attract, develop, and keep in-house talent, starting by actually offering competitive compensation.
The obvious question nobody is asking is: why is the government obligated to pay for a broken service that should clearly be in violation of the contract? And if delivering a broken service doesn't violate the contract, why are these contracts being written so one-sided? There's more at play here than a lack of auditing. In the real world, if you don't deliver what you promise, you don't get paid.
Litigation for breach of contract is something that states can pursue. TFA:
> Deloitte isn’t alone in its tumultuous history with benefits systems. IBM, another major player in the government IT industry, was awarded a $1.3 billion, 10-year contract to modernize Indiana’s welfare system in 2006. The state canceled the contract just three years later after complaints of erroneous benefits denials and other problems. Indiana and IBM sued each other over the dispute; the case has not yet been resolved.
> Yet in 2010, IBM signed a $110 million deal with Pennsylvania to modernize its unemployment benefits system. The contract expired in 2013, ran millions over budget, and was never completed, according to a 2017 audit conducted by Pennsylvania auditor general Eugene DePasquale. In 2017, the state took legal action against IBM for breach of contract. That litigation is ongoing.
The problem is that the wheels of justice move slowly, and in the meantime the system is still broken, and the previous system is also probably not up to par (after all, if it was working there would be no need to replace it).
Big or small, inefficient govt is inefficient. Pushing for “small govt” either misses the point, or is dishonest, possibly with an ulterior motive.
I couldn’t possibly know what that motive is, but seems clear to me that a small inefficient govt would need to spend more on private sector contracts than a big inefficient govt, if it wants to do the same work; it would also have less capacity for audit, as gp points out.
>but seems clear to me that a small inefficient govt would need to spend more on private sector contracts than a big inefficient govt, if it wants to do the same work
The key phrase here is "same work." But I don't believe a small government should be doing the same work as a big government. It should be doing less work, have less responsibilities, and taking & wasting less of everyone's money.
Really the "small government" party likes to use crony "capitalism" to line the pockets of the "private-sector" with as much tax payer dollars as they can.
1 billion dollars later healthcare.gov almost works good.
Sure I need to click the same 14 screens just to get to the Health insurance options screen, and sure any time I open a URL on page 4 and hit back, it takes me to the top of page 1....
But government healthcare would surely not run into issues that seem to trouble every other government service. (Can I post /s or does that delete your post?)
It's no wonder why projects managed by large consulting companies are total disasters.