"Scaling Food Distribution Is Really, Really Hard"
No, it isn't. It is fairly straightforward and thousands of companies have scaled across the world.
Scaling a two-sided network focusing on providing a niche seasonal product to a niche audience is really, really hard. It wasn't the food distribution that wasn't scalable, it was the business model.
"We were motivated by enthusiasm for our mission and eagerness to bring Good Eggs to more people. But the best of intentions were not enough to overcome the complexity."
No, you were motivated by aggressive expansion targets to provide returns for the dozens of millions of dollars you took as investment and to make a lot of money.
"The single biggest mistake we made was growing too quickly, to multiple cities, before fully figuring out the challenges of building an entirely new food supply chain."
This is a repeat of Webvan. Webvan had a good idea, but tried to "scale" fast. They had about 3% market share in 30 cities, and needed 30% market share in 3 cities.
The downside of venture capital, entrepreneurs need to remember that investors prefer trying for billions and failing, and consider trying for millions and succeeding a failure.
I've read a few bitter stories about people with ramen-profitable businesses who got a late-stage investment into their bootstrapped company. Basically you spend years slowly kindling a small fire, and the VC comes over a dumps a truckload of lumber on it. Might get the bonfire they're looking for, but they don't really care if they just smother your campfire instead.
There's a reason webvan was tempted into scaling too fast.
For mass market groceries, there are two virtuous cycles a retailer needs to trigger. The first is:
1. Get higher customer density
2. more deliveries per driver shift
3. reduced delivery costs
4. attract more customers with lower delivery prices
5. go to 1
The second is:
1. Get more customers
2. better economies of scale and more negotiating leverage with suppliers
3. reduced grocery costs
4. attract more customers with lower item prices
5. go to 1
Now, the first virtuous cycle only works at the city level - operating in two cities doesn't improve your customer density. But for the second virtuous cycle you need overall market share, not just local market share. To get Wal-Mart market share (~30%) and prices you need Wal-Mart buying power.
Of course, that's no excuse for building loads of warehouses full of tech that can't achieve the performance you need and to 'iterate' costs a few hundred million per warehouse! But it's understandable why the business plan would call for national expansion at some time - and I can also understand why people immersed in the SV community would find it hard to slow down and not break things :)
I live in Good Eggs central - san francisco, and I foresee problems that are still ahead for this startup. Don't get me wrong: I really want Good Eggs to succeed, but they have quality control issues that need to be addressed and scale very poorly.
Specifically, I've been a customer of theirs for years, and in the past two years in particular, I've seen prices go up pretty dramatically and quality in some cases go down.
It used to be the case that Good Eggs produce and meats were farmers-market quality at really good prices, delivered. Now it's the case that the produce feels like farmers-market seconds (the produce I get from them regularly has only a few more days of freshness left and are decidedly not the fruits or veggies I would pick if I went to the store on my own), and the meat and fish prices have gone well beyond the cost of the same meat from other sources (say a trip to a regular market). Top this all off with persistent packaging and logistics errors - my last order contained someone else's frozen lunch pies, didn't contain any of the cheese I ordered, and was delivered an hour before the actual delivery window I requested (and contained fish, so if I hadn't already been home, this would have been a problem).
Good Eggs is really good at fixing these problems - I email their customer service and they credit me for stuff I don't have and throw in a $5 off coupon for next time or whatever. If inconsistent quality or mispacks were rare, this'd be perfect. The problem is, it is not rare at all. I have definitely fallen off in my use of their service - if I'm paying the same price, I might as well just go to Andronico's, Whole Foods, or the Ferry Building (or Gus's or Rainbow, or Berkeley Bowl ... there's a lot of great grocery stores in the SF region) so I can pick out what I want, and sometimes (especially at Berkeley Bowl) get a way better price.
They need a better process for vetting the product they get from distributors (and possibly a better contract when it comes to rejecting low-quality stuff) and a better process for getting people's stuff into their bags. That's not scaling, that's core competency. It may be the case that this doesn't scale, can they deliver food that's sufficiently high quality that I could say, 'If I went to the store, I'd have chosen that tomato' at a price that's competitive with the store? That's a difficult problem to solve and requires well-established relationships with vendors.
It takes a lot more than good code, talented engineers and passionate customers to scale food tech startups that deal with getting perishable food from local farms to people’s doorsteps.
It feels like some kind of tech-myopia that anyone should think code was the key to a successful food company. The old "when you're holding a hammer..."
Yeah, they could've hit Monad Transformer Hell, or they could've wasted all their time learning how to write practical applications in Haskell and end up bankrupt in a very lovely ivory tower.
I've heard this argument over and over again and it never is good advice. It ignores the context completely as well as whether the developers know Haskell or not.
You can screw things up in any language, especially if a large part of your business model involves stuff not related to code, like shipping perishable food.
I'm guessing the logistical and financial troubles described in the OP were pretty clear 2 months ago...was expanding service to Manhattan a Hail Mary to attract additional investment?
When it's clear at the top that core strategy needs to change (as in this case), this usually doesn't impact the company in bits and pieces — it all comes at once. That's because making a major change at a large company requires coherent planning, and you don't get coherent planning by only saying "we need to reduce opex across the board, we're running out of money". You get it by making a new top-down plan with entirely new goals and a new P&L.
As a result, 140 people got laid off all at once, rather than in waves. Had it been done in waves there would have been mass confusion, low morale, and an inability to buy into the new plan because there wouldn't have been a new plan yet.
And for the same exact reason plans to roll out to Manhattan, which were surely started many months ago and enshrined in yearly and quarterly goals, were axed abruptly.
It's possible it was a cynical ploy for fundraising, but I really doubt that. It's very hard to pull off momentum games at GoodEgg's size — investment scrutiny is much higher, and decisions are based more on spreadsheets. Not that hype doesn't factor into it, but it's not going to float you if your core economics are busted.
Part of their operation was driving a truck up to Fishkill, NY and filling it with goods from Hudson Valley producers, many of who are too small to be commercially distributed already. I bet someone could keep doing that and make good money. But probably not enough money to make VC's happy.
Facepalms all around. Supply chains are hard, and the idea that you can just write software to "solve" them is a level of hubris that defies description. Supply chains aren't a problem you can solve with big data. They aren't a problem that be solved with machine learning. In fact, you could combine the worlds top researchers in Computer Science, Economics, Operations Research, Mathematics, and Statistics, put them in a room together for a decade, and still not "solve" even 1% of the problems that Supply Chains face. I know the ITA presentation on Air Travel Planning Complexity[1] has been circulated here many times, but it always blows peoples minds how complex that problem space is, and that type of problem is miniscule in comparison to end to end supply chain optimization. I work in a problem space that is the combination of 9 different NP-Complete/Hard problem spaces (I've counted) on a regular basis...we are barely scraping by with millions of lines of code and entire data centers of our servers running at peak utilization 24/7, and that is with state of the art heuristics solvers and hundreds of PhD-level researchers, and that is only a subset of the Supply Chain problems that my company faces. To say that Good Eggs bit off more than they can chew is a MASSIVE understatement.
Supply Chain Management is a field that is heuristic-driven because there isn't a solution. The heuristics that drive modern food distribution are the result of a real world genetic algorithm that is already thousands of generations deep...and miniscule incremental improvements have been the subject of PhD Theses for over a hundred years now. You don't optimize on top of that by writing a web service. Supply Chain Management may not be technologically advanced in its usage of enterprise software, but that doesn't mean that their ideas and processes are stupid. At best, it means there are some annoying frictions in the industry in the way that they interact with other industries (and there are actually plenty of viable startup opportunities to fix this). The presumption of stupidity[2] regarding the existing state of logistics and supply chain management killed this company, just like how it killed Webvan, and is going to kill a whole host of startups in the distribution and delivery space within the next few years. [3]
[3] PS: If you are in VC and have a stake in any startup that is doing last mile delivery, liquidate now. There isn't a single startup in the space, whether they are currently valued at $1 or $1B, that can sustainably provide their existing value proposition that they are promising to their customers. It is possible to force scale some of them into something sustainable, but your valuations make me believe your cash flow neutrality projections are off by more than an order of magnitude.
Not to mention that food itself does not elicit rational behavior. If we were remotely logical about how we consume food, we'd all be voluntarily eating a nutrition paste out of a tube every morning.
Eating food is an animal need, and therefore we have instinctual animal behaviors around supplying it and consuming it that just can't fit into any sort of rational mental model. There are functions around food behavior built into our DNA and we haven't had enough generations since the invention of agriculture to evolve beyond them yet. Heck, I'd even guess we would all benefit if food supply chains were more decentralized and chaotic than they are already.
This irrationality is unfortunately a difficult thing for engineering-minded people to accept, so it makes sense that the valley would just keep on making these same mistakes with every generation. So it goes
Are you suggesting human dietary behaviors have not changed since the inception of agriculture thousands of years ago? Because they have, and that clearly demonstrates that how and what people eat is driven by complex cultural processes. But the only source of behavior you identify is in-born and determined by genetics.
Our genetics have been influenced by agriculture, already. I don't know what it means to 'move beyond' behavior. Humans are very good at changing their behavior at time scales far, far shorter than any evolutionary time scale. Even for behaviors that one would consider particularly basic, like sex or... eating behavior.
All I see is an assertion that some behavior is 'logical', actual human behavior is 'irrational', and the reason why humans are acting so irrational is because its ingrained in their DNA. The allegedly logical argument for eating Soylent is left unexplained. I doubt there is one. I doubt modern humans could yet design a single nutrition product that outcompetes a diverse diet in all benefits. I think you just use the word logical to designate things as good, and irrational to designate things as bad.
Instead of actually considering why people eat the way they do, you're not saying anything at all except that you think it's dumb. It's a very boring point to make and it looks flimsy and downright silly when your justification is Argumentum Ad DNA.
Boring, boring, boring, and unfortunately incredibly prolific on the Internet.
You seem to be incredibly smart and well educated and would do the world a great service if you worked to shift your energy towards compassionately sharing in knowledge exchange rather than venting anger at why others don't have the same knowledge you do.
I want to open this up again and give you free reign here to explain to me why my thoughts are incorrect, but with the condition that it's done without attacking the structure of my words or me personally.
I have two theories about neuroscience, but I have no formal education in science and my knowledge comes from the university of HN, wikipedia and google. I want to know why I'm wrong about these theories and become educated on the reasons why they are wrong. If you can help me, I would be very thankful.
The first theory is that grocery stores and home cooking are our last proxy for primitive food scavenging behavior (I believe this is called dopamine reward seeking). This is why Soylent will never be successful and why online food shopping will always be a niche business.
Second - I've never been sold on the idea that our brain ships without any software preinstalled. It's hard to debate in favor of this since the science is murky and, of course, I'm no scientist. The only thing I can point to is this http://www.telegraph.co.uk/news/science/science-news/1048647...
If you can help explain to me why I'm mistaken with these thoughts, you'd make a positive impact on my life and hopefully improve your own outlook as well.
Or you can tell me to fuck off. But know that I did give a shit about what you had to say and it's stuck with me for almost 24hrs now.
No worries. Honestly don't sweat responding to my comment, there are some very silly reasons why I'm interested in the topic in the first place.
I wish there was a better medium to read your work than HN comment history. If you don't already write professionally I hope you are pursuing it. You're very talented and thorough. I want to read more of what you have to say.
But a parting word of advice - that writing talent of yours makes your criticism feel like the intellectual equivalent of being punched by the incredible hulk. If you're pursuing ambitions in technology, anyone on HN could either be a valuable asset or a particularly shitty enemy. Best not to risk crushing the wrong person
It might be a while, I got even more busy with life than when I initially commented to you. Do you have an account at a place with direct messaging? I can send you my thoughts there.
The truth is, it's very easy to talk the way I did, and very hard to talk in a constructive way.
Thanks for the very interesting post. What sort of companies are providing the supply chain / logistics management/optimization consulting services that you mention (or is it mostly done in-house?)? Are any focusing on serving Asia?
I wonder if there's a good startup opportunity in this industry in Asia; many firms possibly haven't figured out the application of mathematics/ops-research to their supply chains yet, at least compared to the level of top US firms.
Logistics (delivery) startups are burgeoning in Asia, but no doubt will face these issues soon - maybe specialized mathematical expertise will help.
Sometimes "do things that don't scale" is bad advice. It's great advice when starting out (because you should be focused on building an appealing product before you worry about scaling it), but you need to be using your scale-up to test ways of breaking out of doing the "things that don't scale". The idea is to turn it into a forcing function for the company: if you can't find a better way to scale, you're done.
Disruptive innovation is about ripping out the assumptions that incumbents made and operating under a new set of assumptions that may not have been possible when the incumbents started. But sometimes, the incumbents' assumptions/constraints are still valid, and new entrants don't have as much of an advantage (or the new entrants' advantage is easily copied).
I think the sweet spot is going to be a Trader Joe's-like online brand. Good quality, organic options, low prices, curated selection. Instacart, GoodEggs and others are missing the mark.
1. A lot of people who care about organic and farm-to-table ethos already live in areas with significant market coverage. Those people will tend to have existing behaviors — GoodEggs sought to displace these. This can be hard if existing players do good-enough job.
2. The paradox of some delivery services is that the people who need them most are sometimes the most economically unviable customers (because they are geographically spread out, because they are lacking in funds or both).
I wonder if the problem has to do with providing a desirable product? It's already pretty easy to get good food, good organic food and good organic locally grown food from a variety of vendors. With GoodEggs it seems you're getting a clunky ordering experience, inability to select produce, an extra shipping charge, moderate/high prices, etc.
- obviously, code won't overcome the challenges of immature, lower-scale logistics and business operations. Roll up your sleeves like everyone else!
- no, it's not about "big data" as much as business process software.
- but yes, code can help in some ways:
- analytics/BI/decision-support: smarter decisions, fewer meetings.
- running ops: fewer managers, automated monitoring and project followup.
- use code to help manage your supply chain.
- but yes, "tech thinking" does help:
- faster, safer process change
- remote monitoring
- application of classic compsci algorithms vs reinventing wheels.
cheers!
adam
Buyer's Best Friend
leading provider of office snacks to silicon valley
http://bbf.io/snacks
Good luck to them. Hope they figure it out. Churn can really hurt your business and affect growth tremendously if you don't address it. I'm guessing that's the case here. It's like taking three steps forward and two step back. I'd rather take one step forward at a time.
$200M Wouldn't be enough. They are trying to solve a problem that doesn't exist for 95% of the country. Sure, there is a market for this, it's just not the size they are thinking.
As the article says, "Figure Out Your Model Before Scaling".
I really liked Good Eggs, and their customer support always seemed great. Unfortunately, my last order came an item short so they gave me a $5 credit on my next order. A day or two later, they close down in NY and now I won't be able to redeem the credit.
Not a big deal, but it was a bit annoying coming from a company I'd always used specifically for their great customer service.
It's a little ironic that goodeggs is considered a food technology firm, when it has nothing to do with technology, and in fact has a rejectionist approach to technology in the first place.
Monsanto and Cargill are food technology firms. Advances in food technology are what allow us to feed 9 billion people with a fixed resource (arable land). There are serious innovations in food production and distribution that have saved millions of lives in the last half century (Norman Borlaug and the Green Revolution).
An online seller of local produce is not "food tech". It is simply using the web to sell food. Add to that their dislike of synthetic pesticide/fertilizer and other real advances in food technology (advances that in the developing world mean a real reduction in deadly diseases that spread through the use of animal manure [natural fertilizer]), and it makes them just another anti-tech firm that happens to peddle goods on the Internet.
I'm not sure if you just have a problem with organic food, or if this is just another reiteration of the tired old "if you're not trying to solve world hunger..." argument that gets used to criticize business ideas, but none of your points are really relevant to a company whose goal is to conveniently supply food to some of the richest cities in the world.
I think folks underestimate how much agricultural science and technology there is that doesn't fall under the very narrow categories of genetic engineering and chemical pesticides.
Modern society is built on top of the massive productivity improvements in agriculture that occurred before those were introduced, and there is plenty of room to innovate without using those methods.
GMO crops are limited[1] in Europe, yet they continue to be a leader in agricultural innovation.
They are 'effectively' banned in Europe though. As you can see from this infographic[1] approval of a GMO requires a political vote in the European Commission. This takes place after a very expensive risk assessment, and you can't get that money back. It's a tough business case to spend a ton of money on the risk assessment and then be at the whim of politicians who are publicly against the technology and know they win votes by being against it. My company makes GMO plants and until the situation changes there's no way we'd considered trying to get approval.
I conjecture they setup this system because an outright ban, which would be unscientific, would violate international trade treaties, instead they just use this process.
There quite a number specifically approved, which is inconsistent with an "effective" ban.
> My company makes GMO plants and until the situation changes there's no way we'd considered trying to get approval.
Demonstrably -- that is, from the list of products that are approved -- that's not true of the major players in the industry (e.g., Monsanto, Bayer, Dow, Syngenta) and even some smaller players.
Your company not having the resources to compete in a given market doesn't mean that that market has banned GMOs.
Those 2013 articles are outdated, Monsanto had a number of permitted approved this year. They either never withdrew their pending applications as those articles stated they intended to, or they got new applications approved in less than two years start to finish.
Of course there are other aspects to agricultural science and technology. However, there is nothing in this company that indicates they are advancing technology in any sense.
This is not to say they are not providing a worthwhile service. Even my local farmer's market provides a worthwhile service to me, but I don't call it a "tech market" just because some of the vendors have an online presence.
I see where you're coming from, but I think there are still a lot of possibilities for innovation here.
We waste 50% of food, with a big portion of that waste happening at the retail level. Food distribution services like Good Eggs have a pretty good shot at eliminating that waste.
Just to give an idea of the potential, attacking only food waste, without any other improvments to our food system, would comfortably allow earth to have a population of 10 billion.
The world produces ~three times as much food as we need to feed the world. Sure, we can push efficiency up further, but there is no real need for more calories and current practices create huge short and long term problems.
Not that fuel ethanol is drinkable, but you can get similar amounts of energy with other processes. Anyway, just the US ethanol production is over 2 gallons of per person on the planet. We are talking ~23,000 calories per person on the planet produced in a horribly inefficient process which is then burned.
You have to know that organic vs. Monsanto is a hot button to push, so throwing that into your initial comment is a sure way to derail the conversation into arguments.
If you really just want to talk about the "tech" label, then you should just stick to that point.
It's a hot topic, yes, but an important one. We are in an age where policy is decided by a scientifically illiterate public, with potentially devastating consequences for people outside the developed world. That needs to be confronted head on, not shied away from.
One of the problems is this idea that opposition to GMOs and growth in demand for organic food is fueled by scientific illiteracy.
There an oft-repeated lament of "I can't believe people think GMOs are unsafe!", but why not instead start by asking why purchases of organic food increase alongside increased education level?
start by asking why purchases of organic food increase alongside increased education level?
Is that true even if you control for income levels? Because it's not really a major mystery why the poor uneducated classes don't buy the (typically) more expensive organic food.
(Yes, I know there are wealthy uneducated people, and poor but educated people, but wealth and education are known to be correlated)
"Advances in food technology are what allow us to feed 9 billion people"
Without a citation I'm disinclined to believe this. Populations have managed to feed themselves without Monsanto/Cargill-style advances for pretty much all of human history. Something like a third of food goes wasted as it is.
If that were true, I'd be proud to work there. Food innovation is of critical importance for mankind, in the same realm as space exploration or improved information accessibility.
Certainly a lot more important selling ads, or providing an internet-based service to pick up your laundry.
I am familiar with what Monsanto does and have pretty much agreed with GP. I am also familiar with the anti-science youtube documentary watching, "do-your-own-research" conspiracy junky crowd accusing strangers of being shills on public forums. Your profile is blank so that is all I have to go on as far as "what you do".
When I first heard of this company I thought it was some kind of joke. Turns out I should really learn how to bamboozle investors into handing me $53M for shitty ideas that don't make even basic logic sense.
No, it isn't. It is fairly straightforward and thousands of companies have scaled across the world.
Scaling a two-sided network focusing on providing a niche seasonal product to a niche audience is really, really hard. It wasn't the food distribution that wasn't scalable, it was the business model.
"We were motivated by enthusiasm for our mission and eagerness to bring Good Eggs to more people. But the best of intentions were not enough to overcome the complexity."
No, you were motivated by aggressive expansion targets to provide returns for the dozens of millions of dollars you took as investment and to make a lot of money.