I work in crisis management, sometimes on tense, hours-long outage calls. Very often, engineers or others on the call will say things like "It would be nice to run test $x before we move forward."
It basically always pays to ask a counter-question like "Will we want to change our action in some way based on the result of this test?"
Often the answer is no, so we can skip doing the test and get to remediation faster.
Once you notice this pattern (it's clearer in outage situations where moments matter), you frequently catch people (including yourself, if you're honest) seeking information that they don't even /plan/ to use as an input to some decision or action. In other words, there is no conceivable future where the answer to some proposed question would have an effect on their actions. If you find such a situation, you can at minimum remove answering that question from the critical path, and possibly just never bother finding the answer at all.
That doesn't mean that having such information is bad, just that we should think of the cost of gathering suit against its likelihood of mattering in terms of our actions. In fact, possibly one of the central purposes of IT broadly is to lower the cost of answering such questions, so that we can afford to ask more of them.
My favorite when making long term decisions is "Do we need to make the decision between these choices now, or can we proceed keeping all the choices open and make the decision later?"
Sometimes there's a belabored debate between two choices and answering this question can reveal that the choices don't have to branch right now, that they have (or could have) a shared 'path prefix', enabling us to make the choice later when we have more information.
The optimal time to make a decision is the moment before it hits the critical path. Where I work we make this actionable by saying to defer decisions and pull forward unknowns.
Sometimes there’s an underlying agenda to these conversations. The people involved are like lions trying to find out who’s the “alpha” - they’re less interested in the decision than establishing a hierarchy in the decision-making process.
Agreed. This question is a superpower. Not just 'the result of this test', this applies to all kinds of questions.
It's also relevant when people ask for ill-specified 'dashboards', or 'reports'.
Will what we do change substantially based on what we see on these dashboards? If not, why make them? And if so... can we just automate that behavior so we don't have to rely on someone noticing the dashboard?
I think some dashboards are meant for some form of situational awareness. At least "I can see when something is different" ideally even "I can recognize if everything is going well".
For those purposes it is difficult to answer 'what will we change' and even harder to automate the process.
Yes, not arguing dashboards are a bad idea in all cases.
If there are five things you always need to go and check when someone asks you ‘is the system running okay?’ Or an alert goes off, by all means put those things on a dashboard.
But when someone just says ‘we want to be able to see these numbers’ it’s often worth asking what they plan to do with them.
A related principle is that often you’re really not blocked on things you think you’re blocked on. Waiting on that part / component / email response? Pretend you just got it. Chances are the first thing, or even first few things, you’d do didn’t require the thing you were waiting for.
This happens in a surprisingly large number of contexts. People raise questions and invest time and energy looking for an answer without the slightest idea how or if the answer will inform future decisions. Seems to be a human fallacy.
The issue came when people decide that "emergency" and "normal operations" should be the same. In an emergency such corner cutting approach might allow for quicker resolutions since in most cases emergencies happen in a relatively charted/pre-computed territory, in ordinary condition impede a proper development resulting in much more emergencies, and more and more larger/tough.
Unfortunately too many fails to appreciate complexity so "if something work" it must work anytime, at any scale, for any scenarios and so on. Teaching complexity IME generally fail.
The important thing is to make sure that it doesn’t fail silently fail or with unknown consequences (e.g. data corruption). So make sure that whenever something doesn’t work, it fails loudly and doesn’t proceed as if everything was okay.
“It basically always pays to ask a counter-question like "Will we want to change our action in some way based on the result of this test?"”
I like to ask this too. This question is also very useful in a medical context. They often order tests where the outcome doesn’t really matter for the path forward.
I completely agree with the sentiment of this post for the purposes of solving business problems, and more generally for making rational decisions in one's life. I find it intriguing though, because I think a way of rephrasing your argument is that one should make sure that all of one's questions and approaches should be instrumental toward a goal. In every-day life, the only kind of person who truly takes this approach interpersonally is a sociopath. This kind of person only interacts with others as a means to their own ends, and generally the goal is efficiency and material gain.
It makes sense, because at some level the success of a corporation is totally devoid of purpose outside the maximization of profit. I just think it is interesting how that bleeds into the habits of employees. How much of that do we take home?
I think we always take some of it home. You can try all you like to keep your work and private lives separate but there will always be some spillover. Plus, no matter what you do, there is always in theory some dollar value attached even to your free time, or some limitation on the amount of time off you can take, which means that you really do have to be goal-oriented about taking a vacation, spending time with family and friends, practicing an instrument, etc. If you're not, it might not happen at all.
I don't know what the answer is but I agree that market-worship is bad and makes all of us involved in market-driven work (i.e. basically everyone) a little less compassionate.
> Finally, the third, and most pernicious situation in which there is too much planning is when planning becomes its own end. This can happen because individuals in organizations get more reward for planning work than actually executing on it. It might be seen that the execution is the “easy part”, and can be done by anyone, whereas the grand visionary (or “architect”) is the real cause of success.
I've experienced this happening due to an influx of people into an organization who simply can't execute -- middle managers come in with impressive resumes but little understanding of the problems that need to be solved. When results aren't delivered, they default to extensive planning processes because it creates an appearance of work. There's lots of tangible outputs (market studies, reams of wiki pages and documentation that are written, fancy slide decks, plenty of presentations...) yet nothing that provides actual value to customers. They know execution is the hard part, they just can't do it, so they stay in planning mode endlessly in order to provide an illusion of productivity, and to keep their job.
Over Planning is also an institutional response to risk. In these cases it is actually "the easy part" and so preferred by environments where any mistakes are costly. As in the "If at first you don't succeed, then Sky Diving is not for you" kind. In these cases I would not call planning an end in itself but rather a defense mechanism deployed to avoid the risk of action until absolutely necessary. Not that I justify what can become analysis-paralysis situations, but it is better to understand the beast you are dealing with if you want to navigate a path through.
To further elaborate on this point, sometimes that risk is internally created. If you get negative feedback for doing something everyone agrees is valuable, negative because it's not what your manager would have chosen as the highest priority, or because the architect would have planned the implementation slightly differently, etc., then the natural response to that negative feedback is "OK, next time, let's do some more planning so that you'll be happier with the work that's being done." What makes customers happy and what makes managers happy is not always the same thing.
Great organizations embrace risk. Not reckless risk, not the kind where you say "eh, this might result in two-day downtime for all our customers, YOLO", but the kind where you say "I prefer that my workers make imperfect decisions that we can fix later if we really need to so that the whole organization can move faster". 80% of the time, the decision does not need to be fixed. 80% of the remaining 20%, you don't have time to fix it, and usually, that actually works out OK, all things considered.
It's a military saying but I think it applies equally to business. I think that's the whole point of rubrics about how much you need to talk to customers.
When seconds count, the police are minutes away.
Said by pro gun people. I'm of mixed feelings about that but I still like the saying very much to encapsulate an idea about making hard decisions in critical situations when time is of the essence. It's similar to the military saying Sometimes, a 90 percent solution now is better than a 100 percent solution later.
> No plan survives contact with the enemy.
I've also heard a variation like "It is certain that the plan will fail, however, without the plan everything fails."
When Mike Tyson was asked by a reporter whether he was worried about Evander Holyfield and his fight plan he answered; “Everyone has a plan until they get punched in the mouth.”
The driving parable reminded me of my one and only time driving in the USA. The thing that absolutely did my head in was the intersections with four-way stop signs. There was no clear indication of who had right of way which I would have thought would cause a lot of collisions. It just seemed like the biggest bully goes first.
It's first come, first served, so you have to note who is already there when you pull up. If two cars arrive at the same time, you yield to the car on your right. Most of the time it works fairly well. If a car goes out of turn because they weren't paying attention, it's usually just a minor annoyance.
The rule for four-way stops is simple: Right-of-way is established in order of arrival at the intersection. First to arrive at the stop sign can go first, followed by the second, etc.
Not everybody follows the rule, and technically they're breaking traffic laws by doing so.
And you're right. I prefer roundabouts too, they're objectively safer. Massachusetts has a fair few roundabouts, but other parts of the country seem to abhor them.
There are other rules. Always yield to pedestrians. Two cars can sometimes go at the same time if they're not on a collision course. If pedestrians are blocking the person whose turn it is, you can go instead. Order is based on arrival time, unless there's a lot of cars then it's yield to the right if you're unsure.
These rules combine in such a way that it's genuinely unclear who should go in some situations - the above rules might contradict, or you might need information you don't have cause you got there after something broke "the order". But if you think it's your turn and start to go, people will let you, and ultimately no one gets stuck forever.
not in Illinois: here, after whoever arrived first goes, the right of way goes to whoever is to their right (or, next, going counter clockwise), regardless when they arrived (as long as they arrived before the first car went)
Most people dont know how to drive in roundabouts even in cities where they are plentiful. I like how it is most of Europe; always yield to the right in the absence of signs that say otherwise (priority sign or traffic lights)
I'll agree roundabouts are tricky if you are not use to them. And those giant ones in Europe look insane. When teaching my kids to drive it was one of the trickier skills they had to master - being able to read the traffic and know when they can go.
You'd be surprised how compact you can get a workable roundabout[0]. The brits have taken this to the extreme in some places,[1] Though I don't know effective a bit of paint is in actually making people treat it as a roundabout.
Actually I think traffic lights are better for very busy and/or very large intersections. But roundabouts are superior to four way stop sign intersections in every way. They don't even need to be all that large.
"None of these take into account the needs of the drivers, so should characterize them as unjust. For example, one driver might be late to catch their flight (costing them hundreds of dollars), while another might be on a leisurely cruise around town with no particular destination."
I don't know how to put this into words, but I hate that "doing nothing" is equated by being worth less, or less urgent, or having less rights. Your emergency is not my urgency, or something along those lines. If I choose to live slow leisurely, I shouldn't loose any rights over it.
It'd also always give right of way to people who postpone doing things until the last minute (so that they're always in a hurry ... Hmm I feel a bit guilty).
Which isn't the best incentives to build society on
My organization has this disease bad. It’s awful. It’s so ingrained that our official leveling competencies stop recognizing differences in execution skills (coding, debugging, etc) after 2-4 years into a career. All growth from there on out is planning, design, vision, influence, leadership.
This design might have made sense in a world with voracious hiring of junior talent to do the execution. But hiring has slowed and had shifted towards more senior roles even before it slowed. So we are awash in grand plans and perpetually short of resources to execute them. Those who do get stuck executing plans are of lower than average skill, since the competent implementers are promoted to planning. So even if you can get resources for your grand plan, chances are it will be executed badly.
The author thinks the parable they heard concerning silly traffic priority negotiation would be too horribly impractical to occur in real life; but they clearly have never been on a winding (mostly) single lane coastal road in Southern Italy, e.g. Amalfi. Silly and impractical negotiation between drivers is exactly how it works, or rather, doesn't. Although in this case it manifests the complete absence of planning rather than too much (the point of the article), because nobody sane would plan to allow tourist busses to travel bi-directionally on a geographically-constrained winding road that barely enables small cars to pass two abrest in many places.
I understood the parable differently than you. The Amalfi coast is not the same as a small town main street in the United States. The crux of the parable is that planning and negotiation is useful when it its cost is outweighed by the benefit and not useful if it's more costly than the gains.
I’ve developed a way of working, where I start with a “rough napkin sketch,” and begin writing, quite early[0]. I also try to get a high-quality (but incomplete) working prototype available as quickly as possible.
A very important part of my process is the “don’t try this at home, kids” part. It requires a great deal of architectural and implementation experience. Lots of scars and a pronounced limp.
If I describe this to folks, they tend to freak out, but it works on my machine…
What is the name for the fallacy in which you make an analogy that does not reflect the real-world territory at all, and then draw a conclusion from it?
“Doing” is not at all enough. (Well, it shouldn’t be.)
Peer into any random engineering org and you’ll inevitably find loads of engineers driving around in circles with the wind in their hair, smiling with a sense of how “fast” and productively they’re moving!
It is the most common way of trying to cope with novelty: by means of metaphors
and analogies we try to link the new to the old, the novel to the familiar.
Under sufficiently slow and gradual change, it works reasonably well; in the case
of a sharp discontinuity, however, the method breaks down: though we may glorify
it with the name "common sense", our past experience is no longer relevant, the
analogies become too shallow, and the metaphors become more misleading than illuminating.
This is the situation that is characteristic for the "radical" novelty.
On the cruelty of really teaching computing science Edsger Dijkstra
This issue is discussed by Donald Knuth in one of his interviews [1].
"I know that every large project has some things that are much less fun than others; so I can get through the tedium, the sweeping or whatever else needs to be done. I just do it and get it over with, instead of wasting time figuring out how not to do it. I learned that from my parents. My mother is amazing to watch because she doesn't do anything efficiently, really: She puts about three times as much energy as necessary into everything she does. But she never spends any time wondering what to do next or how to optimize anything; she just keeps working. Her strategy, slightly simplified, is, "See something that needs to be done and do it." All day long. And at the end of the day, she's accomplished a huge amount." - Donald Knuth
The author is suggesting replacing a bad plan (everyone discuss each intersection) with another bad plan (just drive however you like).
(And as an aside, stopping at each intersection to discuss is not ideal, but at least no one is going to die.)
Anyway, I’m unconvinced. Most of the planning failures I’ve seen came about because people were bad at planning, not because they were talking too much. I’ve seen the opposite (not enough talking, not enough looking ahead) far more often.
I do like the slogan though. Planning is for doing.
>The author is suggesting replacing a bad plan (everyone discuss each intersection) with another bad plan (just drive however you like).
The author is suggesting replacing it with the normal rules of the road, which are as close to not planning as possible (following basic heuristics like go on green). The example in the parable didn't have "unjust" stop signs or traffic lights for a reason.
Actually, I would say that coming up with the rules constitutes planning, whereas negotiating who drives first each time constitutes a comparative lack of planning (and of rules). To me the parable rather demonstrates a case where planning (to decide actual rules) would be worth the extra time, because it saves time each time the rules are applied. The parable thus didn’t work for me, although I largely agree with the conclusion.
>The common way of determining right of way on roads is some combination of right of way, stop signs, and traffic lights. None of these take into account the needs of the drivers, so should characterize them as unjust.
In modern societies having equal rights is 'unjust'.
It often can be. But a slightly wider point of the article is that figuring out a more just process is so harmful that 'equality' turns out to be the least unjust.
And that counts even if you assume everyone is honestly participating rather than selfishly manipulating the system. Once you take that into account the least unjust approach quickly becomes equality.
The William James quotation at the end should not be presented as a single one. An ellipses should only be used if they were from the same paragraph. Actually, the two parts are from distinct passages several sections appart.
good article.
my questions now is how can we avoid over-plannings and under-plannings?
a lot of time I feel like the diminishing return of a planning is pretty steep for many situations but it's hard to tell how much time we should spend for the planning beforehand. (this is a planning for planning and maybe this itself is over-planning lol).
It basically always pays to ask a counter-question like "Will we want to change our action in some way based on the result of this test?"
Often the answer is no, so we can skip doing the test and get to remediation faster.
Once you notice this pattern (it's clearer in outage situations where moments matter), you frequently catch people (including yourself, if you're honest) seeking information that they don't even /plan/ to use as an input to some decision or action. In other words, there is no conceivable future where the answer to some proposed question would have an effect on their actions. If you find such a situation, you can at minimum remove answering that question from the critical path, and possibly just never bother finding the answer at all.
That doesn't mean that having such information is bad, just that we should think of the cost of gathering suit against its likelihood of mattering in terms of our actions. In fact, possibly one of the central purposes of IT broadly is to lower the cost of answering such questions, so that we can afford to ask more of them.