Hacker News new | past | comments | ask | show | jobs | submit login
Jeff Bezos explains the perfect way to make risky business decisions (businessinsider.com)
283 points by bookbild on April 12, 2017 | hide | past | favorite | 108 comments



The lede really stuck out to me:

"Most decisions should probably be made with somewhere around 70% of the information you wish you had. If you wait for 90%, in most cases, you’re probably being slow"

At what point is the information you have available for a key decision enough? You always want more.

This exactly mirrors a MAJOR decision-making framework taught in the Marine Corps--we actually call it the "70% solution."

All of our warfighting doctrine is built around the principle of rapid and decisive action, but the problem is you're in an uncertain, rapidly changing world where there are opposing actors operating around you and you never have all the answers. They emphasize in officer training that a key limiter is being paralyzed with uncertainty and wanting to wait until more information is available, and that you must actively counter this feeling. Indecision is a decision (and often the worst one).

I have personally found this one of the most difficult and perpetual factors I've encountered in both military and civilian careers. It's interesting that Bezos espouses this principle.


Nice comparison. I'm also familiar with "70% solution". One thing to add: officers, and hopefully business leaders, are taught how to build a theoretical list of what approximately 100% of the information, questions and answers might be...That way you know what 70% is in reference to. This is taught and one can easily memorize the checklists of decision making questions (it's a mnemonic device).

To see why this matters imagine: Are we looking for 70% out of 5 questions? or 70% of 10 questions we have? If you think you only need to answer 5 questions before making a decision and answer 4. That's 80% and you would go...But if there are 10 questions to ideally check off before deciding and you only answer 4, that's way too low and dangerous. This is dangerous in the military and in startup land. Knowing the full list of questions to ask is more important than knowing the answers.

Lastly, both Amazon and the military have huge structures, plans and support teams in place to encourage 70% as an OK threshold. Imagine if you have air support you can call in the marines should things go wrong, or you have your parents couch to crash on if your startup fails, those things would help a person be comfortable going on only 70% of the information.


> Imagine if you have air support you can call in the marines should things go wrong, or you have your parents couch to crash on if your startup fails, those things would help a person be comfortable going on only 70% of the information.

I would extend this to say that 50% of the knowledge needed for making a decision is understanding the downsides / risks of moving forward. For example, if you know you'll be OK if your fail, it is much easier to say yes to a big choice. However, if you're risking homelessnes, then it's probably better to find a different option.

Even in less extreme scenarios, minor decisions I make are often based on understanding the risks, more than understanding the upsides. Maybe it is too conservative of an approach, but it is (unfortunately) based on experience.


Mind sharing the list/mnemonic?


>But if there are 10 questions to ideally check off before deciding and you only answer 4, that's way too low and dangerous.

Unless the time required to answer those 6 questions is enough time for a competitor to take your place...?

Which is the better move? Do you gamble by waiting and hope they guess wrong, thus giving you more information that might help answer your 6 questions? Or do you roll the dice and make the best decision you can given everything you know?


I like to borrow from the decision theory field. There are massive frameworks around the "value of information" or "value of clairvoyance"[1].

The best summation of their framework is "Make a decision when the probability of new information changing your mind is minimal."

Of course, we need to quantify "new information" and "minimal", but it's a good framework to consider what new data points would change your existing course.

[1] http://lesswrong.com/lw/85x/value_of_information_four_exampl...


My favorite must-memorize Patton quote that sums this all up is:

"A good plan, violently executed now, is better than a perfect plan next week."


It also helps if you've got superior firepower and human resources compared to your adversaries, because otherwise Operation Barbarossa was quite the violent plan executed at just the right time (there are many reports saying that the Soviets were also preparing to attack first) but which failed miserably in the end. Nobody quotes generals Model or Guderian, who were just as competent if not more compared to Patton. Survivor bias at its best.


I agree with the survivor bias imbued with this sentiment, but I don't know if Operation Barbarossa is a good parallel to draw in this case. The Wehrmacht was targeting Moscow, the capital, then Hitler changed his mind, and overruled the generals. Combine with his later orders to never give up Stalingrad, I doubt his case is applicable for the poitns 3) and 4) of the article.


I thought about Hitler not pushing hard enough for Moscow after I had submitted my comment, and then I remembered that taking Moscow had been in vain for Napoleon (basically the guy who invented modern warfare), so not sure if that would have helped the Germans that much.

I'm a big fan of Clausewitz and of strategy discussions in general, but I think that we tend to focus too much on providential men who, naturally, can only take providential decisions (like Bezos, or Napoleon), and we tend to ignore the "longue durée", as the French historians called it (https://en.wikipedia.org/wiki/Longue_dur%C3%A9e), the long course of history which should have informed Hitler that his country did not have the necessary resources (logistical, military etc) to defeat both the Soviets and the English (and the Americans after the Pearl Harbour attack).

It may sound deterministic, but IMHO most of the battles (either in war, finance, in politics) have been won before the two sides started engaging in conflict. Some of those involved just didn't know it yet.


I would argue that Hitler's capture of Moscow would have yielded result scores different in both strategical significance and political, than Napoleon's, and it would have handed him a great advantage in 1942.

The strategic importance of the city was hundred-fold different. The Napoleonic wars was a time where the modern states first started engaging in national fights with significant casualty, but still continue the war effort. In this sense it was quite the first "modern" warfare, but the telecommunication technology was still in its infancy; scout messengers were still used and reconnaissance was still relatively rudimentary. Gathered information was a fraction of what the Wehrmacht could gather, and it meant the control center relied on smaller Staff office. Combine it with Napoleon's providential talent, you could essentially have a mobile command center and the army could still function. In 1940s the sheer amount of intel and volume of telecommunication wouldn't be handled by no other place than central capital. Imagine if Napoleon couldn't dispatch orders in time and get the intels that he needed. Moscow needed to be taken specifically because taking it would have damaged Russian effort more than it did in Napoleonic times.

Note that in Napoleonic invasion Russian army set Moscow ablaze and left them to an empty city; in the Barbarossa they were willing to defend it to last man. Even Stalin refused to move away. Production prowess of modern city Moscow wasn't also nothing to scoff at, too.

Von Clausewitz did write that war is a political means(I remember this from first few chapters from On War, though I forgot which translation I read) and one of the primary goals must be to 1) preserve one's own strength, while 2) deteriorate the opponent's fighting capability. Taking Moscow would have done exactly that. It also would have yielded a political gain, as Soviet army at the time was heavily dependent on a few personnel, especially after the Great Purge. Once you take Moscow and demean Stalin's infallibility, you also damage Soviet morale.

I too believe that most of the battles are determined before the engagements. We know of the contrarian ones because they are rare. German generals(Model, Manstein, Rundstedt, Guderian to name a few) were perhaps the most competent ones in the second world war, but they hardly get mentions. (Perhaps deservingly for the ends they were aiming at)


> They emphasize in officer training that a key limiter is being paralyzed with uncertainty and wanting to wait until more information is available, and that you must actively counter this feeling. Indecision is a decision (and often the worst one)

Analysis paralysis happens frequently with analytic types of people.



Based on the results of the last 15 years of foreign policy it seems like we're working from far less than 70% of the information. :(


Or it's possible your (persumably you're American) foreign policy is indeed working with full information - the goals are different from what you think it is. After all weapons manufacturing capabilities expensively started up cannot simply be closed down once your own needs are met except in case of uniquely advantageous stuff like f22 or b2


Or you just don't notice / ignore / know about all the things that went right / according to plan.


How do you even know what % of information you have?


I really like that this is an example from the real world. In the real world, people have skin in the game and their decisions matter. In the military field, of the boots on the ground variety, this distinction is crystal clear. People can die from the wrong decision. What happens if your Decision Framework Theory is wrong? Most likely nothing.

If Bezos didn't develop this heuristic himself - through his own risk taking - I wouldn't be surprised if he got it from something like the Marine Corps.


This is very valuable. My overall impression is that many people tend to be maximally risk-averse, regardless of scope. For example, rejecting a course of action due to fixating on a single unlikely thing that could go wrong. Rather than holding out for a choice that comes with no anxiety, I believe it's better to work on not being anxious.


If you structure things so you can change or revert your decisions later, it is also much cheaper to start executing them with not that much information.

Not always possible, of course, but once you start thinking in those terms, you can get good mileage out of that habit.



> Indecision is a decision (and often the worst one).

This is a great guiding philosophy!



I was thinking of the same thing! 70% is "too slow" as well.


It's a bit like speed of light, you can't reach it.


>> This exactly mirrors a MAJOR decision-making framework taught in the Marine Corps--we actually call it the "70% solution."

This is ironic.

Taken as an entire group, people who go to fight for their respective countries are routinely lied to, rarely have the full picture of the motives of those who are sending them out to fight, and end up causing destruction which is completely disproportional to the amount of knowledge they possess.

As the saying goes, the first casualty of every war is the truth. Besides, is the 21st century war, where one side drops bombs on a technologically hapless group of people via UAVs, actually a level playing field from which you want to learn business principles?


> This is ironic.

Well, to be fair, there are all sorts of sub-level scenarios within the larger political decision to go to war that call for independent decision making.

Officers cannot be blamed for political decisions


Given how much of success is luck, I try and play a game whenever I see successful people asked "how to be successful" where I try and 'spot the overfitting.' Is this really what made you successful? Or is this just what you wanted to do, and because you were successful, you think it has to be why you succeeded because of course your success was earned?


When the odds of success are favorable if you keep playing, eventually you will win, even if each individual event is luck.

For example, is it luck that the Vegas casinos make money on gambling? Nope, even though the outcome of each individual hand is luck.

Many people who are successful went through repeated failures to get there. They never gave up.


I find that a very weak analogy. The house wins in Vegas over the long haul because the odds have been tilted in their favor. Any one hand of blackjack is luck, but the house winning over the course of a million is a mathematically certainty. It has nothing to do with 'not giving up.'

What are the odds of becoming a billionaire? And how does someone tilt those odds in their favor?

Perhaps more to the point, Amazon was the first company Bezos founded. MicroSoft was Gates' first, and Michael Dell founded Dell while still at school. In what sense did these men go through repeated failures?

That's not to say these aren't exceptional men. Of course they are. But luck played a huge role in their success.


It isn't clear just what Bezos did before Amazon, but he engaged in a number of different occupations.

Traf-O-Data was Gates' first company. There's a reason nobody has heard of it :-)

Ray Kroc failed for 30 years before founding McDonalds. Colonel Sanders tried many things, and did not succeed until he was 62. T. Edison failed constantly.

Most musicians wander the desert for years before becoming "overnight" successes.

Mark Cuban had many attempts at business, including being a disco dance instructor (!).


Can we say survivorship bias? There are millions of hardworking and talented individuals who try repeatedly and end up with at most moderate success.

The number of breakouts is so small that there isn't any statistical significance to what those individuals did differently. Mostly, they happened to execute well - but nailed the timing perfectly. Bill's success with MS-DOS (on the eve of the PC revolution), the rise of car culture and suburbia for McDonald's. And Cuban got really lucky at the height of the bubble.

I'm not saying they weren't smart or talented. But to perfectly nail timing requires the ability to predict the future. And if you read "The Road Ahead, 1st ed.", it's obvious even the talented BillG isn't much better at this than anyone else.

Which leaves us with luck.


If "moderate success" is achievable without extreme fortune, then that's a perfectly acceptable target to aim for. I agree that trying to become a billionaire is a fool's errand but if you're trying to become a millionaire, you'll end up more likely to be a billionaire.

I've heard Mark Cuban say numerous times how incredibly lucky he was with Broadcast.com. But he's unapologetic in his belief that he'd be successful. He knew he'd become a millionaire and the work it took to get there, coupled with the right product and the right time is what made him a billionaire.

I have to agree with that. I'm sure Gates, Zuckerberg, Bezos, etc. would all be successful if Microsoft, Facebook, and Amazon didn't become incredible successes. They just wouldn't have been as successful.


This reminds me of something coaches often said. "Practice doesn't make perfect, only perfect practice makes perfect."

Sure luck has something to do with it, but if you're learning from your mistakes & constantly improving, you're chances of success get better all the time. It's just like throwing a ball at a hoop or any other thing in life. After a while you quit making as many mistakes that lessen your chance of success.


Timing is is easier if you try all the time


Exactly. If you wait for the wind to blow before building a sailboat, you'll never catch that "lucky wind".


The people who quit before they start fail 100% of the time.


>95% of people that use this quote end up with a lot less time and money than those who quit.


"Quitters never win, winners never quit, but those who never win AND never quit are idiots"

-I have yet to find the origin of this.


You only get perfect timing if you see things coming miles ahead. That's probably what they did better than anyone else.


> Most musicians wander the desert for years before becoming "overnight" successes.

No, most musicians wander the desert and never end up leaving the desert.


This explanation of survivor bias is about the best and most accessible I've found.

http://www.smbc-comics.com/comic/path-of-a-hero


Ray Kroc didn't found McDonalds, Richard and Maurice _McDonald_ did.

Also survivor bias but also you obviously will never accomplish anything without trying.

I think it is entertaining to try to find the abstraction that leads to success but it's probably more harmful than helpful to believe in something like that. Just do the best you can with what you have. Maybe it leads to success, maybe it doesn't but don't let that affect your perception of it, you did your best and that's all anyone can expect of you.


> Richard and Maurice _McDonald_ did

Technically you are correct, but that business was going nowhere under their management. It was Ray Kroc that made McDonald's in the practical sense of the word.


> And how does someone tilt those odds in their favor?

Study how to run a business and study the market one wishes to enter. Make a plan. Work every day to move forward with that plan.

Right there is already way ahead of most other people.

What lowers the odds to pretty much zero is doing nothing because you believe it's all luck so why bother.


> What lowers the odds to pretty much zero is doing nothing because you believe it's all luck so why bother.

You keep saying things like this, but nobody else is arguing that.


> What are the odds of becoming a billionaire? And how does someone tilt those odds in their favor?

Becoming a billionaire will rarely be the result of one lucky decision, but rather the result of thousands of individual decision where a larger than average number of them turn out to be "lucky".

I would think that the way to tilt the odds in your favor would be to study and listen to those who have successfully done this, and see if you can get some hints to how they made all these decisions.


But you can't just study what the successful people have done, you have to study what actually makes them different from everyone else. Think of it as a Type I error:

https://en.wikipedia.org/wiki/Type_I_and_type_II_errors

Let's say you interview 10 billionaires and 9 say they did X. Does X increase the likelihood of being a billionaire? You don't know until you look at the entire population of people TRYING to be billionaires. If X occurs in 90% of the billionaire population but 99% of the non-billionaire strivers population, then what you conclude is that in fact NOT doing X makes you more likely to succeed.* You can't study what makes you successful without including people who aren't successful.

* Yes I know that the difference between 90% and 99% is hard if not impossible to measure in a sample of 10.


> what actually makes them different from everyone else

They don't suffer from "I cannot succeed so why bother trying."


So do millions of other ambitious people. Many of us do achieve moderate success.

Walter - do you think there's a vast gap in ability between you and Zuckerberg? Or better yet, between you and Dustin Moskovitz? (He's worth 10.4 billion).

Or did they just also happen to be in the right place at the right time?


In the 70s I was with a company that designed SBCs. They had everything needed to be an Apple - talent, money, expertise. Everything but a Jobs who saw the opportunity.

Hence I'm not a billionaire.

In the years since, over and over and over, hindsight tells me I was in the right place at the right time, but was either too blind to see the opportunity or too chicken to take the risk. It's a running joke in my family.

Ironically, the ventures I have been successful at were all pooh-poohed by everyone when I started them.

If you can get up out of bed in the morning, have a brain that works, and live in a free country - you have all the luck you need to be successful.


> If you can get up out of bed in the morning, have a brain that works, and live in a free country - you have all the luck you need to be successful.

In general I agree with this, though of course there are exceptions. One of my brothers had a mental disability that required him to have 24/7 care. His main caretaker was my mother, and as a sibling I helped where I could. This was bad "luck" not only for my brother, but also my mother. She was always ambitious but my brother's disability needs took over most of her time and energy, leaving her with very little to pursue any financial/career success. Anyway, that's just one example.

I think there are many clear paths in the U.S. to becoming a millionaire. Most U.S. citizens can pick up a career guide and learn about high-paying careers and what it takes to get into it. I did this with becoming a software engineer. Then it was just a matter of following the well-beaten path. Go to school -> get computer science degree -> take up one of the many opportunities -> properly manage finances and save/invest.

Of course, this all requires some degree of luck. I was born a white male in America. This alone put me off to a great advantage over someone born in a poor corrupt country. I haven't had any known health or genetic issues, which is another fortunate circumstance.

Like you said, as long as you have a reasonable/average amount of health, then there are a lot of opportunities. As individuals we just need the courage and be willing to put in the work to make it happen. But let's never get too full of ourselves, because there is luck involved with any success.


Maybe we're in agreement. You fit my definition perfectly - success for sure (even your own wiki page), but never that massive breakout. And hindsight isn't great for predicting those once-in-a-lifetime convergences. Maybe Steve Jobs himself could have led your 70's era company to be Apple instead. Or maybe the happy set of coincidences would have played out differently and it would have been a moderate player but not the most valuable company in the world.

I too consider myself moderately successful (without quite your track record). I've also been close to those who do experience that next level of success. They are uniformly talented but their huge successes were once-in-a-lifetime. Some of them even tried to replicate and they achieved only moderate success the second time - what you'd expect of someone of their talent - despite their obvious advantages from having done it once before.


That's a bit of a strawman here, no? I'm saying "not all things that successful people do are the reasons they succeeded and successful people tend to suffer from hindsight and survivor bias so it's tricky to tell what led to their success." I said nothing about it being impossible or futile to try to succeed.


I share your view on luck. It's possible to position oneself such that a profit is made on a lucky event. One such event occurring is luck, but the positioning -- in order to profit if the event happens -- has nothing to do with luck and everything to do with skill.


if you interview lottery winners, you're not going to get very good advice.


No, but if you interview casino owners (people who make decisions that means they regularly get a lucky outcome, even if they don't win all the time), you will probably get pretty good advice.


casino owners compete with each other. Lots of casinos fail, but you don't get interviews with failed casino owners. (edit: I just realized that there is one, very notable exception to this)

If the only difference between a successful casino strategy and an unsuccessful one is failure to control for some variable, then the variance in that variable will yield some very successful casinos, and some unsuccessful casinos, regardless of the actual expected value of the casino business.


I'VE GOT THE BEST CASINOS - THEY'RE TREMENDOUS.

That kind of notable exception?


Moderate success is usually earned: hard work + talent. Breakout success typically requires a lot of luck on top of everything else. The difference between a successful career and that billion dollar exit mostly comes down to timing.


but has anyone ever actually work out the stats for this ? Do we really know the (hardwork+IQ+talent)/luck ratio for the top 20 most successful business leaders ? Who's examined the other universes where Bezos made all the same decisions but had bad luck ?

It seems like this whole area reeks of astrology.


Meh. Positive exposure is a choice. Knowing how to create that exposure, is a choice. If there is an application of effort at an opportune time, such that one is successful, that still necessitates the effort.

Admittedly in the startup world and the early days of computing this was probably more true, that factors outside of your control influenced your success. It happened to be the case that computers were something people wanted to invest in, even though some random hacker in his garage was just doing it for fun. The volatility plus the hype exaggerated any effort that was put into the game in the first place, thus making the coupling between effort and success seem discontinuous. But a human still had to motivate themselves to build the thing.

"How did my actions create positive exposure and how did I capitalize on things that appeared to me?" would be the operative questions here, not "what was the sequence of actions I controlled the world with?"


I'm not sure what to do with your insight. Should we not listen to successful people's advice? Should we listen, but be more skeptical?


I would say, yes -- "do not, or heavily discount, all successful people's advice" -- is advice that is life +EV.

Most successful people are not statisticians, nor drawing their conclusion from a sample larger than, say, themselves and perhaps the few successful friends closest to them, when giving the sort of advice in the OP. Moreover, given how easily humans tend to conflate "What did successful people do?" with "What did successful people compared to the set of people who tried to be successful?", the risk of bad advice is even larger. At best, you're hearing white noise, but more than likely, you're hearing a case of survivorship bias (even worse than white noise).


So, there's signal and then there's noise, right? Successful people do know things about success. But they have various biases -- hindsight bias, selection bias, etc. -- that add noise to what they know. So yeah, "listen, but be more skeptical" I think is the right approach. And one thing I think (but I could be wrong!) is a thing to listen for is successful people who acknowledge their good fortune. It speaks to a self-awareness that suggests they may understand their success more than someone who isn't aware of their biases.


How much of success is luck? I honestly don't know. It feels like another feel-good mantra coming out of SV culture. Most successful people I know or have heard of do not fit the profile of an average Joe (including qualities that are pre-set at birth and cannot change with success).


I don't think you can write Bezos off via survivor bias. Amazon has a track record of success in different industries; it's reasonable to assume that the man in charge isn't just randomly lucky.


Love the "disagree and commit" part. Perfectly encapsulates the situation where "I'm convinced this is a bad idea, but I'm going to trust you and support you."


As I worked in amazon previously and know this first hand, you can translate it to: "Shut up, and do as be told".

It is a very top down hierarchal type of company, and engineers usually have little influence in general, and often are kept in dark on key decisions. Disagreeing in general is seen as a nuisance. A lot of it would disappear if managers gave concrete details on why a decision was made, but it usually takes a skilled manager to do that, and Amazon, as a large company, has a combination of good managers and plenty of not-so good ones.


I think you worked in a crappy team, because I was there for nearly a decade, and I didn't find that to be the case.


You were on a bad team, sorry!


he [Bezos] notes, when you disagree and commit, it's not about holding an "I-told-you-so" over other people's heads. It's a chance for people to hear an opposing point of view but to move ahead with action

I actually can't see it as anything other than a (potential) "I-told-you-so." Why are you specifically highlighting that you disagree, if you are then giving the go-ahead? You are the boss. If you approve something, ultimately it's your responsibility.

If you really trusted the person you would give them full authority on the decision without dropping your own agreement or disagreement into the mix.


No, it works both ways. Ive personally been in the room where my VP and ~2 others favor product/feature A. The other ~6 people favor product/feature B. Everyone reports in to the VPs org. A round of discussion and it ends with "Ok, I dont understand/agree with points B.1 B.2 & B.3 but the room is clear. Lets move ahead with B, and make sure to follow up on answering those points early."

Much more commonly it's applied in the context of group decision making. Imagine Alice presents design Alpha for implementing a sprint story. Bob asks some pertinent questions, outlines an alternative, Bravo, and explains his reasoning. A round of discussion with Carol and Dan commences. Whether the group decides on Alph or Bravo doesnt really matter. Everyone has had the opportunity to ask, and answer, relevant questions. Everyone (hopefully) has greater context and understanding of the context, risks, and intended outcome. Everyone, including both Alice and Bob, is expected to stick with the decision and move forward with that single implementation.

When IC's play "I told you so" theyre poisoning team morale and disincentivizing others from speaking up. When managers resort to "just do it" they're fundamentally failing to communicate context the team lacks, or failing to understand the knowledge they're missing. Neither is an acceptable pattern.

PS: six years at amazon as an IC


I've been in a situation where the manager two levels above me vehemently disagreed. He allowed me to continue anyway.

It was never a situation where he wanted to hold it over me, rather it was something that should give you pause and cause you to consider if you are really correct. It also gave me a lot of respect for the manager - that he would trust me even though he was convinced it wouldn't work.

In the end it was a success - but had it failed I don't think it would have been held over me. If you don't ever experiment you'll die as an organization.


I agree 100% with this. The 'disagree' part should at least come with some form of strings attached. Something like 'we need to be able to pull the plug early with this', 'you start with a limited budget', or similar. Otherwise it's useless to mention the fact that you disagree.


You can trust that someone will make the decision and still think that they are doing it wrong. Voicing said disagreement doesn't mean that you deflect responsibility, it just means that you think that they could use the feedback.

Like this:

A: Hey B, could you look into problem P since you are the expert on this?

B: Sure, I think that we should do X.

A: Are you sure, I would go with Y.

B: I have considered Y and I think it is worse.

A: Then go ahead, you know this better than me.

Alternatively:

A: Hey B, could you look into problem P since you are the expert on this?

B: Sure, I think that we should do X.

A: Are you sure, I would go with Y.

B: I hadn't considered Y, thanks for bringing it up!

A: No problem, glad I could help.


The example he gave was about consensus, and when you want to stop fighting. No one else can actually force you to stop fighting and move on to commit.

You don't have to win or draw out every argument, and you don't have to be convinced of everything. Sometimes you just put a pin in it and go.


How does the leader give feedback in your model? Can the leader never chime in with some advice that might be helpful to the situation?


I read about all the crap Amazon employees have to deal with. So it's hard to view anything this guy does in a good light. IMO, anyway. I wonder how much this philosophy contributes to the horror stories we hear out here.


The horror stories don't resemble anything that I experienced during my five years at Amazon. I think they're largely propagated by NYT and friends because Amazon out-competed them in their own space.

Amazon is the best run company I ever worked for. By far.


Amazon out-competed the New York Times? Shouldn't they be more upset at Google or something?


Bezos does own the Washington Post


Any particular litmus-test items to distinguish good Amazon teams from bad ones?


Ask them for the ticket count metrics - number of tickets, severity and age of open ones. Overly ambitious managers tend to build unstable systems that will break under load.

I am going through my second stint at Amazon and I enjoy working there. However, it is not the best place for everyone.


That's one (important) aspect. When looking at a team or individual I check wiki contributions, design documents, commits, code review feedback provided to others, ticket queue depth/volume, "cm" volume, customer forums, broadcast sessions, and org role/level/tenure blend. Also for teams get a referral or just send a "cold" email to have coffee with an IC on the team. I'd hazard that most of those data sources work for most of the teams at Amazon.


This is the correct answer. My advice for amazon SDEs is: either your manager puts operational excellence as their top priority or you need to find a new manager.


Just to elaborate On the priority bit. Page count during on-call rotations is probably a huge metric to watch.


I would ask about how long people have worked on that team, how many have come on and left in the last year, and how many headcount they are getting in OP1.

This will tell you a lot about whether this team is growing and successful, stagnant, or on fire.

Getting paged a lot is a hallmark of teams you don't want to be on, but in the last couple of years, Amazon has instituted policies that make it easier to internally transfer (no strict time limit on how long you have to stay with a team).

Getting paged is sort of a proxy for the real thing you want to know: how many people are voting with their feet.


Check the policy again. As of 2017-04-01 (for reals) there are effectively no restrictions on internal transfers. Including "performance" restrictions.


Find out how many senior engineers are on that team and are participating in the oncall rotation that you will be a part of. Or, find out how many SDE IIs with 3+ years are on that team. If you don't have either, then you should be concerned.

Ask how long the various managers have been in that org/under the same senior manager/director. A manager who you like that gets along with senior leadership is super important.

Ask questions during your interview. If you interview for an org, you'll be working with most of those people. Most of them won't lie to you because you'll just leave asap or be pissed off that they were full of crap.

This is all much more difficult to do if you're external, once you're in, it becomes easier to figure out if you want to work somewhere. I'm not willing to even consider an internal transfer unless someone I know will vouch for it. Too many horrible teams and a bad manager can mean you need to find a new company.


Also probably depends on the department you're in. Warehouse packaging employees seems to be under lots of pressure.


After the NYT article restrictions around transferring teams were relaxed significantly:

Previously I believe you had to stay on a team for a year before you could switch. Now it's something like a month.

If you find yourself on a bad team, pop open jobfinder and don't be afraid to abandon ship! Don't needlessly suffer.


They have a very toxic reputation in the Seattle area from the few devs I've spoken to that live there. Retention rates are also supposed to be amazingly low


+1


Funny, I've heard some of them too and they only made me want to apply there! I haven't given it too much thought but Jeff Bezos is probably one of the industry leaders I admire the most. ducks


That seems more like a values thing than a process thing. I doubt that this process would produce terrible work conditions if Amazon actually cared about their employees well-being.


Capitalism is a game of extracting as much labour surplus value as possible. By definition an article praising Bezos' success must have employees not realising anywhere near their value add.

Bezos does not do the actual work, he facilitates the labour of others. And if he can do that and underpay them more than someone else would then we have a winner.

This is the system.


Implying that skillfully faciliating and organizing other's labour isn't work? Well, early on, soviet government and their 1918 predeccessors thought the same. They ended up with a lot of 'liberated' workers and farmers, and no managers and business owners.

Strangely, they had to restore capitalist system and then to invite foreign specialists and managers to make the industry even a little bit operational. Turns out, there isn't much you can do with dumb labour alone.


It is work but workers are not in a position to bargain for their value added. Bezos keeps it.

We should tax land not income and stop forcing people to under-sell their labour.

Spare me the "Russia" extreme example. This isn't quantitative, we can dislike that extreme and also the extreme celebrated by the USA which fails so many.


You're saying that USA fails many as if it's a problem. Which implies that in your opinion, no one in a state should fail. Why?


Take what you read with a grain of salt. Amazon can be an amazing place for some and horrible place for others. The horrible stories are more interesting to tell and stick with you but doesn't make them the rule.


Bezos might have got this from Colin Powell's 40/70 decision rule. http://integratedleader.com/articles/40-70rule.pdf


The % of information missing versus the importance of the missing information... between the two I've learned the hard way that you wait for the right information. Perhaps Bezos way works well for capital rich companies, or it could be my instincts aren't great. I only know that every time I rush into a decision I pay more that I should and often have regrets.


I don't see waiting for 70% rushing at all. The point is that there is always another piece of information, so if you're waiting for the impossible 100% you'll never get anything done.


I thought of OODA Loop (Observe, Orient, Decide, and Act)

https://en.wikipedia.org/wiki/OODA_loop


In principle, the ideas sound great. In practice, the biggest decisions seem to have high switching costs (e.g. home purchase), by design, to reduce flexibility.

So it might prove useful to pursue courses of action that provide maximum flexibility.


The big problem is that the majority of people with 70% of information will go into f*ckup 100% certain. You need to have a talent of some sorts to make right decisions or at least reversible ones.


this is some self-help guru type of bullshit.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: