EP256 - Colin Bryar former Amazon Chief of Staff and author of Working Backwards
Happy episode 256, our penultimate 8-bit episode! (we’re going to have to migrate to 16-bit for next week).
Colin Bryar is author of “Working Backwards: Insights, Stories, and Secrets from Inside Amazon” (affiliate link). This is the definitive book about the unique processes, programs, and culture implemented at Amazon which have enabled its remarkable track record of speed and innovation. Colin had a 12 year career at Amazon and served as Jeff Bezos technical advisor (chief of staff) shadowing Jeff in all meetings for a year. His co-author, Bill Carr ran Amazon Prime Video and sat on Amazon’s S-Team. These two had a front row seat to many of the most seminal moments in Amazon’s history.
If you work with Amazon, need to compete against them, or just want to duplicate their success in another field, you owe it to yourself to listen to the episode and then read Colins book.
Episode 256 of the Jason & Scot show was recorded live on Wednesday February 24, 2021.
Join your hosts Jason "Retailgeek" Goldberg, Chief Commerce Strategy Officer at Publicis, and Scot Wingo, CEO of GetSpiffy and Co-Founder of ChannelAdvisor as they discuss the latest news and trends in the world of e-commerce and digital shopper marketing. Transcript
[0:24] Welcome to the Jason and Scott show this is episode 256 being recorded on Wednesday February 24th 2021.
I’m your host Jason retailgeek Goldberg and as usual I’m here with your co-host Scott Wingo.
[0:41] Hey Jason and welcome back Jason and Scot show listeners
Jason as you know one of our favorite topics here on the Jason Scott show is Amazon their culture and different business strategies tonight on the show we are really thrilled and excited to welcome Colin Breyer he’s an ex Amazonian and co-author of the brand spanking new book working backwards
welcome to the show call.
[1:03] Thanks for having me on the show happy to be here.
[1:07] We’re excited to have you Scott in particular as a huge Amazon Fanboy so this is a he’s trying to be cool but this is a thrill for him.
And so Colin Scot introduced you as an ex Amazonian but maybe you could tell us a little bit about your background and how you came to Amazon and what you did there and then you know what you’re doing now.
[1:29] Sure I moved out to Seattle and 1990 and worked at Oracle for about five years I was
so consultant red wire and after 5 years I realized I didn’t even know what Seattle look like so I left Seattle and started a company with.
With two other folks Charlie Bell and Kevin Millar and what we were doing at the time so this is in 95 were helping companies
take all of their internal data and help them expose it on what was then the nascent world wide web and most companies you know
we’re struggling to do that and we worked with a bunch of larger companies out here in Seattle Microsoft Boeing and then some companies are like tear W outside of Seattle.
And one of the small companies we work with was called amazon.com
and we realized that there was a really special place so from the moment we stepped in through the door and so we decided to join Amazon and that was so I joined Amazon in March of 1998
and Amazon was only a Bookseller
just based in the US and there were probably about a hundred people in the corporate department and 500 people and total and customer service in the Fulfillment centers
so it was a.
[2:51] You know very special place in you could tell that something something was going on and it wasn’t sure if it was going to work yet but things are moving fast and customers were validating what we were doing and
the press and pundits sometimes agreed and a lot of times they didn’t but it was fun to see Emma’s on transform from they did.
147 million dollars in Revenue when I started to and now this last quarter was a hundred and twenty five billion dollars in Revenue so it’s been fun to be part of that transformation.
[3:21] Yeah they’ve had to stretch the the cells on the Excel spreadsheet a little bit since you started and a hundred employees so I’m trying to think would you have gotten a desk that was made out of a door or or did you have actual furniture by then.
[3:35] No I had a door desk and you know he’s still get to our desks and I was lucky enough to my the email address was just Callin to so it was a pretty small place then.
[3:46] That’s that’s very cool and then.
You did a couple different roles in Amazon but one of them in particular is a pretty cool role and you might have if I am going to pretend like I didn’t read your book but I did you were the second person in the in that role right.
[4:03] Yes so I started out in the software group and worked there for about five years and then I was Jeff Bezos is
the internally the roll is called the Shadow or technical assistant and externally it’s more akin to a chief of staff and so I got to spit was very very fortunate to spend.
10 hours a day with with Jeff participating in the meetings and you know seeing how.
He thought in and was planning on on building a very large organization to be what he termed Earth’s most customer Centric company.
And it wasn’t just technical issues I got to experience everything from the Fulfillment centers legal PR
the commercial group the retail groups and also she has some pretty deep dive technical issues too.
So I was very fortunate to have done that and then I spent after that I went to IMDb The Internet Movie Database is their CEO which is a wholly owned subsidiary of Amazon.
[5:02] Yeah and as I understand it the the the technical adviser primarily entails training up Jeff in tennis to get him ready for celebrity tennis tournaments is that.
[5:13] No that was a very small part of my job was less than a day yeah.
So yeah that that was an interesting Adventure but really that what the job entailed was two things you know when Jeff he asked if I wanted to be this next technical advisor and
you know I did rather than jump at the opportunity I said can I take the weekend to think about it but first can you please tell me what
um success looks like in this role because it’s too important of a role and I’m to take if I don’t think I can be successful in this role and he said well the first thing is you know just to help.
Jeff Lewis helped him be a better CEO and you’re making sure that the right issues and teams get in front of Jeff and.
And I could go places in the company that he couldn’t but then the second part is the way he put it is a we want to be able to model how each of us think.
So after this role ends when you go somewhere else in the company
that’s going to have a pretty good idea of his vision and your core tenants and Leadership principles and be able to move into the company so it is a rotating role and I was in that role for about two years.
[6:27] That that’s amazing and the present your predecessor in that role was Andy Jesse wasn’t it.
[6:32] Yes Andy was the first technical advisor and you know relied on his advice and counsel do you do too
tell me what I was getting into and you can see the phenomenal job and he’s done since then you know there aren’t too many people who get a chance to
to do this role so I realized I am very lucky to have been one of those people and you know one of the reasons bill and I decided to write this book was to talk about some of the principles and processes
that really are you know that secret sauce of what makes Amazon work.
[7:10] Cool let’s let’s dig into the book The the book is split into kind of two parts you have
the first section which is being Amazonian and then the second one which is kind of case studies of applying that called invention machine at work
in the Bing Amazonian section you guys go into a lot of my favorite topics and sadly we only have you for an hour I feel like I could
essentially just talk about this forever but you talked about the Amazon leadership principles the six-page note that is the
kind of the Keystone of every meeting the bar raiser program
left I thought it would be appropriate maybe just started the title concept which is working backwards give an overview of what that means and you know maybe an example
don’t use AWS causal ask about that later but as an example maybe a of how that gets used inside of Amazon.
[8:07] Sure so working backwards it’s a very specific process used at Amazon
to look at ideas to vet them and decide whether to bring them to Market be it a feature of opening up a new business
and if you have to remember one thing about the working backwards process it’s this it’s that you start with the desired customer experience and then you work backwards from that.
It sounds simple it’s actually pretty hard to do and it’s different from how a lot of organizations make decisions a lot of organizations use What’s called the skills forward approach they look at things and ask questions such as
what are our core competencies we know what are we good at what are our competitors doing.
And how can we nudge into this adjacent Market if we get 10 percent market share what’s that going to look like and you know a SWOT analysis the strengths weaknesses opportunities and threats is a typical analysis that.
[9:01] Companies use to decide what to do next but often a word that doesn’t get mentioned in that analysis at least up front is the word customer.
So Amazon decided to invert that and and say we want to make sure that the customer is front-and-center from the very germ of an idea.
And so Amazon developed this working backwards process.
And the primary tool that Amazon uses for the working backwards process is the pressure release and frequently asked questions document so it’s a type of narrative called the pr FAQ document.
[9:37] So that if anyone has an idea and again this works for something as small as a new feature on the IOS app or if Amazon is deciding to get into a whole new line of business or move into a brand-new geography
the first thing that a the person who has the idea or the team that has the idea does is they write a one-page press release so it has to be one page or less.
If which forces you to really crisply Define the idea and the press release has a couple different components one is a clearly defined what is the customer problem you are trying to solve.
And you know and that also it can take a couple of iterations that in the next part is you have to explain to the customer.
Why they might be interested in using the feature or buying the the product.
And then you go on in that press release you typically you can use a quote from a customer or if it’s a something for a partner the partner talks about why how this actually solve their problem.
And if you and so this is an iterative process once you write your press release.
You read the press release and if you don’t want to go out and buy that product or use that feature the service you stop and you rewrite the press release until you get it right.
[10:52] And then the the next step in the process is the FAQ process and you can break that up into two primary components and external FAQ and an internal FAQ.
And that external FAQ are think questions that you ask an answer that would typically go to customers or if the Press.
How much is this product going to cost.
Why should I use this product versus what’s out there on the market why should I change my behavior and what’s in it for me if I’m going to go through some extra steps to go use this product or service.
[11:25] The internal FAQ is a series of questions and answers about.
What are the tough problems that the company is going to have to solve in and how are they going to organize.
[11:36] To actually get together and solve these problems to bring this idea to Market so if some examples there could be.
Kim can’t how and can we build this product with the bill of materials that’s less than 200 dollars to get out to the market at the desired price.
Taking me technical issues what are some unknown technical problems that we need to solve.
And how are we going to organize and approach and solve these problems Legal Financial issues or
privacy issues or if it’s a sales B2B this requires your Salesforce do we use a direct sales force or we’re going to partner with someone.
And there is this is all an iterative process and a lot of ideas don’t actually make it through the end of the working backwards process.
[12:21] And the ones that do have gone through many many iterations of meetings where people weigh in you’re missing a key.
Fack you know so let’s go ask an answer that and come back next time.
And if this is by Design By the way because it’s meant to one ensure that the customer is not forgotten
but do it saves time because it saves you from moving in the wrong direction you know people talk about speed a lot is important
but velocity I think is is important in velocity is the vector a and the vector of speed and direction so this helps you make sure that you’re moving in the right direction at the beginning and conserving what’s typically
companies bottleneck resources which are Technical Resources.
[13:09] The so it seems so then you’ve got this idea Factory right and everyone’s creating these things.
And I imagine they’re all pretty good then at some point someone’s gonna have to decide like there seems like there’s always going to be an abundance of them even the ones that you know even given that some don’t make it through.
So then does does Bezos just essentially say all right here’s the top 100 we’re going to draw here like who’s sorts these things in part eyes as I’m how does that work.
[13:37] Looked it’s typically it goes to who’s ever controlling the the set of resources that need.
You know that are needed in order to get this done and for a very large.
Initiative or what Amazon calls a one-way door once you do it you know it’s very expensive and difficult to roll back that will float up much higher in the organization but it could be you know as simple as we want three people to go work on this new feature on the website so whoever controls those whatever the
appropriate management level is that controls those resources that’s where it gets done and you’re right in that very
most people have good ideas it’s just is this idea we’re doing is it big enough and is it in is it worth doing now those are the the types of questions that you have to ask
given the resources and constraints that the company has so for low very large projects that goes up to the.
Esteem The Witches Jeff’s you know.
Direct reports the management team but most of them are smaller than that because Amazon usually works Inseparable teams and so who’s ever controlling those single-threaded separate both teams typically makes that decision.
[14:53] Yeah let’s so
Let’s do an example a simple one in this is Jason’s favorite let’s say house plants he always used to use this one to something that Amazon I guess wouldn’t do and now they do so so I have the idea to do houseplants I write a press release
Amazon now ships houseplants and I talked about how we have.
They arrive at your door fresh and you know a selection of thousands and you know but then do I have to tell you like how do you sighs that opportunity versus I don’t know.
You know B2B cogs are widgets and cogs.
[15:33] Yeah so
that you first of all have to one of the questions is in the fact you have to address what’s the town that the total addressable market so how big is this idea you know some typical questions would be.
We don’t have life things in our fulfillment center right now how do we handle you know how do we handle that and you know what how long can they stay in the Fulfillment center before they need attention you know Electronics depreciate you know
some of them appreciate 10% a month plants die if you don’t water them I would so you know you’d have to address issues like that how are we going to keep our inventory alive before we get it out to customers and then
but in terms of it’s a great question about how do you balance that with a B to be completely unrelated project and.
[16:24] That that prioritization is really tricky and It’s Tricky for a couple of reasons because.
[16:30] Even out the the pr FAQ stage you don’t know
really how big of an idea this is because you don’t know customer adoption that’s very tricky to predict even if you have a great idea and you don’t know how long it’s going to take to build and deploy the
the technology or that you know the heavy lifting infrastructure to handle plants in this example and so Amazon
a lot of times what they do instead of made it making that prioritization decision they take a step back and make a resource allocation decision for given areas and for this one that would probably have been done in a yearly planning cycle
to say we are going to devote for our B2B efforts we’re going to devote this many people are you know these many organizations or groups are going to
very large now but are going to focus on B2B issues and here’s the
you know new category expansion for the the retail business and if you do that up front and then you have your the teams are at that point separated
then you don’t have to it’s hard to prioritize between apples and oranges
and so Amazon doesn’t want didn’t want to make that prioritization decision because you often get that wrong and so just taking a step back and use a different decision making tool which is resource allocation and you do that you don’t have to do that every project
you can do that once a year or once a quarter and then balance resources as more data comes in.
[17:59] Yeah and then and then what’s the so alright I’m in the B2B group and I’ve got my you know is it,
what’s the unit of allocation for resources at people hours is it dollars is it gummy bears you have seen people do all these really weird things where they’re kind of like you know you get this mini you know
seats on a train if it’s engineering how does is there such a thing at Amazon like that.
[18:23] Well there are you know there are some constraints and you know setting hard constraints at the beginning of a planning process actually
it typically saves iteration you know if you say hey send me all the any ideas you want your ask is going to be much bigger than you could ever do
and so setting some constraints about here’s the
free cash flow that we anticipate we’re going to have to invest back in the business some of it are just you may want to do something but
you may not be able to hire a hundred software engineers in the in the time you need to hire those people so you could say here’s our staffing and here’s our hiring rate for the year so.
We have what we have at this point in terms of Technical Resources
but it’s a combination of what it’s resource constraints and in some cases it could be dollars and some cases it could be bottleneck resources like software Engineers or data scientists
or it could be you know fulfillment center capacity so you have to know what your bottleneck constraints are and now that would be how you make those types of decisions.
[19:28] So in my plan example I’m going to say I need you know I need the retail team to kick in and create a category for me and I’m going to need three developers to add all the attributes for houseplants and I’m going to need
a photographer to take pictures of them and I’m going to need a greenhouse outside every fulfillment center and I’m going to need a I don’t know
what is a.
Plant person to you know an expert is it kind of like that like some of them I’m drawing resources from other teams and others I’m hiring or how does that get expressed in those that process you’re talking about.
[20:03] So for for some of them will touch other other teens and
Phantom Zone we talked about Loosely coupled teams not completely separate and independent so there are some shared resources and you know especially for smaller organizations you’re not going to have.
Illegal rep you know that for each of the small groups you’ll kind of share those
those resources across there but you would need to identify here are all the things that we need to get done and you know in terms of Transportation Logistics design and and for this those shared the pools of resources that you are going to have to
get some of those allocated for that period of time but that having been said if the issue if the the idea is big enough
you know you you can justify getting those resources on your own and one of the things that when a great frequently Asked question is that’s been asked several times at Amazon is
what things are outside of your control.
Do you wish that you had under your control and how are you going to organize and how would you organize to to bring those things under your control and that’s a continuous process you know if it’s your always short design resources for instance
and you work too.
[21:24] Get design resources on your own team somehow if that’s the right answer or if that Central group you needs to grow or double in size you know there’s not there’s often not one.
Universally right answer for every company but knowing what things that you’re it’s.
Amazon wants people to be control in control of Their Own Destiny and so asking
what things are outside of your control that and because it’s hard to ask people to be accountable when you don’t give them resources to get things done so Amazon tries to make sure that that happens.
[22:01] Awesome and calling you’re being like really polite by humoring Scott but the reality is his one-pager would never get off the ground because if the
if the one pager was Amazon now ships plants that’s not very exciting the one-pager should be Amazon just shipped its billionth plan.
[22:17] Yes yeah you have to attack it does have to be a big idea to work you know.
[22:23] Ink I think that’s increasingly true right as Amazon becomes a bigger Enterprise those those new Ventures have to have to be bigger to be relevant.
[22:35] They do have to be bigger to be relevant one thing Amazon is is unique is their patient you can plant seeds so some of these things take years to grow
into something big and you have to.
Yeah Jeff put this in one of his shareholders that you have to have the institutional memory to know what it’s like to go from you know a 1 million
Dollar business to 10 million dollars to fifty million over relatively short period of time that’s not going to move the needle in a hundred and twenty five billion dollar quarterly business but.
Given enough time and if the total addressable Market is big enough.
Then it is worth doing and you do need to be excited about those types of things and pay attention and to them and you know cult
I guess the seed analogy is apparently a proposed since we’re talking about house plants but Amazon is patient and if it’s big enough they’ll wait and work to get it done.
[23:39] The so I want to Pivot a little bit the if we didn’t spell it out up front the the book is really a tool to give some insight.
Into how Amazon organizes.
[23:56] It steams and runs its business and and the structures and Frameworks that Amazon has in place to be shockingly Innovative in spite of their.
Now tremendous size and success and you present it in a way to try to help others.
Decide if and how they would Implement some are all of these things in their own organization so it’s sort of a business manual if you will.
And you you kind of go through a bunch of stuff though wieder ship principles which I think there’s 14 leadership principles now the bar razor program Which Scott and I have had several of our razors on the show is guess.
Um which is Amazon’s hiring
process but one of the things that comes up a lot in the book and that you’ve referenced a couple of times already tonight is this concept of single-threaded separable teams and sometimes referred to as though I to Pizza teams for example.
I’m wondering if you could talk a little bit because I think you were there while that sort of before that that philosophy was fully embraced and as it was implemented can you talk a little bit about.
Um how that came to be and how that has served as a Advantage for Amazon.
[25:12] Yeah sure sure I can do that and you know the the areas that you mentioned that one of the reasons we wrote the book was because a lot of this work is you know it’s under the tip of the iceberg it’s things that people don’t see but in people ask always ask how many
Echo devices are sold this quarter or how many Prime members are there going to be and.
Well that may be interesting it’s not that helpful to most organizations and we felt that Amazon had made some significant advances in the field of management science.
[25:44] And that’s why we tried to put all of these Concepts together to say here’s how
you can take a small organization and use some of these principles some of them were inspired and you know stood on the shoulders of other companies before us
but so we tried to organize them in a way that’s useful and helpful for the reader and in terms of separable single-threaded leadership
that one was a journey
you know for instance writing narratives you can just say we’re going to switch from slides to narratives and make that change which is what Amazon did you may stumble a little bit and it takes a while to write great narratives
but that’s an easy change figuring out how to get to this single separable single-threaded leaders
was a multi-year journey and you know when I started at Amazon you know I’ve started in the software group and and
Amazon was growing so fast but it was already so large that there weren’t any there were very few commercial software applications that you could buy
to help solve the problems that we are trying to solve we are already well beyond the tolerances of most commercial software so we had to
build and build fast just in order to keep the lights on and this is in addition to opening up a new geographies or internationalizing the code base and.
[27:09] Moving into these different categories which you know they have different attributes so you got it
you have to change how its search looks like with the order pipeline looks like if you want to do a Peril you need to have size variations like size and color you know books don’t have variation so it was very easy but we realized that we’re adding a whole bunch of people
and we weren’t moving all that much faster and even worse we are spending more time coordinating than
actually doing and you know so that ratio is was getting out of sync and one of the things that Jeff is
appraisals is particularly good about is you can take a look at a trend in then project it out when what is this going to look like five years from now is he 7 years from now and
the prognosis was not good we had a tangled.
Code base and you know it was all one executable even at one point for the website was called overdose.
But and so that meant you had a you know.
[28:09] Couple hundred software Engineers working and stepping on each other’s toes someone would change something that you didn’t even know and it would break your stuff for or vice versa and then if you wanted to get something done
you often would have to control another group to say hey can you work on this you on this library and so how can you change some of these things
and we realized that what so we had a huge technical problem
we also had an organizational problem you know those same the same dependencies existed on the organizational side you’d have to go ask for resources from the design team or from the Fulfillment center team and to figure out how to
to get things done and a lot of companies what they would do is they would build.
[28:53] Better processes and we build solutions for collaborating and communicating and Jeff
did said just the opposite he said I would like to have an environment at Amazon where we don’t have to collaborate and communicate.
And we have small separable teams and we started off with the idea of a two Pizza team
and and the reason it was called to Pizza teams as because two pizzas should be to be able to feed the entire team so you couldn’t really have a team more.
[29:24] Scot and I would have to be individual teams that.
Yeah that was that exercise those left to the team owner but how much PC would allocate but Engineers don’t like to go on hungry stomachs so and so we tried separating these teams and.
But in order to do that you’ve got to change your technical architecture and you also have to teach people how to be autonomous because in this prior environment you know.
You couldn’t do anything really on your own you had to go ask so many people that it was more top-down here’s the next thing we’re doing and you know so next quarter you’re going to be working on an initiative that you didn’t even know about the quarter before it was kind of disheartening.
And so we had to untangle and the code base build What’s called the services based architecture
a lot of people do that now and you know it sounds easy didn’t really exist at the time so we’re it also inventing a lot on our own on how to build this type of architecture
and then we had to separate the decision-making process for the or two.
[30:28] So some roles in organizations like a chief product officer kind of go away
because you want those product decisions to push them down to these small separable teams you don’t want to have one person or group make all of the
product decisions and you know same thing with that engineering decision so we had to decouple and distribute that
and you know white where said it was a journey that stuff was hard to do we also had these things called Fitness functions which were basically a composite metrics that would a single metric.
[31:01] Which is a composite of individual ones that would measure the progress of a team and we realized we were spending
whole bunch time arguing over you know should it be twenty percent speed of the service and 60 percent revenue and you know and you know 20% something else and it just was a waste of time and we so we
we stopped doing that in the fitness functions and it turned out what was the.
[31:24] I would say the high order bit that made them work as their separable teams but a single-threaded leader
and the best example I can give the this at Amazon is there was a project called
self-service order fulfillment and we don’t have exciting names for some of these internal projects but what that meant is we we knew that we wanted to expose some of our functionality in the warehouse the logistics centers
to third parties and so we wanted to make it self serve where people could fulfill orders and it
was it good idea but it never got done and when I was working with Jeff as as technical advisor we would do we would go in for an update on it
and it would be yeah we have to talk these eight other teams and we’re making some progress the next update six weeks later there’d be a different
person giving the update and your different leader and it was kind of this rotating thing and so finally Jeff said Bezos said to Jeff Wilke who was running the opposite group at the time said you need to assign a senior leader
can’t you do to make this happen and I want that person to work on self service order fulfillment and nothing else but self-service order fulfillment and so Jeff Wilkie chose Tom Taylor who is a VP in the group
Tom had a big job at that time and you know Jeff Wilke went in and said your big job is no longer you’re worried you’re going to work on a project that
is risky it doesn’t generate any revenue and it’s and you’ve got to go figure out how to do this but.
[32:53] I’m woke up every day figuring out how to organize and get this thing done and not you know is it year year-and-a-half later
it launched into what was now called fulfillment by Amazon so very big business and I’m not sure if that would have gotten done it certainly would have gotten done at the time you know.
The time it took to build something that big without Tom and a single-threaded leader so Amazon took that and use that as a model to for how to get other things done.
And Dave Lim who is the senior vice president of devices now at Amazon has a great quote
and he says that the best way to fail at inventing something is by making it someone’s part time job.
So that that is an example of where your Amazon just took a slightly different approach on how to organize around.
Really working on the things that matter and that will drive the needle.
[33:43] That I love that and it one of the things that’s fascinating to me about it is it seems like it’s worked bofur.
Technical as a technical solve like it like you guys organized software that way and apis and and the sort of space architecture and all that and you’re organizing.
The human resources that way as well and it seems to apply equally to both I do have one question though.
From what I hear the one thing that doesn’t seem like a jives perfectly with that is it seems like you hear a lot of people talk about the s team and and you know the biggest decisions in the company getting elevated to the s team
in a way that s team sounds like kind of the antithesis of.
Single-threaded readers if there’s you know like at the s team it sounds like the finance guy can critique the software approach or vice versa or those thing am I misunderstanding.
How the s team works.
[34:35] The way that that the the operating Cadence at Amazon is there’s a yearly planning cycle where you have some tent poles.
[34:44] About just what are the constraints that the organization has to face.
And you know each team then or group comes in with their it’s called the operating plan one their op1 plan.
About what can they do and they come in with the resource ask and at some point you do have to rationalize.
If the ask is bigger than the individual resources you do have to figure out how you’re going to take a fixed pool.
Of capital so I think you know a comment we get a lot is
well Amazon has unlimited capital and unlimited software Engineers that can just appear magically whenever you need them at the door whenever you need them that’s not the case there and it was actually it was very difficult to get resources allocated to
a project that you are working on so there was some friction there but it was by Design.
But what did not happen is there wasn’t a lot of thrash after that after you make that allocation so here’s our yearly plan and rather than say what the team needs to do and how they’re going to go do that the team would commit
given this set of resources that I now have for the upcoming year.
Here’s what I’m here’s what I’m going to commit to and here are the set of initiatives that I think are going to get me to you know to achieve these goals so you do what what I’ve seen.
Some people go overboard with these separable teams at and just make them totally autonomous
and and I think that you need to come back.
[36:14] You know once a year sometimes even once a quarter just to check to make sure you’re moving in the right direction the right direction and staying true to what you you know sanity check on are we making progress on the goals the company goals that we want.
So there is a true true up. But it’s on a yearly basis for the most part in the operating Cadence of Amazon.
[36:37] Well that’s interesting you run on these annual Cycles but let’s say.
I don’t know some earth-shattering new thing happens in the middle of that cycle what’s the process for kind of is there a like in scrum or agile software there’s a way of kind of just saying.
Scrap everything we’re going to reorient it is is that a thing at Amazon or know you stick to these annual cycles and don’t deviate.
[37:00] You need to take a look at the data that comes comes in and adjust and in so I.
Don’t know if I’ve ever seen a yearly plan executed a hundred percent exactly if you were good to go back year that everything
happened the way that we thought it would you have to move fast you have to move with
less information than you would like you know about 70 80 percent of the information you have to end to make the decision so you also need to pay attention to what’s going on and to be able to adapt quickly you know there are some times where.
You’re like Amazon Prime for instance is a good example where.
There are exceptions to the rule hey we’re going to go launch Amazon Prime Jeff said this and it would there’s an October and we’re going to launch it by the end of the year it was a
you’re not the biggest project Amazon did that year but it was it was a substantial one and it was a fairly short period of time so there are exceptions to that rule and you do need to
you know to be agile
the group do that has committed to achieving certain goals that STM doesn’t really tell them how they’re going to go with cheve those gold so if something changes the group you know the group.
[38:16] And question adapts and they can say hey I’m no longer going to work on Project a because Project B or does new project that I didn’t even think of you know back in op1 comes in the fall
is now worth doing so I’m going to set these other things aside and you make that exception the planning tools to help you make the right decisions but if more information comes in over the year
to tell you that hey if you stick with this thing you’re going to make the wrong decision you know you change the plan.
[38:47] Got it and then says that’s been super helpful to walk us through those different
principles and then second half of the book you kind of think of them as case studies and that’s the invention machine at work.
I was going to ask you about AWS but then it occurred to me I’ll make it your choice so anything you want to talk about what would be a good example for listeners of you know
in your in your memory of how Amazon applied some of these things and any fun stories in there always always welcome.
[39:18] Yeah we’re going I’m going to talk about AWS and and so you know Kudos huge kudos to Andy jasion and his team for inventing cloud computing but
a couple things that are I think know too little about the evolution of AWS so Andy you know there were signs before well before
S3 and ec2 and the
queuing service were the first three AWS services that came out well before that there were signs that either something going on here with web services it’s just a better way to build software
and our internal software and you know Engineers were using it we were using it with third-party Sellers and with affiliate program and Andy had put this plan together.
It’s been said you know there may be something that here and we should adopt this you know model and go try it.
And any could have had any job that he wanted.
At the company at this point he had just spent I think it was a year and a half working as Jeff’s technical advisor and he chose to go to a non-existent business that had a.
High level of risk
we also didn’t know where other companies were on the path of inventing cloud computing we were looking at the same data that people.
Microsoft Oracle IBM Google we’re looking at boot go to some developer conferences and see the same Usual Suspects there.
[40:47] So we had no idea what they were doing so what I think one notable thing is that it’s okay at Amazon to go take a risk it’s not a career.
Breaker to go from a big business to a small business or from you know.
Job where you have a lot of head count to go start to build the new idea and invent something and that’s one thing and then two is that.
[41:10] We talked about the working backwards process
especially for something caught the term cloud computing didn’t even exist then and the initial ideas that we had about web services is what we call them in the beginning
we didn’t know what the fundamental units were and you know so ec2 is the elastic compute cloud and that’s
your computer you buy compute units of compute power in the beginning we thought that was really going to be
provisioning which was a problem that we had internally a teams would write their software and then
they wait six weeks for the hardware to arrive and for people to provision the hardware and then push it out there it was also hard to also get the right software on each of the computers that you needed and then if the I didn’t.
[41:57] If the idea didn’t take off reclaiming that Hardware so you could send it to someone else was another big heavy lifting project but so we thought it was going to be provisioning
but there is a journey that we went through and for AWS we basically wrote documents for about a year and a half
and reviewed them it was you know Andy and whoever the if it was the compute team or the the storage team
and Jeff and I would be in a room and we’ll be reviewing documents sometimes we wouldn’t get past the first page because we realize hey there’s an issue here that we don’t really understand or we haven’t gotten to the
you know really to the core of the issue or defined what it is for the customer
and you know they’re in some metaphors just popped up during this time where one very powerful one is that we want to provide the same world-class Computing infrastructure to a college student in a dorm room
then someone who works at a company like Amazon and and that really clarified things and you know the other thing with S3 is,
you how does three fail you know you can either have it fail for an hour a year which is bad if you have.
Hundreds of thousands of businesses relying on it or if it does have to fail you can have it fail gracefully.
[43:14] But just and you know one transaction every you know
couple million transactions a oh go try it again that’s it those are two different failure modes and you have to build something very different.
And we also knew that outgoing once this thing got out the door it was going to be hard it had to get better as it got bigger
so you couldn’t throw this thing over the fence and then decide what to fix because so many people would be relying on it’s a much different relationship that you’d have.
So just the I think in the notable thing about AWS is it was an experiment.
And we felt weird in a land rush we wanted to get out there first it would by no means ensure success but it sure would help.
[43:57] But we stuck true to what we don’t haven’t really defined you know you were using this working backwards process we haven’t really defined what,
what we’re trying to solve and and really identified that the core technical issues and then you know there was also some astounding engineering
work and it advances that the web services team did that went along in the background because we knew sometimes what we wanted to do but we hadn’t figured out how to do it.
[44:23] So just the Journey of getting a brand new idea and you know for a company to be able to say this is not our Core Business.
But it is something that we have we think we can do as good as anyone else on the planet and it’s worth worth trying
there are some Skeptics inside Amazon and even at the board level about you know why
why are you doing this when you’re still trying to get your retail business working and improving on the retail business you know prime it just lunged at that time so be willing to be misunderstood.
For a long period of time if you go back and look at some of the quarterly announcements Jeff would say well we’re working on it was web services and digital and and.
Said that for many many quarters and those turned out to be two very large pillars of the company but they were started out from you know.
Risky ideas most companies had made the transition from physical delivery of goods to be a pure digital player in terms of movies
Books and Music also so that was just another transformation that happened.
[45:32] Happy to go into more detail on it on any of that but I think the notable things are what Andy did and then also just sticking to the working backwards process because ultimately you want to solve customer problems and if you’re in and if you solve customer problems
it will work out in the long run so Jeff firmly believes in he
you know told us all that in in the in the long term the interest of customers are perfectly aligned with the interests of shareholders
and so if you do what’s right for your customer it will work out in the long term and you’ll build a company that he can be proud to tell your grandkids about.
[46:07] Got it when did maybe you you left before then or got moved to something else but like when did when did you know or Amazon know that the cloud thing was going to be pretty big.
[46:23] We had you know we had a suspicion that it was going to be big and and.
[46:31] I think that it wouldn’t it wasn’t really proven until you when S3 first launched it was.
It wasn’t an overnight success.
But once another service ec2 came with it where you know you weren’t you didn’t use a storage service and then have to move over to your own data centers to handle something
once ec2 and S3 started working in conjunction it was a lot easier to build some pretty cool applications and you know that was another
tenant that we developed during this working backwards process.
Present a single service in itself is it going to be all that useful you need to have you know a critical mass of services that work together in the cloud
in order to really make larger organizations you know to jump on the bandwagon and start using it.
So I would say you know after ec2 and launched and then you’ve got to see what people did with the ec2 and S3.
We knew that that rocket ship was going to take off.
[47:34] Yeah this is where the Tam things tricky right because I’m sure the original paper the tan was pretty small and you know now it’s probably like thousands of times bigger than that original tan anticipated.
[47:47] You know for success is this large you know you can think big but that’s total addressable Market we did know that there’s going to be a new paradigm on how to build and deploy software.
And if we could do it it’s basically the business-to-business software Market you know that that’s huge.
And answer so we knew that it was a large number you know virtually.
Unconstrained if you want to think of it that way in terms of if you can get it right there’s a lot of work and you know even right now you take a look at the total compute.
Our you know that that’s going on are the software development it’s still there still a lot of Runway ahead of AWS.
[48:31] Yeah and another thing I don’t know where this lands in the principles but then there’s this very unique to Amazon think other people are copying it now but this whole idea of you know Walmart would have taken that.
That infrastructure and they would have viewed it as this super proprietary kind of a thing that they would use internally right where does that culture of opening it up.
To external users where does that come from.
[48:58] I would say the root of it comes from customer Obsession and I’ll give you an early example of something like this which is where Amazon wasn’t you know owned all of its inventory.
And and so when you went to a detail page a product page on Amazon there was only one seller it was Amazon there’s only one seller on the platform.
And it was a controversial issue to say should we open it up to on the on that product page to third-party sellers we had tried in auctions.
[49:27] Product on a separate tab at the time and if you remember those then there was something called Z shops and turns.
Known went over to that neighborhood because all the cool kids were over on the product page of detail pages of Amazon and in you know the for instance you.
The head of the retail group or the head of the electronics category would say are you kidding me
I’ve done all this work to get my scarce allocation from these vendors you know try to get sharp prices on them and try to keep them in stock and now you’re and and I’ve created this great detail page for this electronica
item and now you’re going to let any third party cell right inside my store now you know how is this does this make my job easier and and how is this good for Amazon
and once you know and it was Jeff who said looked at it and said well
where how big amazon gets it’s still going to be a small part you know percentage of overall retail and and ultimately we’re in the business of allowing customers to make purchase decisions so if we don’t have the product in stock.
[50:38] We want to eat we want that we still want the customer to be able to buy that Earth we don’t have the lowest price we still want them to be able to come to that detail page and conduct a transaction to find out more about this product.
And buy it and if you want to make that product page to be the best place on the web for that particular item you have to have multiple sellers you have to have the best item Authority information about that and yeah by the way you now
is the general manager of the electronics group your job is a little harder but you know it’s it.
Making these things making your job easy isn’t what Amazon is all about you know we’re trying to solve customer problems and this is the best way to solve the customer problem so that you know I think if you look at it from that point of view
then you say oh yeah we have to open up our product pages and create this Marketplace
initiative which is now you know now outsells the owned inventory business on Amazon is as you guys will know.
[51:37] Yeah yeah it’s crazy it’s annoying how many stories like that Amazon has of these.
Things that in hindsight are enormous successes like the marketplace but at the time like had to be hugely controversial difficult decisions.
One you know as I was reading the book one of the things that kind of recurring theme was a lot of these business structures and processes.
I feel like they were really invented to help Amazon scale Beyond Jeff right like.
You know to maintain Jeff high standards once he couldn’t meet every employee personally we need a bar razor program for hiring and we need the business principles to sort of indoctrinate everyone in the company.
The big news this quarter is all the Jeff’s are leaving Amazon and so I’m sort of curious like.
Do you believe that all of all of this infrastructure and culture that that you guys all put in place.
Are going to enable Amazon to sort of keep clicking at the same level you know when when Jeff was like a little more involved.
As he sort of disengages and spends more time on Rockets or something or or do you think that’s gonna be an inflection point for Amazon it’s hard to in my mind it’s hard to still be.
Day one company when your your founder retires after 27 years.
[53:00] So you know Jeff has he had spent since the time I was working on he devoted a whole lot of time.
[53:08] To try to instrument the company and encode some of the knowledge has and principles that he you know where he wanted to take Amazon
in and make them repeatable processes you’ll it for there’s no one at Amazon who could say let’s turn around and be competitor Focus rather than customer-focused it’s just it’s in Amazon’s DNA
so first of all I think Andy he’s young he’s the right guy for the job if I had to write the Amazon CEO job description it would be someone who is steeped in Amazon’s culture
able to build
you know large multi-billion dollar businesses and work with small teams you know and jump in between the two and bonus points if you built.
Business from zero to ten billion dollars faster than Amazon did and he did all of those you AWS got four to ten billion dollars faster than
you know Amazon the company and so I you know I think it’s the Amazon is in good hands with with Andy but I think if you look at the legacy of.
I’ve what Amazon is in Jeff is going to leave
you’re at the end of the day these hockey pucks and cylinders we have in our kitchen or the two-day and delivery is going to seem laughably primitive
sometime in one day delivery will seem laughing laughs Ali primitive sometime in the near future but what is a lasting thing is really this in the ninja machine and it’s Jeff’s term.
[54:31] That he created at Amazon
and he was always very upfront about it and he would talk about some of these things about long-term thinking about you know you read the shareholder letters about separable teams and you know he’s been up front about the working backwards process so I think that it’s
these are processes where you don’t.
Have to use the stick to get people to use them it is more carrot approach because once you start using them you realize this is just a better way of building a bit and operating a business,
you know you don’t have to Once people start writing narratives if you were to tell them to stop that go dumb it down and use slides to you know
convey a complex idea you they look at you like a deer in headlights no matter who that person was so I think that there’s still a lot of innovation to come from Amazon and
you know whatever company or initiative Jeff Bezos is working on it will be very fortunate to have them but you know there’s there’s.
There’s a lot of people at Amazon who will continue to operate and tweak and improve this invention machine.
[55:35] Yeah one of my favorite lines from your book Colin was.
You talk about how many people say oh sure Amazon is successful but you have unlimited resources and Jeff Bezos and and you and Bill pointed out like.
Hey for most of the time we are there we are heavily resource-constrained that’s not not true at all and you know all of these processes can absolutely work without Jeff.
Although if Jeff’s available the work on your project we would both highly recommend him.
[56:02] Yeah and that still holds true.
[56:07] When the one thing that’s been interesting Colin and you’ve been writing a book so maybe haven’t seen this but Shopify is really kind of ascending and getting a lot of play as kind of a
you know an alternative to Amazon and they talked about arming the rebels and this kind of thing and then there’s also a wedge in there in that
Brands don’t love Amazon because Amazon.
They love they want the brands there but they want to control the price and there’s kind of I’ve had brand say to me it’s a love hate hate hate relationship kind of thing
so it’s going to be this really interesting battle we talked about this on our show a lot and then recently it was in the press that Bezos was getting more involved in the business
to kind of formulate a Shopify strategy that was right before he kicked himself upstairs what you know.
But in listening to you think about the customer it almost seems backwards for someone at Amazon to have an initiative that’s kind of like you know what are we gonna do about this competitor Shopify how would you kind of project what do you think they would do and with.
What’s going on there.
[57:08] Well I don’t have any first-hand data or information.
Give you here just to be very clear and if it does get back to really.
[57:20] If something’s worth doing you first of all have to identify the customer problem that you are solving and the customer problem isn’t to go
take over Shopify and you know so it’s it’s how can we serve our customers better be they third-party sellers be they bite you no buyers on the site you know and and and what could how can we organize to solve those problems and so you know
that that’s just the way.
Ideas are developed you know I will say Amazon does occupy a different place in society than it did you know five ten years ago and you know some of these things
are going to be worth putting in the public dialogue and you know that that’s part of being a company that’s you know
at a half a trillion dollars in yearly Revenue but you know I can’t predict what’s going to go on there but but
Amazon whenever there’s tough decisions what people at Amazon do is they fall back onto these 14 principal leadership principles because that’s what they’re there for its they’re there to make the tough calls.
[58:32] And and so while I don’t know what that what Amazon will do I know that after they do it if you read these leadership principles and then you listen you know
listen to yourself to say in the back of your mind that the long-term interests of.
Customers and shareholders are completely aligned it probably will make sense.
[58:55] You know we are running up on time but I do have sort of one last last thread for you I know
that the book is obviously intended to help help folks adopt some of these best practices from the Amazon and if I have a right I think you and Bill.
Also consult with some companies and and sort of help them adopt some of these processes.
I’m curious how successful or difficult outside entities fine some of these things I got I’ll give you a personal antidote.
I’ve hired a lot of X amazonians in my life.
And I’m always super excited that I’m going to get these people that you know come in and write these like you know super detailed six-page narratives and stuff and and what ends up happening is no they all do I really crappy PowerPoint because they’re all.
Like tired of reading the neck so I like part of me wonders like is there some Secret Sauce in Amazon like you know obviously we all believe some of these things can be useful in many other companies but.
Is there an endemic in managing Amazon and.
Cohesively doing all of these things together that make them work better than than individual bits and bytes do outside of the Amazon that’s fear.
[1:00:13] Well I would say that the first two that you’d have to do if you don’t have them and you know some smaller organizations don’t is defining who you are and the leadership principles you know so the idea is not to copy Amazons and.
You need to come up with your own about who you are and then the second part is that bar razor process the hiring process
is how do you vet new people coming into your organization because you want to use the new people you want coming in you want them to reinforce your culture and if you’re not deliberate about what your culture is and how you decide
is this person going to reinforce my culture or change it
if your culture will change because you’re going to get a culture as your company grows it just is your choice if it’s whether the one you want it to be or whether it will become bait you know whatever will become based on the new people who are coming in if you go from 5 to 20 people and you don’t have a deliberate hiring process with the leadership principles that’s how you get people who say it’s just not like it used to be last year so those two things I would say you have to do when you’re and.
[1:01:17] You know in order to stay true to your roots the other ones you can you I would not recommend doing them all at once I think some of them are easy but sometimes our Journeys so no you don’t have to do them all at once but where we’ve seen
it work in organizations and what I guess will receive a not work as if.
That the head of the organization you know the CEO or if it’s a large company if it’s a you know Division if they’re not on board it’s probably not going to work
you know if someone says hey I’m going to write narratives for the group and then the the.
[1:01:52] PPR the CEO says yeah that’s great but just give me a PowerPoint when when you’re done with your narratives did then we’ll make the decision you know that’s it’s probably not going to work so I think you have to buy into some of these and
principles and processes and they give them a chance to work at the right level.
[1:02:11] That makes total sense Colin and that’s going to be a great place to leave it because it’s happened again we’ve used up all our allotted time.
As always if folks have comments or questions they are welcome to follow up with us on our Facebook page or on Twitter.
And as always if you enjoyed this episode we sure would appreciate it if you jump on the iTunes and give us that five star review.
[1:02:34] Can we really appreciate you taking time out of your busy schedule to walk us through the book we strongly encourage readers to go not only by the book but read it like Jason I have we both thoroughly enjoyed it you got four thumbs up from us.
Obviously you know Amazon carries the book so that’s the logical place to look and then if folks want to find you online do you pontificate about things or are you big on Twitter or SnapChat or inserting.
[1:03:02] We have website working backwards.com so I’ll one word and you know that’s a good place to go to print out from there.
[1:03:10] Awesome we really appreciate having having you on the show.
[1:03:14] Thanks again for having me.
[1:03:16] We really enjoyed it and until next time happy commercing.