With all due respect to the author, I think the article kind of romanticizes start-up founders and their perils.
Amazon famously dubs itself with being the biggest startup in the world. There's a company that espouses Clay Christensen's philosophy of building enduring businesses. Over the years, a lot has been said and written about how Amazon manages to do it, I mean, this is better demonstrated by Jeff, its founder, whose wealth went from $10B in 1997 to $1B in 2002/3, but still bet big on AWS (2003), Prime (2005), and Kindle (2007) in an unprecedented run of series of innovations that'd could have killed the company.
--
"working backwards"
I want to draw some parallels to the YC application process with how Amazon operates, from what I've read and what I've experienced as an ex-employee:
A lot of product development is funded at Amazon after review of what's called a PR/FAQ doc [0] (this is the "working backwards from the customer" part). The PR needs to clearly articulate in a headline or two what the product is about; whilst the first paragraph must present a complete summary of what the product would be in its v1 form at launch. The next few paragraphs detail the current problem and the proposed solution interlaced by imaginary quotes from would-be customers; and the concluding paragraph has a clear call-to-action on exactly how customers can make use of the proposed solution.
If you've ever filled out a YC form, you'd find yourself going through a similar exercise.
And on what merits is a product funded [1]?
- If it works, would it be really big?
- Is the customer / target market well-served today?
- What in Amazon's approach is the key differentiator? And is that compelling enough?
- Should / can Amazon build it in-house, or do they need to buy some / all of the expertise?
Again, pretty similar to the process YC has [2].
--
"two-pizza teams" [3]
The team that's put together to run is encouraged to own the product end-to-end ("single-threaded owners" aka STOs), in the truest sense of the word: That is they're free to duplicate effort, not be beholden to another team's priorities, build whatever they need to, buy whatever they need to, and so on... Other STOs running existing but overlapping business or businesses at the risk of being cannibalized by this newer one do not absolutely get any say. This approach to incubating newer products within Amazon is what led them to build AWS in the first place, because they didn't want various internal engineering teams to be truly duplicating their efforts in building "undifferentiated parts of their businesses" which, on the Internet, is building all that Infrastructure required to start small and yet be able to scale. AWS, interestingly, itself was removed / isolated from Amazon's Infrastructure team at the time and was completely a separate under-taking (there's probably a Harvard case-study in there somewhere demonstrating the effectiveness of STOs).
--
"disrupt yourself"
The other thing Amazon does is it is truly customer-focused as opposed to product-focused or competitor-focused. If I were to take the example of Android: How many customers do Google have a direct line to? If you are a FireOS user, you could chat with customer service about its annoyances, send an email to kindle-feedback@ or even escalate it to jeff@ and all those complaints are root-caused and fixed to whatever extent deemed necessary, with FireOS MayDay being an extreme example of this customer-obsession. Amazon believes in listening to its customers and disrupting its own cash-cows if it means it delivers value to the customer. No one flinches a bit in taking these decisions.
--
"you can't fight gravity" [4]
As opposed to reacting technology changes constantly and riding the wave, Bezos instead believed in focusing on universal constants (like gravity) that never change: For example, for Amazon's e-commerce business, those constants are customers would always want lower prices, larger selection, and faster deliveries". That was never going to change. But this simple framework then lets his management team decide on what bets to take with respect to technologies that help move the needle in the right direction, because if they don't, someone else will eat their lunch by doing those three things.
--
"the best way to predict the future is to invent it"
To truly create an atmosphere of invention within Amazon, there are a lot of processes in-place, to make sure bureaucracy ("a single no" vs "a lot of yes") doesn't kill a promising idea. Of course, there's nuisance here, in that some decisions need to be carefully vetted ("one-way doors") vs ones that needn't be ("two-way doors") and the key is knowing which is which (and escalate when in doubt).
--
"simplify"
If you follow AWS, you'd know how primitive and lacking the v1s really are: For instance, Lambda launched with just NodeJS support with no observability story of note, no "local development" environment, no support for other runtimes, and just 1 minute of execution time. This stems from the PR/FAQ process (distill down the v1 to the absolute minimum but deliver comprehensive value to the under-served) and two-pizza teams (too many resources to work on a problem is never approved of, so it is paramount to do things that don't scale for the v1).
--
This isn't to say Amazon hasn't been disrupted at all: It has been, by instacart, daipers.com, doordash among some examples that come to mind.
I am no expert (either in the ways of the likes of Amazon or the nimble start-ups), but I believe that to reduce innovation / invention being a playground for start-ups just because they've no access to a "fancy cafeteria" is telling half the story [5] and probably misinterpreting symptoms for cause.
[2] YC also fund all sorts of "uninteresting" ideas too (from outside, what looks like a spread and pray, but is likely a heuristic that they are working off of from).
[5] I mean, at the end of the day, Waymo wasn't even a startup by the time Google acquired it and also it isn't like Google doesn't have a track record of successful acquisitions...
Amazon famously dubs itself with being the biggest startup in the world. There's a company that espouses Clay Christensen's philosophy of building enduring businesses. Over the years, a lot has been said and written about how Amazon manages to do it, I mean, this is better demonstrated by Jeff, its founder, whose wealth went from $10B in 1997 to $1B in 2002/3, but still bet big on AWS (2003), Prime (2005), and Kindle (2007) in an unprecedented run of series of innovations that'd could have killed the company.
--
"working backwards"
I want to draw some parallels to the YC application process with how Amazon operates, from what I've read and what I've experienced as an ex-employee:
A lot of product development is funded at Amazon after review of what's called a PR/FAQ doc [0] (this is the "working backwards from the customer" part). The PR needs to clearly articulate in a headline or two what the product is about; whilst the first paragraph must present a complete summary of what the product would be in its v1 form at launch. The next few paragraphs detail the current problem and the proposed solution interlaced by imaginary quotes from would-be customers; and the concluding paragraph has a clear call-to-action on exactly how customers can make use of the proposed solution.
If you've ever filled out a YC form, you'd find yourself going through a similar exercise.
And on what merits is a product funded [1]?
- If it works, would it be really big?
- Is the customer / target market well-served today?
- What in Amazon's approach is the key differentiator? And is that compelling enough?
- Should / can Amazon build it in-house, or do they need to buy some / all of the expertise?
Again, pretty similar to the process YC has [2].
--
"two-pizza teams" [3]
The team that's put together to run is encouraged to own the product end-to-end ("single-threaded owners" aka STOs), in the truest sense of the word: That is they're free to duplicate effort, not be beholden to another team's priorities, build whatever they need to, buy whatever they need to, and so on... Other STOs running existing but overlapping business or businesses at the risk of being cannibalized by this newer one do not absolutely get any say. This approach to incubating newer products within Amazon is what led them to build AWS in the first place, because they didn't want various internal engineering teams to be truly duplicating their efforts in building "undifferentiated parts of their businesses" which, on the Internet, is building all that Infrastructure required to start small and yet be able to scale. AWS, interestingly, itself was removed / isolated from Amazon's Infrastructure team at the time and was completely a separate under-taking (there's probably a Harvard case-study in there somewhere demonstrating the effectiveness of STOs).
--
"disrupt yourself"
The other thing Amazon does is it is truly customer-focused as opposed to product-focused or competitor-focused. If I were to take the example of Android: How many customers do Google have a direct line to? If you are a FireOS user, you could chat with customer service about its annoyances, send an email to kindle-feedback@ or even escalate it to jeff@ and all those complaints are root-caused and fixed to whatever extent deemed necessary, with FireOS MayDay being an extreme example of this customer-obsession. Amazon believes in listening to its customers and disrupting its own cash-cows if it means it delivers value to the customer. No one flinches a bit in taking these decisions.
--
"you can't fight gravity" [4]
As opposed to reacting technology changes constantly and riding the wave, Bezos instead believed in focusing on universal constants (like gravity) that never change: For example, for Amazon's e-commerce business, those constants are customers would always want lower prices, larger selection, and faster deliveries". That was never going to change. But this simple framework then lets his management team decide on what bets to take with respect to technologies that help move the needle in the right direction, because if they don't, someone else will eat their lunch by doing those three things.
--
"the best way to predict the future is to invent it"
To truly create an atmosphere of invention within Amazon, there are a lot of processes in-place, to make sure bureaucracy ("a single no" vs "a lot of yes") doesn't kill a promising idea. Of course, there's nuisance here, in that some decisions need to be carefully vetted ("one-way doors") vs ones that needn't be ("two-way doors") and the key is knowing which is which (and escalate when in doubt).
--
"simplify"
If you follow AWS, you'd know how primitive and lacking the v1s really are: For instance, Lambda launched with just NodeJS support with no observability story of note, no "local development" environment, no support for other runtimes, and just 1 minute of execution time. This stems from the PR/FAQ process (distill down the v1 to the absolute minimum but deliver comprehensive value to the under-served) and two-pizza teams (too many resources to work on a problem is never approved of, so it is paramount to do things that don't scale for the v1).
--
This isn't to say Amazon hasn't been disrupted at all: It has been, by instacart, daipers.com, doordash among some examples that come to mind.
I am no expert (either in the ways of the likes of Amazon or the nimble start-ups), but I believe that to reduce innovation / invention being a playground for start-ups just because they've no access to a "fancy cafeteria" is telling half the story [5] and probably misinterpreting symptoms for cause.
[0] https://www.youtube.com/watch?v=aFdpBqmDpzM
[1] https://www.hbs.edu/forum-for-growth-and-innovation/podcasts...
[2] YC also fund all sorts of "uninteresting" ideas too (from outside, what looks like a spread and pray, but is likely a heuristic that they are working off of from).
[3] https://www.youtube.com/watch?v=XavPl5t9dS8
[4] https://www.youtube.com/watch?v=O4MtQGRIIuA
[5] I mean, at the end of the day, Waymo wasn't even a startup by the time Google acquired it and also it isn't like Google doesn't have a track record of successful acquisitions...