Weaving a Rope of Smaller Advantages
Compound small edges into a moat by making every business feed the others.
- Difficulty
- Advanced
- Time to result
- ~ongoing to results
- Steps
- 7
- Confidence
- 80%
Brad Stone's reading of post-2013 Amazon is that the flywheel story was replaced by something less legible: a set of interlocking, self-reinforcing businesses. The mechanism is deliberate consumption. Prime's two-day shipping stopped differentiating once fulfilment centres ringed every major city, so Prime had to become a content club, which is why video mattered. Prime Video and Alexa are both consumer applications of AWS, so they load the cloud business. Retail is AWS's first, largest and most demanding customer, which gives AWS scale, a beta tester, and economies retail then buys back cheaply. Advertising sits on top of retail traffic and drives it back. No single link is a moat; the rope of them is. Bezos framed it to Tim O'Reilly as weaving a rope of smaller advantages, and pushed executives to ask what they were doing for the other businesses.
Origin
Described by Brad Stone on Acquired as the central insight of Amazon Unbound, replacing the flywheel diagram of The Everything Store. He traces it to a Bezos line quoted in the earlier book, and says the realisation crystallised while writing the Prime Video chapter.
Core principles
- 01Almost no company gets one giant advantage; most get a pile of small ones.
- 02A small advantage only compounds when another business unit consumes it.
- 03The shared customer, not the shared market, is what links unrelated businesses.
- 04Today's differentiator decays; plan the replacement before it does.
- 05Interlocks have to be deliberately built, not hoped for after the fact.
How to run it
- 1
Inventory the advantages you already own
List the infrastructure, data, distribution, customer relationships and cost positions the business already has. This is the raw material; nothing gets woven that is not on the list.
Pro tip Include unglamorous assets — traffic, warehouse footprint, servers already paid for.
- 2
Date the expiry of your current differentiator
Identify what customers currently choose you for and estimate when competitors will match it. Amazon's shipping speed stopped being a differentiator once fulfilment centres sat outside every major American city.
Watch out Teams defend the decaying differentiator hardest, precisely because it built the business.
- 3
Re-specify the customer bundle around the next differentiator
Decide what the relationship has to become when the old edge is gone. Prime moved from a two-day shipping club to a one-day shipping club to a content club.
Pro tip Change the bundle before the old benefit is commoditised, not after.
- 4
Build each new business on infrastructure you already run
New bets should be consumer or commercial applications of existing platforms. Prime Video streams off Amazon servers; Alexa was pitched as a cheap device whose brains sit in the cloud.
Pro tip If the new bet needs an entirely new stack, the interlock is theoretical.
- 5
Make an internal unit the first and largest customer
An internal anchor customer gives the new business early scale, real requirements and a permanent beta tester, and gives the internal unit near-wholesale pricing and priority at peak.
Pro tip Price internal supply so the link is actually used rather than routed around.
Watch out An internal customer that cannot switch away will tolerate a bad product; keep an external benchmark.
- 6
Make the cross-business ask explicit and recurring
Ask every leader what they are doing for the cloud business, for the voice business, for advertising. The interlocks appear because someone is repeatedly asked for them.
Pro tip Put the question in the operating review, not in a strategy offsite.
- 7
Audit the rope, not the strands
Review the portfolio for units that have drifted into standalone P&Ls with no link to anything else. A collection of businesses sending cash to head office is a conglomerate, not a rope.
Pro tip Draw the map as edges between businesses; a unit with no edges is a candidate for exit.
Watch out Dense, opaque interlocks are also the thing regulators will ask you to unpick.
In the wild
Employees, board members and investors read Bezos's Hollywood push as a personal indulgence. In Stone's account it was structural: once fulfilment centres sat outside every major city, shipping stopped differentiating Prime, so Prime had to become a content club. Video then reinforced retail through Prime, ran on Amazon's own servers as a consumer application of AWS, and later fed an ad-supported streaming tier and the advertising business.
→ One investment simultaneously defends the subscription, loads the cloud business, and feeds advertising.
Amazon retail was AWS's first and biggest customer. That gave the cloud business immediate scale, economies of scale, and a beta tester for every new service, while retail got wholesale-style internal pricing and the best cloud provider behind it during pandemic and holiday traffic peaks.
→ Both units get an advantage neither could have bought on the open market.
Common mistakes
Buying adjacency instead of building interlock
Expanding into a near-adjacent market feels safe but adds no new strand if the new business shares no infrastructure or customer relationship with the old one.
Letting the internal customer opt out
If the internal unit is allowed to buy elsewhere on price, the new business loses its anchor demand and the link never forms.
Mistaking a conglomerate for a rope
If the units only send capital to head office, you own a holding company. The point is that businesses consume each other's outputs.
Is it for you?
Best for
Operators running several business lines who need each one to make the others harder to attack.
Not ideal for
Single-product startups still searching for one advantage worth having.
From the transcript
“we don't have many big advantages, so we have to weave a rope of smaller advantages”
“And so you've got this set of really opaque and hidden connections between all these businesses that is at the center of how Amazon operates.”
“And Prime would have to be something more. It would have to be a content club.”
From the episode
Special: Amazon Unbound (with Brad Stone)