Beachhead Category Selection
Pick the one category that lets you earn the right to sell everything.
- Difficulty
- Moderate
- Time to result
- ~weeks to results
- Steps
- 7
- Confidence
- 87%
Bezos and David Shaw agreed early that the everything store could not be launched as the everything store: you pick one category, build the consumer brand, the website and the traffic, and add categories on top. Bezos then ran a structured comparison rather than choosing by affinity. Books scored well on five dimensions at once — they are perfect commodities, so a paperback is a paperback wherever bought; only two real distributors held US inventory, so supply was a pair of accounts; there were roughly 4,200 publishers, so no gatekeeper could cut him off; three million titles were in print against the roughly 80,000 a superstore stocked, so the long tail was enormous; and neither Barnes & Noble nor Borders held even 12% share. Music lost on producer concentration: six labels.
Origin
Reconstructed on the Acquired episode on Amazon.com from Bezos's category research at D. E. Shaw in 1993-94, when he compared books against music and other candidates before choosing books.
Core principles
- 01You cannot launch the everything store; you earn it one category at a time.
- 02Commodity products remove the need to prove your version is the good one.
- 03Concentrated distribution makes supply a two-phone-call problem.
- 04Fragmented producers mean no single gatekeeper can shut you out.
- 05A long tail creates a reason to use you that incumbents structurally cannot match.
- 06A fragmented incumbent field means you fight for share, not for survival.
How to run it
- 1
Write the horizontal ambition first
State the end-state business plainly. The entry category is chosen to serve it, so the ambition has to exist before the choice is made.
Pro tip Amazon's name was chosen for the ambition, not the category: earth's largest river, earth's largest selection, A to Z.
- 2
Score commoditisation
Prefer categories where the product is identical regardless of source. Commodity products remove any need to convince customers your version is good.
Watch out Non-commodity categories force you to build trust in the item and the seller at the same time.
- 3
Score distributor concentration
Few distributors means you can access the whole market's inventory with a handful of accounts. Books had Ingram and Baker & Taylor.
Pro tip Physical proximity to a major distributor is a real operational advantage; Ingram's Oregon warehouse influenced the choice of Seattle.
- 4
Score producer fragmentation
Many small producers means no single party can decide you do not exist. Books had roughly 4,200 publishers; music had six labels, which is why music lost.
Watch out Concentrated producers can and will refuse to supply a new entrant that threatens their existing channel.
- 5
Score catalogue depth
Compare the total catalogue with what an incumbent can physically stock. The gap is your only structurally defensible reason to exist on day one.
Pro tip Three million titles versus eighty thousand on a superstore shelf is the shape of gap you are looking for.
- 6
Check incumbent concentration
Confirm no player holds dominant share. Barnes & Noble and Borders each had under 12%, so reaching their scale was a share fight rather than a war of survival.
Watch out Being right about the category still invites a fight; Barnes & Noble sued Amazon three days before its IPO priced.
- 7
Confirm the assets transfer
Verify that the brand, traffic and operational muscle built in the entry category actually carry into the adjacent ones you plan to add.
Pro tip Amazon's ISBN-driven authoritative catalogue became the ASIN system for every product it ever sold.
In the wild
CDs looked at least as good as books on shipping weight, standardised packaging and commoditisation — arguably better on several dimensions. They lost on one criterion that turned out to be decisive: six record labels controlled the catalogue, so a new entrant would need all six to agree. Books had roughly 4,200 publishers and two distributors, meaning the entire commercial catalogue was reachable through two accounts and no single party could block entry.
→ Amazon launched with books and used the resulting brand and traffic to add music, DVDs, software, toys and electronics within a few years.
A founder wanting to build a vertical marketplace for a broad trade sector scores candidate entry categories on the same five axes and finds that the obvious high-revenue category is controlled by three manufacturers, while a smaller adjacent category has hundreds of makers, two distributors and a catalogue far deeper than any physical stockist carries.
→ The smaller category becomes the beachhead precisely because supply cannot be shut off.
Common mistakes
Choosing the biggest category first
Market size dominates the analysis and hides the supply-access questions that actually determine whether a new entrant can operate at all.
Ignoring producer concentration
A category with a handful of suppliers gives each of them a veto over your existence, which is why music lost to books despite better shipping economics.
Picking a category whose assets do not transfer
If the brand, traffic and operations you build do not extend to the next category, you have chosen a business rather than a beachhead.
Is it for you?
Best for
Marketplace and retail founders whose end-state is many categories but who must start with one.
Not ideal for
Single-product businesses with no expansion ambition, where category choice is the whole strategy rather than an entry decision.
From the transcript
“you can't just start with the everything store you need to pick one category build that build the consumer brand”
“books are the ideal category for a few reasons one they are perfect Commodities”
“with that huge diversity of products 3 million books in print you could build a store online that simply could not exist in any other…”
From the episode
Amazon.com