1994 theory

Amazon

“Start with books — because the Internet will eventually sell everything.”

Contrarian beliefs

“What do you believe that others don’t?”

Your contrarian belief

For an online bookstore, shelf space isn’t the product — selection is. A single web-fronted operation can list every book in print, drop-shipped from existing distributors, with almost no warehouse. For this bookstore, customers will buy without seeing or touching the book — because books are commodity SKUs (ISBN-indexed, no sizing, no color), and reviews can substitute for browsing.

Common beliefs in your space

If you’re building a bookstore, your shelf space is the product. Costs come from inventory, real estate, and foot traffic. Even Barnes & Noble’s biggest superstore can carry only ~150k titles.
Customers won’t buy a tactile product they haven’t held. Returns and ‘showrooming’ make remote retail structurally unprofitable for anything other than commodity goods.

Core problem

“What single problem does the belief let you see clearly?”

Physical bookstores can stock only a tiny fraction of titles in print. Readers outside major cities can’t find obscure books. Even Barnes & Noble’s best superstore carries <200k titles vs. 3M+ in print.

Who bears the greatest cost

Book buyers, especially those in smaller markets or searching for long-tail titles; independent publishers invisible to physical retail.

What keeps it from being solved

Physical-bookstore economics make stocking the long tail unprofitable, so the only existing answer is mail-order catalogs that can't search or merchandise. Distributors won't disintermediate themselves for a startup. Consumers are still nervous about credit cards online, so even if the storefront existed, buyers wouldn't trust it without strong social proof.

Subproblems

“If the core problem were solved, what smaller challenges must be addressed first?”

Breadth of selection — list every book in print
Make true
Why hard to solve — Listing 3M+ titles requires data we don't have, distributor relationships we haven't earned, and an ops process that doesn't fall over when someone orders a book we've never stocked.
What it enables if solved — Once breadth works, every customer eventually finds a book they can't get anywhere else — and starts coming to us first by default.
Consumers will buy a book they haven't held in their hands
Verify
Why not yet proven — Every existing data point on remote book sales (mail-order catalogs) is heavily skewed toward people who can't get to a bookstore. We need to know whether mainstream readers will do this when bookstores are still convenient.
What it enables if confirmed — If true, books are the right wedge category and the long-tail thesis holds. If false, online retail is a discount play and the flywheel argument falls apart.
Shipping economics that don't destroy per-book margin
Make true
Why hard to solve — Carriers price aggressively only at scale we don't have yet. Drop-ship from distributors is fragile. Building our own fulfillment too early sinks capital we need elsewhere.
What it enables if solved — Sustainable per-order economics let us offer prices below physical retail — which becomes the wedge for everything else.
Buyers will trust a brand-new online retailer with credit cards
Verify
Why not yet proven — In 1994 'will my card get stolen' is a real fear, not a tail risk. There's no third-party trust mark customers recognize yet, and we have no track record to point at.
What it enables if confirmed — Once trust is established, every category we eventually expand into inherits it — and the funnel stops collapsing at checkout.

Your Theory

“How do your inputs become value?”

Wordsmithed

Bookstores are fundamentally constrained by shelf space — even the biggest superstore can carry maybe 150,000 titles out of three million in print. The internet erases that constraint: a single web-fronted store can list every book ever published, drop-shipped from existing distributors, with almost no warehouse. Lower prices and broader selection breed customer loyalty, loyalty breeds volume, and volume gives us better vendor terms — a flywheel that gets cheaper to run as it gets bigger. Once it's spinning on books, the same flywheel extends to music, then DVDs, then everything.

If-then

If internet adoption continues to grow at 20%+ annually, and consumers will adopt credit-card-online payment as standard, and distributor (Ingram/Baker & Taylor) economics hold so we can drop-ship most titles, and customers will trust a brand-new online retailer with enough social proof — and we can solve breadth of selection, shipping economics, customer trust, and discovery — then an online bookstore with effectively infinite inventory can offer broader selection at lower prices than physical retail, building a customer-loyalty flywheel (more volume → better vendor terms → lower prices → more loyalty) that extends from books into every other category until Amazon is a retail operating system.

Actions

“What will you do — to test, to acquire, to find?”

Run experiments

Experiment 01
done
Buyers will trust a brand-new online retailer with credit cards
Test — Launch the site in beta to 300 friends. First external order: a copy of 'Fluid Concepts & Creative Analogies' (April 1995). See whether real customers will actually put a credit card online with a brand-new retailer.
Success — Successful fulfillment + payment with no fraud.
Result — Worked. First week: $12k in orders. Second week: $15k. The team couldn't fulfill fast enough — problem of over-demand, not under-demand.
Experiment 02
done
Consumers will buy a book they haven't held in their hands
Test — Allow user-submitted reviews on all products, including 1-star reviews. Measure conversion vs. review-free listings.
Success — Conversion higher on reviewed listings.
Result — Reviews dramatically lifted conversion. Trust came from honest reviews, not curated marketing — a non-obvious result that publishers initially hated.

Now build your theory.

Start →