teardown 02ubercab, 2008
UberCab’s 2008 deck: twenty-five slides, no business model, and it still raised
UberCab’s 2008 deck runs to twenty-five slides with no business model, no team slide and no ask. It broke every rule the advice gives, in the same year Airbnb’s eleven-slide deck kept all of them. Both raised. What the two share is specificity, and the reason Uber’s ordering survived is that it was read in a room.
What the deck is
UberCab, late 2008. Garrett Camp published it on Medium on the company’s ninth anniversary, so this is the founders’ own copy rather than a leak. It sits behind roughly USD 200,000 of first funding. The service it describes is invitation-only, aimed at professionals in San Francisco and New York, running Mercedes sedans, and compared in the deck to NetJets.
Twenty-five slides cover the problem, the concept, differentiators, operating principles, the app, the website, use cases, environmental benefits, fleet models with fuel-efficiency figures, the initial service area, demand forecasting, market data, target cities, potential outcomes, US smartphone charts from August 2008, marketing ideas, and progress to date. There is no business model slide, no team slide and no deal terms.
Three things it does well
It reads as a build document, not a sales document
Car models with fuel-efficiency figures. The launch service area. Demand forecasting driven by day, time, weather and traffic. Smartphone penetration charts dated to a single month. None of it is aspirational, and that is the whole effect: it reads as a document written by someone who had already decided to build the thing and was working out how.
Specificity is the one quality a template cannot supply and a generic draft cannot fake. It is also the quality that survives being wrong. A named wrong number invites a correction; a vague right one invites nothing.
The outcomes slide carries the bad case as well as the good one
Best case, market leader with over a billion dollars of annual revenue. Worst case, a small executive car service that never leaves San Francisco. Almost nobody writes the second one down, and writing it down buys more credibility than the first one does: it tells the reader you have already thought about the version where this stays small, so they do not have to raise it themselves.
The last slide is progress to date
After twenty-four slides of intent, the document ends on what already exists. That is the right place for it in a room, where the last thing said is the thing remembered.
Three things that would fail today
The market arrives at slide seventeen
Sixteen slides of product before anyone is told how big the opportunity is. In a room that ordering survives, because you are sitting there to answer the question the moment it is asked. Emailed to an associate on a Friday, the deck is closed long before slide seventeen.
This is the whole difference between a document that is presented and a document that is sent, and it is the question we ask second in every brief.
No ask, no model, no team
In 2008 a seed round was often a conversation between people who already knew each other, and those three could live outside the deck. A 2026 seed deck arriving without them reads as unfinished, and the reply is a template.
Twenty-five slides that need a narrator
Several slides carry material that only works while somebody talks over it. Sent cold, a slide with nine bullets is not read: it is scanned for the one line that decides whether to keep going.
Three slides rebuilt
None of the original artwork appears here. The schematic view is our own diagram of what each slide carried; the rebuild is our work. Toggle between them.
Published decks only. We reproduce none of the original artwork: the “before” panels are reconstructions we draw ourselves.
01 · The problem, split by who has it
Same facts, one structural decision. Splitting the column shows the two-sided market without a sentence explaining it.
02 · The outcome range
The bravest slide in the deck, given the size it deserves. Naming the downside is what makes the upside readable.
03 · The running order
An appendix is not a demotion. It is where the thirteen slides that answer questions go, so the twelve that raise them can be read in four minutes.
The law it proves
A deck read in a room and a deck read alone are two different documents.enzwa · craft law 02
The room version can front-load the product, because a question gets answered the moment it forms. The sent version has to answer questions in the order a stranger asks them, and the first one is always some form of “how big is this, and why you”. Same material, different sequence, and usually a different length.
Most decks are built once and used both ways, which is why so many of them work in the meeting and go quiet afterwards. It is the second question in our brief for that reason.
Read alongside the Airbnb teardown: two decks from the same year, breaking opposite rules, both funded. The note on what that leaves is the shorter version of the argument.
Sources
Every teardown uses a publicly published deck. Schematics and rebuilds are our own work.