16 · The campaign
$8,000, all or nothing
Where every dollar goes
Every line below is costed, and the arithmetic is shown rather than asserted. The two deductions come off the top, and the six lines they leave add up to the remainder exactly.
All six are development, testing and the other technical costs of building version one.
Off the top, before anything is built
- Kickstarter's fee
- $400
- Payment processing
- $264
5% of the total raised. Fixed, and charged only if the campaign funds.
3% plus 30 cents per pledge. Shown at eighty backers — the per-pledge part moves with the number of people, not the amount.
What is left to allocate
$7,336
Every line is costed, and the arithmetic is shown rather than asserted.
What is left to allocate, after both fees.
$8,000 raised, less $400 to Kickstarter and $264 to payment processing. That is the money the lines below are drawn from, and it is the number the roadmap has to fit inside.
What it is spent on
Six lines, adding up to $7,336. Each one is a real thing that has to be paid for, and the largest by far is the work itself rather than the machine it runs on.
- Building the two things still missing
- $4,000
- Moving Zwaptz onto a proper computer
- $400
- The computer it runs on
- $600
- A spare copy, somewhere safe
- $150
- The AI, while we build it and test it
- $1,200
- Trying it with real people
- $986
Real sale prices behind every valuation, and meeting places suggested from facts about actual places. This is the work itself, and the biggest line by a long way. Group trades is already in the prototype, so the same money buys two things instead of three.
Everything goes across: the records, the photographs people have uploaded, the settings. The one it runs on now is sized for a demonstration. Days one to four of delivery week.
The machine version one is put together on, tested on and opened on, for the build and the week it opens.
Building the spare copy, then proving it works by putting everything back from it. A backup nobody has ever restored from is a guess. This is the building of it, not renting space forever.
Every valuation it runs while we are building the pricing and checking it against things whose real second-hand value we already know. This is what it costs to get the number right before anybody leans on it.
A proper round of testing before it opens, with real people and real objects rather than us testing our own work, and the material that goes out with it. It ends when version one opens.
The order they get built in, and why comparable sales comes first, is on the roadmap.
The eighty-backer figure in the processing line is an assumption, not a forecast. It changes that line and nothing else: the percentage part is the same at any number of backers, and only the thirty cents a pledge moves.