This is where the longer version of things lives: the reasoning behind a confidence tier, why we priced overdue work the way we did, what changed and why. If you want the short version, the product itself is the short version.
Blog
Notes from building CostFlow.
Writing about the problem we're solving, how the pricing methodology actually works, and what we get wrong along the way.
Planned, not yet written
Why every friction number is a range, not a point estimate.
The actual reasoning behind expected-value-with-a-range instead of a single confident-looking number.
What we mean by 'unpriced,' and why we'd rather show nothing than guess.
The confidence-tier system and the decision to leave things unpriced instead of inventing an assumption.
The first version of CostFlow's cost model was wrong. Here's what changed.
An honest post-mortem once there's something real to write about.
Reading a Jira board like a queueing problem.
The mental model behind treating 'waiting' and 'stalled' as distinct, priceable states.
Nothing published yet
Check back soon, or see the product in the meantime.