Single-job utilities for product work under uncertainty — estimation, prioritisation, and the conversations around them. No accounts, no tracking, no build step; your data stays in the URL.
Back-of-the-envelope arithmetic that's honest about what you don't know. Every input is a 90% range; the answer is a P10/P50/P90 distribution — with threshold probability, A/B scenarios, a driver-tree view of what moves the answer, and a read on which input is worth researching first. A cashflow mode prices NPV, IRR, payback and runway the same honest way.
Open →Weighted prioritisation scores are guesses wearing a suit. This wobbles your weights and scores over thousands of simulations and reports which ranks survive — and which are ties you shouldn't argue about.
Open →Write your roadmap as plain text; get a deck-ready now/next/later (or monthly/quarterly) graphic with snapshot-to-snapshot change tracking, SVG/PNG export, and zero box-nudging.
Open →Expected-value trees with honest uncertainty: options, chances, and payoffs as 90% ranges, Monte Carlo rollback, the recommended path highlighted — and exactly what would flip the decision.
Open →Your opportunity solution tree and your roadmap as projections of one text file — columns derived from discovery status, with audits for roadmap items that have no why and committed bets nobody has tested.
Open →Assumption maps, stakeholder grids, futures matrices, and risk registers are one physical act: items on two axes, meaning read off named zones. Write the list, drag to place, get the method's verdict.
Open →Gauge the room before the discussion anchors it. Write estimation questions, share a link, everyone answers privately — probabilities and 90% ranges — then reveal the whole spread at once. Questions never leave the link; the server sees only numbers, for 24 hours.
Open →A programme timeline that refuses false precision: every milestone is a P50–P90 range drawn as a diamond with an uncertainty whisker. Snapshot before each board pack — the compare renders the slip slide.
Open →Little's Law made visceral. Sliders for demand, team, WIP limit, and variability drive a living queue — watch cycle time explode as utilisation nears 100%. Then price your batch size on Reinertsen's U-curve, and triage a real backlog: which lever clears the pile fastest.
Open →In the workshop: a throughput forecaster ("when will it be done", honestly), a dependency-aware sequencer that prices your delivery order in cost of delay, and a calibration trainer. Same rules for all of them: one job, opinionated, no accounts.