Why Metriq is estimate-free
Most planning tools begin by asking your team to guess. How many points is this? How many days? Metriq never asks. There is no points field, no sizing ceremony, no number you assign before work begins. This is not an omission we plan to fix later. It is the single most deliberate decision in the product, and understanding it is the key to understanding everything else.
The problem with guessing
Estimation rests on a quiet assumption: that a person looking at a task before it starts can predict how long it will take. Decades of delivery data say otherwise. Estimates are systematically optimistic, they vary wildly between people, and they get re-litigated in meetings that produce no working software. Worse, once a number exists, it becomes a target — and a target distorts the very behaviour you were trying to measure. People pad, they argue, they cut corners to hit the figure. The number you invented to create predictability ends up destroying it.
The deeper issue is that an estimate is a forecast made from no data. You are predicting the future from intuition. There is a better input sitting right in front of you: how long your team's work has actually taken, over and over, across dozens of items.
Counter-positioning: measure, don't guess
Metriq inverts the usual order. Instead of guessing first and measuring never, it measures continuously and forecasts from that measurement.
Every time a card moves across the board, Metriq records its cycle time — the real elapsed time from "started" to "done" — and its throughput, how much work the team actually completes. Once it has finished a body of work, you have a distribution of how your team actually behaves. That distribution is not an opinion; it is your team's recent history. From it, Metriq runs a Monte Carlo simulation to answer the questions you actually care about: When will these ten items be done? How many can we finish before the demo?
The answer comes back not as a single date — single dates are lies dressed as confidence — but as a probability. The p85 forecast, for instance, is the date by which you'll be finished in 85% of simulated futures. This is a fundamentally more honest claim than "8 points, so two days." It carries its own uncertainty in it, and it is built entirely from evidence.
This is why measured flow is more trustworthy than story points, not less. A story point is a shared hallucination — a number a room agreed to feel comfortable. A cycle-time forecast is a calculation over what really happened. One is calibrated by reality on every completed card; the other is calibrated by nothing.
The constraints are the product
It would be easy to read "no estimates" as a feature that's missing. It is the opposite. The absence is load-bearing. By refusing estimates, Metriq removes the ceremony, the gaming, and the false precision in one move — and it forces the data to do the work that guessing used to pretend to do. The same logic governs the fixed four-column board: a constraint that looks like a limitation is in fact the mechanism that keeps your flow legible and your forecasts grounded.
A tool that let you add estimates "just in case" would quietly invite the whole apparatus back: the points, the padding, the planning theatre. Metriq holds the line because the line is the value. The opinion is the product.
What this means for you in practice
- You will never be asked to size a card. Start it, finish it, and the data accumulates on its own.
- Forecasts appear once you have enough completed history for the simulation to be meaningful; until then Metriq shows nothing rather than a number built on too little data.
- Plans are expressed as probabilities and ranges, not as committed dates. You learn to read "p85 = next Thursday" the way a meteorologist reads a forecast: as a calibrated bet, not a promise.
If you've spent years in estimation meetings, this can feel like something is missing. Give it a few weeks of real cycle-time data. What's missing is the guessing. What replaces it is a forecast you can actually defend.
Next: How forecasting replaces estimation walks through the mechanics — what Monte Carlo does with your cycle-time data, and what p85 means in plain language.