Estimate the Unknown

Turn an unknown into an estimate you can defend.

Estimate the Unknown

What this fixes

You have to commit to a number - an estimate, a timeline, a sizing - and you can’t know it precisely. So someone gives a single confident figure, because a range feels like weakness. Then the plan is built on a guess that was never really defensible.

Estimate the Unknown breaks the unknown into parts you can actually reason about, estimates each as a range, and shows what’s making it uncertain - so you leave with a number you can defend and the assumptions visible.

How this works

Example: your manager asks, “How long will the billing migration take?” You genuinely don’t know. Instead of saying “17 days”, you break the work into parts, put a range on each, and say what could move the estimate up or down.

A whole unknown feels unguessable - so instead of guessing it whole, you break it down. On the Estimate Worksheet, you list the parts of the estimate, put a low–high range on each, and name the two or three things actually driving the uncertainty. Combine the ranges into one range, then say what would move you to the low end and what would push you to the high end.

The result isn’t “17 days.” It’s “15–20 days - 15 if the integration is stable, 20 if approvals take more than two rounds.” A range you can defend beats a number you can’t.

Team: each person decomposes independently first, then you compare parts, ranges and assumptions, and agree on one range together. Comparing where your estimates differ is where the real uncertainty shows up. One person acts as the facilitator, using the Facilitator Guide, while everyone takes part.

Time required

Solo ✓   Team ✓   Solo: 10–15 minutes.   Team: 30–45 minutes, with a facilitator.

The time required is indicative and will depend on how many parts the estimate has and the size of the team.

For a team run, the facilitator should read the short facilitation guide first - about 8–10 minutes. It is a downloadable PDF on this page.

When to use

Bring one real estimate you have to commit to now - the tool works on a live number, not a hypothetical.

Use it whenever:

  • you’re sizing work with real unknowns - sprint, release or roadmap planning
  • a migration, integration or discovery has to be sized before it’s understood
  • someone’s asking for “just one number” and a single figure would be a guess
  • a past estimate was wrong, and no one had written down what it assumed

What you get

  • A decomposed estimate - the parts, each with a low–high range
  • The two or three real drivers of the uncertainty, named
  • One agreed band, with what moves it to each end

You’ve done it well when: you leave with a range instead of a single guess, the assumptions are visible, and you can say what would make it land at the low end or the high end.

What you’ll notice

  • Guesswork becomes visible - everyone can see what you’re assuming, and what could change the answer
  • The uncertainty has a small number of real causes - not a fog
  • The conversation shifts from “what’s the number?” to “what’s driving it?”

Why this works

A whole unknown is unguessable, but its parts usually aren’t - breaking it down turns one impossible guess into several reasonable ones. And precision creates false confidence: people trust a single number more than a range, even though the range is the better answer. Naming the range and its drivers makes the uncertainty visible instead of hidden. (Based on the principle of decomposition in estimation and research on overconfidence.)

What you need

One real estimate. A number you have to commit to - an estimate, a timeline, a sizing - that you can’t know precisely.

The Estimate Worksheet. The core of the tool: columns for the parts, a low and high range for each, the assumption behind each (something you’re taking to be true), and a confidence note. Use it in landscape orientation, so each part has enough room for its range and assumptions. One printed sheet, or drawn on a whiteboard.

A pen or marker. Solo: a pen and your worksheet. Team: a few markers, plus one large sheet or whiteboard everyone can gather around.

Team - a facilitator. They run the flow, keep each person decomposing independently first, and drive the room to one agreed band. Everyone else takes part.

Steps

  1. Write the unknown (2 min).Team: appoint a facilitator first. Write the one number you have to commit to at the top of the worksheet - “how long to migrate the billing service?”
  2. Break it into parts (5–8 min). List the parts of the estimate, one per row - the pieces of work or phases you can reason about separately. Solo: write your own. Team: each person silently breaks the work into parts on their own worksheet first, then compares lists and agrees on one combined set of parts on the shared worksheet.
  3. Put a range on each part (5–8 min). For every part, write a low and a high - your honest best and worst realistic cases, not a single number. Team: do this on your own first; don’t discuss yet. If you’re tempted to write one exact figure, that’s usually a sign you’re hiding uncertainty, not reducing it - a wide range is information, not weakness.
  4. Name what’s driving the uncertainty (5–8 min). For the widest ranges, write what’s making them uncertain - the two or three real drivers - things that could change the estimate (an unknown dependency, a review cycle, data quality). Team: reveal everyone’s ranges together. Wherever they differ most, ask why - those differences point to the real uncertainty.
  5. Combine and state the range (5–8 min). Combine the low estimates into one lower bound, and the high estimates into one upper bound. Then write the two lines that let you justify it: what would move it to the low end, and what would push it to the high end. Team: agree on one range together.

Output / Takeaways

You leave with:

  • A decomposed estimate as one agreed range
  • The named assumptions and uncertainty drivers behind it
  • The two conditions that move it to the low or high end

Important rules

  • Always a range, never a single point - a band you can defend beats a number you can’t
  • Decompose before you estimate - never guess the whole
  • Name what’s driving the uncertainty - don’t just widen the range
  • Write the assumptions down - an estimate without them is just a guess

Planning to facilitate this activity? Access the resources below to get started.

MIndn
Execution must be smooth. Not chaotic or stressful.
Built by Witmer.