Skip to content
Log in
← Glossary

range estimate

Also called: estimate range, estimating in ranges

An estimate given as a low and a high value with a stated confidence, such as "8 to 14 days, fairly sure", instead of one number. The width of the range shows how much is still unknown.

A range estimate gives a low and a high value with a stated confidence, such as "8 to 14 working days, about 90 % sure", instead of one number. The width is information, not vagueness: a narrow range says the work is well understood, a wide one says something is still unknown and invites the question "what?".

A single number sounds more committed but carries less. "Fourteen days if nothing breaks" contains its own warning, and the bare "fourteen" drops it. People hearing one number treat it as a promise; people hearing a range plan around its top. Ranges come from three-point estimates per task, from how similar past work actually went, or from both, and the confidence should be said in words everyone reads the same way ("fairly sure", "about 9 in 10").

A good estimate message has four parts: the expected value, the date you are fairly sure of, what drives the spread, and when you will update it. "I will know in a week" is a legitimate estimate too: a date by which you will have a better one.

Ranges narrow as the unknowns get answered, and that narrowing is progress worth reporting. Re-estimate what is left rather than saying "on track". A single date becomes a commitment only after it sits near the top of the range, has a shared buffer behind it and has a named owner.

Example

Danylo first said the physics diagnostic test would take "fourteen working days, if nothing breaks". With optimistic, most likely and pessimistic values for its three tasks, the expected total is 16 days with a standard deviation of 2.236. Two standard deviations either side: 16 − 4.47 = 11.53 and 16 + 4.47 = 20.47, so about 11.5 to 20.5 working days.

Sixteen days is roughly even odds. The marketing plan builds on this date, so it should sit near the top: about 21 working days. The message to Leo: "Test expected in 16 working days; 21 we are fairly sure of; the questions are the uncertain part; I will update the range on Friday."

The same move on the chain that sets the opening date, on 30 November: expected 35 working days, two-standard-deviation range 28.85 to 41.15 days, against 43 working days left to 29 January. The top of the range still fits, with under two days to spare, which is why the plan keeps one shared buffer and watches it weekly.

Common mistakes

  • A range without a confidence. "8 to 14 days" means little until you say how sure you are of it.
  • Quoting the low end as the date. Others build on your date; give them one near the top of the range.
  • Ranges that never change. Re-estimate what is left as unknowns are answered and report the narrower range.
  • Making the range wide to be safe. A range padded to cover everything tells nobody where the uncertainty is; name what drives it.
  • Turning "most likely" into a promise. A task landing inside its range is not late; judge the plan by the shared buffer.

Learn it in the course

  • Estimating in ranges · Three-point estimates that keep the spread: the expected value and standard deviation of each task, how to add several tasks, and how to say a date with its confidence.