Estimate Range With Narrowing Cost

Asked for a duration when the answer is genuinely unknown, both obvious responses fail. A fabricated number is treated as a commitment; a refusal leaves the asker unable to plan.

The three-part answer: a range with stated confidence · what drives its width, named specifically · and what it would cost to narrow it.

Why the third part matters most

It converts an estimation problem into a decision. The person asking can spend two weeks buying certainty or plan against the wide range — both legitimate, but only if they know the option exists.

An estimate offered without the option to improve it is a guess presented as a finding, and it will be treated as a commitment regardless of the caveats attached.

Implementation patterns

• Give a range, not a point, and state the confidence honestly — including the value that would not surprise you.
• Name the drivers concretely: the number of undocumented consumers, whether the data converts cleanly, whether the original expert is available, whether the coexistence period is weeks or a year. Vague uncertainty is not actionable; named uncertainty is.
• Decompose into what is understood and what is not. "The build is six weeks and the unknowns are the schedule" directs investigation at the right place and is more useful than any single number.
• Price the narrowing. Two weeks of investigation, a spike, a data profiling exercise, a conversation with the one person who knows.
• Allow for the tail. The last 2% takes most of the effort — the anomalies, the records fixed by hand, the consumer discovered after cutover — and an estimate without it is wrong by a large multiple.
• Re-estimate as information arrives, and communicate the revision rather than silently absorbing it.

Industry example

Migrations in regulated financial platforms such as Groww illustrate the pattern well: the mechanical work is estimable, and the schedule is determined by data quality, undocumented consumers and how long records with long lifecycles must coexist across the cutover. The last of these is frequently the single largest determinant and is made implicitly, which is how an eighteen-month plan becomes a four-year one.

Failure scenarios

• A point estimate, treated as a commitment.
• A refusal, which is unhelpful and reads as obstruction.
• A range with no named drivers, which cannot be acted on.
• No narrowing option offered, so the uncertainty is accepted rather than reduced.
• No allowance for the tail, producing a systematic underestimate.

Trade-offs

Ranges are less satisfying than numbers, and some stakeholders read them as evasion — particularly in organisations where confident estimates have historically been rewarded regardless of accuracy.

The counter is that the alternative is a number that is wrong, and that the narrowing option demonstrates willingness to reduce the uncertainty rather than merely to declare it. Offering to buy certainty is what distinguishes an honest range from a hedge.

Interview question

"How long will this migration take? You have no data. Give me your actual answer, and then tell me what you would do in the next two weeks."