← Back to blog

Same Deal, Two Estimates — Which Number Is Right?

Two senior consultants. Same RFP, same client, same modules. One comes back at 2,400 hours, the other at 3,100. Both have run this kind of project before. Both can defend their number for an hour without repeating themselves.

Which one is right?

In most firms, that question has no answer — not because nobody is smart enough, but because there is nothing to check either number against. Each estimate is a spreadsheet built from one person's judgment, and the judgment isn't in the spreadsheet. The spreadsheet holds the outputs. The reasoning — what scope was assumed in or out, which risks got padded, why testing is 15% on one tab and 25% on the other — lives in the head of whoever built it.

The three ways firms usually resolve this

We've seen all of these. We've done some of them.

• Average the numbers. Now you have a third number nobody believes, built from two sets of assumptions that may directly contradict each other.
• Go with the more senior voice. Sometimes correct. But it settles the argument by rank, and the other estimator's concerns — which might be the ones that materialize in month four — go unrecorded.
• Take the lower number, because it wins the deal. The most expensive option. The gap between the two estimates was a risk signal, and it just got priced at zero.

None of these answers the actual question, because the actual question isn't "which total is right." It's "where do these two numbers disagree, and why?"

Disagreement is signal — if you can see it

Here's the thing practitioners know: when two experienced people scope the same work differently, the difference is information. One of them is assuming the client's data is clean. One of them has been burned by that assumption before. One prices a phase for a strong client team; the other has met the client team.

That disagreement, surfaced at the assumption level, is one of the most useful conversations a firm can have before signing. Buried inside two incompatible spreadsheets, it isn't a conversation at all — it's an argument about totals, and the loudest total wins.

The fix isn't better spreadsheet discipline. It's a shared structure both estimates are built against: the same scope questions, the same role and phase model, so that when two people diverge, you can point at the exact line where they diverged and ask why. Then the senior conversation happens where it should — on the assumption, not on the sum.

What this looks like in practice

This is the problem we built JungleOS Scope around. Scoping runs as a structured set of questions instead of a blank workbook, and the output is hours by role and phase that trace back to the answers given. When an estimator adjusts a number — and good estimators always adjust — the adjustment is recorded with the reasoning attached, under their name, instead of vanishing into a cell edit.

Two estimates of the same deal stop being two spreadsheets that can't be compared. They become two sets of answers to the same questions — and the differences stand out on their own.

The judgment still belongs to your people. That's the point: it's their judgment, made visible enough that the firm can actually use it — compare it, question it, and stand behind the number that comes out.