An Estimate Is a Promise About Scope
Clients hear an estimate as a price. Firms mean it as a price for an assumed amount of work. The gap between those two readings is where overruns live.
- Entry
- 03 of 07
- Section
- Before
- Kind
- Analysis
An estimate is a figure attached to an assumption about how much work there will be. The firm means the second part; the client hears the first.
The commercial boundary in “An Estimate Is a Promise About Scope” becomes more useful when estimated work, recorded hours and authorised scope can be compared without changing the signed terms. Teams researching apps for tracking work hours can use check the current feature details to organise that operational evidence while keeping the engagement letter and approved change record as the contractual source of truth.
Almost every overrun dispute is that gap, and it is closed by writing the assumption down in the same sentence as the number.
Additional context for “An Estimate Is a Promise About Scope” is available from NLRB employee-rights guidance. Confirm current legal, tax and accounting requirements with a qualified adviser where necessary.
What the firm means
"If the matter runs as we expect — one round of comments, no third party, no change of structure — it will cost about this."
- One round of comments, not three.
- No third party joining the matter.
- No change to the structure or the parties.
- No dispute arising out of any of it.
The conditions are real, they were in somebody's head, and they usually did not make it onto the page.
What the client hears
"About this." The conditions, if stated at all, were in a paragraph below the figure and were read once.
Clients are not careless. A figure is concrete and a condition is not, and memory keeps the concrete thing.
Writing the assumption into the figure
"£14,000, assuming one round of comments and no change to the structure" is a different sentence from "£14,000 (assumptions below)".
The first survives being remembered. The second does not.
Stating what is not included
Shorter and more effective than stating what is. Three bullets of exclusions are read; a page of inclusions is not.
Exclusions also tell the client what to come back about, which is the point.
The estimate that was never revised
A matter changes shape in month two and the estimate from month one is still the number in everybody's head.
Revising it is a ten-minute email and it is the single most effective thing a firm can do about overrun disputes.
Why revisions do not happen
Because they feel like asking for more money, and because the person who can see the overrun coming is often junior.
Both are solvable: make the revision a standing step at a set point, not a judgement about when it is bad enough to mention.
The trigger worth setting
At seventy per cent of the estimate, somebody checks whether the remaining work fits in the remaining thirty.
It costs five minutes, it catches overruns while they are still decisions, and it can be delegated to a system rather than a person.
What happens with no trigger
The overrun is discovered when the bill is drafted, at which point the options are to absorb it, to bill it and argue, or to bill part of it and absorb the rest quietly.
All three are worse than a call six weeks earlier.
Estimates and fixed fees
An estimate is not a cap and not a fixed fee, and clients frequently treat all three as the same thing.
Saying which one is being given, in the words the client will remember, avoids a dispute whose subject is a definition.
The range
"Between twelve and eighteen thousand" is more honest than fourteen and is heard as eighteen by the firm and twelve by the client.
If a range is given, the top of it should be the number repeated in every subsequent conversation.
Recording what was said
Estimates are given on calls. Two months later nobody agrees what was said, and the firm's note is in somebody's head.
One email confirming the figure and the assumption makes the estimate a document rather than a memory.
Estimating from data rather than memory
Recorded hours on comparable past matters, by stage. Most firms hold several years of it and almost none extract it.
Memory estimates the shape of a matter and forgets its weight: the procedural steps, the chasing, the third draft.
A firm that pulls the hours on its last ten matters of a type produces a better figure in an hour than any amount of judgement, and the figure is defensible.
Why the data is not used
It is awkward to extract, matter types are not coded consistently, and the person giving the estimate is usually doing it on a call.
All three are solvable, and the first firm to solve them in a practice area finds its estimating improves permanently rather than case-by-case.
The estimate that is a negotiation
Sometimes a figure is given because it is the figure that wins the work, and everybody knows the matter will cost more.
That is a commercial decision and it is legitimate, provided the firm does not then treat the resulting overrun as a client problem.
The dishonest version is giving a low figure, saying nothing, and billing the true cost with an explanation at the end. It works once.
What a client should be told
The figure, the assumption, what would change it, and that they will be told before it does.
Four facts, one short paragraph, and they convert an estimate from a hostage into a working instrument.
The connection to the bill
Every overrun dispute is an argument about an estimate conducted through a bill. Fixing the estimate removes the dispute; nothing on the bill can.