Backed by Y CombinatorFour documents and three sheets that price a new engagement from what your own closed engagements actually took, not from a fresh guess.
Free download · No account needed
What you get
Pack of 7 files: 4 documents, 3 spreadsheets
Consulting Pricing Model TemplateEffort Model
| Engagement type | Bottom-up plan | Closed | P50 | P75 | P90 | Share over plan |
|---|---|---|---|---|---|---|
| [Engagement type] | [Days] | [Count] | — | — | — | — |
| [Engagement type] | [Days] | [Count] | — | — | — | — |
| [Engagement type] | [Days] | [Count] | — | — | — | — |
Percentiles are computed from actual closed-engagement days, never from a fresh re-estimate.
Every consulting pricing template in this category does the same thing: list the tasks, put a day estimate on each, add them up, quote the total as the fee. That total is somebody's honest judgment of what the work should take. It is not a record of what the same type of work has actually taken before, and the gap between the two is exactly where a fixed fee starts losing money. This pack's Effort Model runs the check most independent consultants skip: what every closed engagement of this type actually took.
HM Treasury's Green Book guidance on optimism bias states plainly that project appraisers are proven to be over-optimistic about cost, benefits and duration. Its own generic uplift tables are meant for use only in the absence of more robust primary data. A practice with several closed engagements of the same type already holds data more robust than any generic table. The Effort Model turns that history into a number: the type's own median, 75th and 90th percentile actual days, computed from what happened rather than from a fresh estimate.
Northolme Advisory, a fictional three-person operations consultancy, bids every process improvement diagnostic off the same ten-day plan. Eleven closed engagements actually ran from 9 to 22 days, 73% over that plan, so pricing all eleven at the plan would have lost $77,000 against the type. FAR 16.103(b) states that a firm-fixed-price contract belongs where a reasonable basis for firm pricing exists, and that another contract type should be considered otherwise. The Scenario Comparison decides between a fixed fee and a capped structure before the number reaches the proposal.
Effort ModelNortholme Advisory · three tracked engagement types
| Type | Plan | Closed | P50 | P75 | P90 | Max | Over plan |
|---|---|---|---|---|---|---|---|
| Process improvement diagnostic | 10 | 11 | 12 | 15 | 17 | 22 | 73% |
| Vendor selection support | 15 | 9 | 16 | 17 | 19 | 19 | 67% |
| Org design review | 20 | 7 | 23 | 27 | 32 | 35 | 71% |
All three types run past their own bottom-up plan most of the time. The diagnostic's median actual is 12 days against a 10-day plan, and its worst case on record is 22.
P75 and P90 are linear-interpolated from the ranked actuals, rounded up to a whole day before either prices anything.
Margin by TypeRetrospective: pricing every closed engagement at the plan versus at the type's own P75.
| Type | Closed | Plan price | P75 price | Margin at plan | Margin at P75 | Swing |
|---|---|---|---|---|---|---|
| Process improvement diagnostic | 11 | $22,000 | $33,000 | −$77,000 | $44,000 | $121,000 |
| Vendor selection support | 9 | $33,000 | $37,400 | −$26,400 | $13,200 | $39,600 |
| Org design review | 7 | $44,000 | $59,400 | −$63,800 | $44,000 | $107,800 |
| All three types | 27 | −$167,200 | $101,200 | $268,400 | ||
Twenty-seven closed engagements, three types, and the bottom-up plan underpriced every one of them. Pricing at each type's own P75 instead would have swung the practice from a $167,200 shortfall to a $101,200 surplus.
Scenario ComparisonPricing a new process improvement diagnostic for Bramcote Fasteners, a light manufacturer.
| Fee structure | Priced days | Fee | Margin at 12 days (P50) | Margin at 17 days (P90) | Margin at 22 days (worst case) |
|---|---|---|---|---|---|
| Fixed fee at the plan | 10 | $22,000 | −$4,400 | −$15,400 | −$26,400 |
| Fixed fee at the type's P75 | 15 | $33,000 | $6,600 | −$4,400 | −$15,400 |
| Capped time and materials at the type's P90 | up to 17 | up to $37,400 | $0 | $0 | −$11,000 |
The capped structure is the only one that does not lose money against the type's own median or its 90th-percentile case. It only loses, by $11,000, against the single worst engagement this type has ever produced.
Pricing ApproachPricing Approach
The standard task list for this type adds to a 10-day plan. Eleven closed engagements of the exact same type actually ran from 9 to 22 days, median 12. Nothing about Bramcote's brief looks easier or harder than the eleven engagements behind those numbers.
A fixed fee at the plan loses against the type's own median before anything goes wrong. A fixed fee at the type's P75 breaks even against the median but still loses against the worse cases. Time and materials capped at the type's P90 is the one structure that does not.
Decided: time and materials, billed at actual days, capped at 17 days not-to-exceed.
Every tracked engagement type's actual closed-engagement days, with the median, 75th and 90th percentile computed from history rather than a fresh guess.
A retrospective backtest across every closed engagement, comparing what pricing at the bottom-up plan actually returned against pricing at the type's own 75th percentile.
Tests a candidate fee structure against the type's own median, 90th percentile and worst-case actual days before the number ships.
The mechanism note: a task list prices the work as briefed, and a practice's own closed history prices it as it has actually gone.
The internal working note behind one specific quote, tracing the chosen fee structure back to the type's own Effort Model before the number is signed into the statement of work.
Three structures in the order to try them, and the one number from the Effort Model, how wide the type's own range runs, that actually chooses between them.
Where the day rate itself comes from. This pack holds it constant and takes it as a given input from a rate card built on the practice's own realized yield.
Open the pack and send your closed engagements by type, or download the Word documents and CSV sheets and run the percentiles yourself.
Every closed engagement of a type gets its real day count, sourced from the calendar or the invoice dates rather than remembered as typical.
Sort the actual days and take the median, 75th and 90th percentile, rounded up to a whole day before either one prices anything.
Run the candidate structure against the type's own median, 90th percentile and worst case, so the number is checked before it reaches the client.
Yes. The zip is Word documents and CSV sheets, no account and no card. Edit with AI is the optional half: send your closed engagements by type and the agent builds the Effort Model and tests a fee structure against it. The rest sit in the template library.
Four Word documents and three CSV sheets in one zip. The Effort Model, Margin by Engagement Type and Scenario Comparison sheets open in Excel, Numbers or Google Sheets with every column intact, no conversion needed.
More than three of the same type. Three or fewer is not enough to support a real percentile, and the pack says so rather than computing one anyway. Price the next one or two conservatively and log the actual days the moment each closes.
No. The day rate is an input here, not an output. The rate card and positioning pack sets it from a practice's own realized yield by engagement type, and this pack takes that rate and prices the day count around it.
Only with a specific, stated reason the scope is narrower, such as one plant instead of the two or three most closed engagements covered. Feeling straightforward going in is the exact optimism the Effort Model exists to correct.
The number goes into the proposal as the priced days, then into the statement of work as an agreed term. If the scope grows mid-engagement, the change order template prices the addition the same way.
Written by
Co-founder & CEO, River
Chandler has spent the last six years working in AI, building tools for millions of users and Fortune 100 companies. In 2025, he co-founded River to build a new kind of operating system for AI agents.
Take the Word documents and CSV sheets blank, or send River your closed engagements and get a fee structure tested against your own history.
Edit with AI