Fee Estimator

Four methods, run side by side on the same project, so the spread between them is visible. The arithmetic is the same everywhere. Every rate that goes into it is yours.

Workflow

Step 1. Bootstrap if needed

python scripts/init_workspace.py

Step 2. Establish the basis

Ask, in one message:

1. Project type, size, and construction cost, with what the cost is based on.
2. Delivery method and the services in scope. Basic services against additional services changes every method below.
3. Is this governed by a procurement rule that caps or prescribes the fee? If so, name the rule and its cap, and I will apply it and cite you for it.
4. Where do your rates come from? Past projects are best; give me two or three comparable ones with their fee and cost, and I will derive the percentage rather than assume one.
5. Which methods do you want run? Any I have no inputs for will be reported as not run.

Step 3. Collect the inputs, per method

Ask for each, and never substitute a typical value.

Method	Needs
Percentage of construction cost	the percentage, and where it came from
By area	the rate per unit area, and the area basis
Hourly against a cap	staff rates by grade, hours by grade, the multiplier if you bill that way, and the cap
Lump sum	the build-up you are proposing, or the figure and its basis

Then the splits, each of which is also yours:

• Phase split, as percentages that sum to a hundred, with the phases named.
• Consultant shares by discipline, and whether they come out of the total or sit alongside it. This distinction moves the number by more than most rate choices, so ask it explicitly.
• Reimbursables, as a percentage or a figure, and what it covers.

Step 4. Run each method, showing the arithmetic

Method: percentage of construction cost
  construction cost        <as given>
  percentage               <as given>   [from: <your source>]
  fee                      <arithmetic>

Repeat per method. A method with a missing input prints not run and names what it needs. Never carry a figure from one method into another as a substitute.

If a procurement cap was named, apply it last, show the comparison, and say which figure governs.

Step 5. Compare and split

Show the four figures together with their spread. A wide spread is information, not an error, and the memo says which method you chose and why.

Then apply the phase split and the consultant shares to the chosen figure, and show those arithmetic too.

Step 6. Write the deliverable

Fill assets/fee-estimate-template.md. Head it with the project basis and the source of every rate. Foot it with the methods not run.

Step 7. Update the index

python scripts/append_to_index.py \
  --date <YYYY-MM-DD> \
  --project "<slug>" \
  --type "<project type>" \
  --sector "<sector>" \
  --cc "<construction cost>" \
  --pct "<percentage applied>" \
  --amount "<fee amount>" \
  --source "<slug>-fee.md"

Do not edit docs/fees/index.md by hand; the table alignment breaks.

Step 8. Report

The four figures, the spread, the chosen method, the splits, the methods not run, and:

Every rate, percentage and cap here came from you, with its source beside it. I ran the arithmetic and supplied no figures. Methods marked not run had no inputs.

Rules

• Never supply a rate, a percentage, a split, or a multiplier. Not from memory, not from a search, not from a figure in this file. There are none in this file.
• Never quote a compensation survey. They are paywalled, and a remembered figure from one is both unreliable and not yours to publish.
• Derive from the user's own comparables where possible, and show the derivation, so the number belongs to them.
• Ask whether consultant fees are inside or alongside the total. Getting that backwards changes the answer by more than any rate choice.
• Show the arithmetic for every method, and print not run rather than filling a gap.
• A procurement cap is applied only when the user names the rule and the cap.

Anti-patterns

• Offering a percentage band because the project type is a familiar one. That band is exactly what the paywalled surveys sell, and a recalled version of it is a number with no provenance carrying your authority.
• Filling a phase split or a consultant share from convention.
• Applying a statutory or procurement cap the user did not name.
• Running one method and presenting it as the answer, with no note that three did not run.
• Rounding a construction cost or a derived fee to a neater figure.
• Bundling a rate table, a phase split, or a procurement cap into this skill later. It had all three, plus a script that applied them, and they were removed.

Resources

• assets/fee-estimate-template.md — the deliverable structure.
• assets/fee-index-template.md — the project index structure.
• scripts/init_workspace.py — bootstrap docs/fees/.
• scripts/append_to_index.py — append a row to the index.

No reference files. The rates come from your projects and your contract.