Friday afternoon, three fee requests, one director
An architect sends a request for a structural and civil fee on a primary school extension. Two other requests are waiting. The director opens a proposal from a school job two years ago, changes the name, nudges the fee up because the site looks awkward, and sends it. There is no time to check how many hours that earlier job actually took, and the timesheet system would not show it by stage anyway.
Sometimes this works. Sometimes the practice wins the job at a fee that was never going to cover the work. Sometimes it loses a job it should have won because the fee was padded out of caution.
Why pricing stays a gut call
- Past hours are recorded by job number, not by stage, so the useful comparison is not available.
- Projects are not tagged by type, size or construction form, so finding comparable jobs means memory.
- Proposals are Word documents, so the fee build-up behind them is lost or was never written down.
- Each director prices differently, so the practice has no shared view of what a stage costs.
- Scope exclusions are copied from old proposals and do not always fit the new job.
The information to price properly exists in the practice. It is just scattered across old timesheets, old proposals and people's heads.
There is also the problem of confidence. A director who knows the practice lost money on the last three schools will pad the next school fee, and one who remembers a single profitable job will cut it. Without numbers, pricing follows memory, and memory favours the most recent or most painful job.
What guess-based pricing costs
| Pricing outcome | What follows |
|---|---|
| Priced too low | Stage overruns absorbed, or strained talks about additional fees |
| Priced too high | Lost work that suited the practice |
| Inconsistent between directors | Clients notice, and margins vary for no good reason |
| Weak scope and exclusions | Arguments later about what was included |
There is a time cost too. Senior people spend evenings writing proposals, and repeat clients get slower responses than they should.
How we build a fee build-up tool
- We pull historical hours from your timesheet or accounts system and tag past projects by type, sector, size, construction form and disciplines, with your help on the ones the data cannot classify.
- Where old time was not recorded by stage, we estimate a split from dates and drawing issues, marked clearly as an estimate.
- For a new request, the director enters a few characteristics and sees comparable past projects with their actual hours per stage and grade.
- A build-up screen starts from those figures and lets the director adjust each stage for this job's risk, with a note explaining why.
- Grade rates convert hours into a fee per stage, using your current rates.
- The proposal document is generated from the build-up: fee schedule, stage scope, assumptions and exclusions from a clause library you control.
- Once the job is won, the stage fees feed straight into fee tracking, so estimated and actual hours can be compared as the job runs.
A language model can draft scope text from the architect's request and your clause library for the director to edit. It does not set the fee, and nothing is sent without a person reviewing it.
We normally begin by pulling a few years of timesheet data and sitting with a director to tag the projects. That exercise alone tends to be revealing, because it is often the first time anyone has seen how long a typical stage 2 or stage 4 really took across a set of similar jobs.
How pricing feels afterwards
A director preparing a fee starts from evidence: what the last few comparable jobs really took, stage by stage. The judgement stays theirs, but it is informed and written down, so another director can see why a fee was set as it was.
Proposals go out faster and more consistently, and each finished project improves the next estimate.
Is this how you price today?
- Fee proposals start from a copy of an old proposal.
- Nobody can easily say how many hours a typical stage 3 took on a similar job.
- Different directors would price the same job very differently.
- The fee build-up behind a proposal is not kept.
- Won jobs regularly overrun in the same stages.