PRICING MODEL · ARCHITECTURE PRACTICE
An architecture firm agrees a fee months before it finds out what the job takes. This model builds the fee from the hours the work needs, then checks that guess against what really happened.
Margin the year made
4.1%
Against a 15% target
Hours past the estimate
13.1%
56,404 planned, 63,792 worked
Multiplier achieved
2.66x
Fees are priced for 3.00x
Every job added together, split into the five phases an architecture job runs through.
Site work was planned at 11,280 hours and took 15,311. That one phase is 55% of the whole overrun. It happens because the site allowance gets set as a flat share of the fee, and the fee has nothing to do with how many questions a contractor asks.
Nobody logs the Thursday that went on a site query either, so the overrun never shows up until the job closes.
Same firm, same year, same people. The only thing that changed is how the fee was agreed. Open a row to see the jobs behind it.
Bars run to a 15% target
Fixed fee work earned eight times the margin of work priced off the build cost.
There is a worse problem hiding in that bottom row. Charging a percentage of construction cost pays the architect more when the builder’s costs blow out, and less when everyone does their job well. Everybody in the industry knows this and almost nobody prices around it.
Neither one asks anybody to work harder, and neither needs a difficult call with a client already on the books.
Change the overhead rate, drop the multiplier, or switch the scenario to Downside. Every sheet moves with it.
Open the model11 SHEETS · 605 FORMULAS · NOTHING HARD CODED
Built to show the method. The practice, the projects and the figures are invented so the model can be handed out. The structure is the one that ships. See the rest of the work.