Blog
8 min read
How to set a stage based fee that survives design changes starts with one defined output. The stage number only identifies the stage.
The RIBA Plan of Work organises a building project into eight stages. It gives each stage an outcome, core tasks, and information exchanges. That structure helps you arrange a fee, but it does not set the boundary of your service.
Do not add a vague redesign allowance to solve that problem. Define the stage boundary before work starts. For each stage, record six items:
This is an estimating method, not contractual advice. Get advice on the legal effect of your appointment terms.
Start with the result that closes the stage, not a list of activities.
“Develop the design” has no visible finish. “Issue the coordinated spatial design pack for client approval” does, but it still needs detail. Name the drawings, schedules, models, reports, format, level of detail, and issue purpose.
Use the RIBA stage outcome as your starting point. Your service must reflect the project, procurement route, and consultant team. The fee also needs a testable output.
For each output, record:
A client approval email or approved meeting minutes can show that the stage is closed. Define this evidence before you price the stage.
If the output is still uncertain, do not force it into a fixed fee. The Architecture for London architects' fee guide describes time charging for feasibility work when the brief, build cost, size, or complexity is not yet clear. A paid discovery phase can resolve those unknowns before you set the later stage fee.
Iteration is design work within the agreed direction. Redesign replaces that direction or reopens a settled input. Your fee needs a practical line between them.
Define one review cycle as a sequence. For example:
State who can comment and require one combined response. Separate comments across several weeks can turn one nominal review into several design passes.
Set the permitted scale of the review. You can include corrections and clarifications, plus selection between the priced options. A new accommodation need, another option, or reversal of an approved decision can trigger a new estimate.
This approach builds on defining deliverables that clients cannot expand. Give the stage output a quantity and an acceptance test. Then define the one cycle that moves it to acceptance.
An assumption is a fact that you use to estimate work before confirmation. A dependency is an input or action that another party must provide.
Write both beside the stage fee. Keep them specific enough to test.
Weak: “Client information will be provided promptly.”
Useful: “The client will issue one approved equipment schedule before the spatial design review.”
Useful dependencies can include:
Designing Buildings lists the brief, procurement route, project complexity, stakeholder count, programme, and consultant roles among relevant fee considerations for architects. Convert those broad considerations into project-specific assumptions. Do not leave them in an internal risk note.
If the brief is incomplete, resolve each gap before you fix the fee. Price what is known, use a stated allowance where suitable, or wait for an answer. The incomplete brief decision tree explains that choice.
When a dependency fails, start an assessment. Record the extra output, programme effect, and price before the extra work begins.
A reopening trigger is an observable event that requires you to reassess the remaining stage work. It is not an automatic charge.
“Changes may cost extra” is too broad to apply consistently. Use events that the project record can show:
Link each trigger to a response. Pause the affected work and describe what you must repeat. Estimate the labour and external cost. Record any programme change, then obtain approval through the agreed project process.
This differs from charging for design changes after stage sign-off. That process deals with a change after the boundary has been crossed. This method designs the boundary and its trigger before stage work starts.
Test each fee with one realistic change before you issue the proposal. You cannot predict every request. Instead, check whether your boundary gives the team a clear decision.
Use this stage-fee stress-test matrix:
| Stage output | Included iteration | Client dependency | Change scenario | Reopening trigger | Pricing basis |
|---|---|---|---|---|---|
| Approved project brief and feasibility record | One workshop and one combined comment return | Client confirms priorities and budget, plus site information | A new user group adds accommodation | Requirements added after brief approval | Hours by role for new briefing and testing, plus reissue |
| Concept design pack with one preferred option | Two initial options and one refinement cycle | Client selects one option at the review | Client asks to develop both options further | More than one option proceeds beyond selection | Separate fixed fee from tasks and hours for the extra option |
| Coordinated spatial design pack | One combined review against the approved brief | Consultants issue inputs to the coordination programme | Engineer changes the structural grid after coordination | Issued consultant input invalidates completed coordination | Hours by discipline for affected drawings, model work, and review |
| Planning submission pack | One response to validation comments that does not alter the design | Client approves submission and pays third-party charges | Client changes materials after submission approval | Approved design decision is reversed | Task estimate for drawing, schedule, report, and submission updates |
| Technical design package for the stated procurement route | One coordinated review and correction cycle | Specialist design arrives on the agreed dates | Procurement changes to a route needing different information | Recorded procurement route changes | Reforecast of remaining outputs with role hours and external costs |
| Construction-stage service with stated visit and query limits | Included visits and responses up to the named count | Contractor follows the information and query process | Repeated site queries arise from an unapproved substitution | Visits or responses exceed the count, or substitution causes redesign | Unit rate for extra visits plus task estimate for redesign |
Read across each row. Can the project architect identify when included work ends? Can the client see which input supports your price? Can the practice estimate the response without debating a vague phrase?
If not, revise the boundary.
Use a fixed amount when the added output is clear. Use role hours or unit rates when its volume is not clear. Reforecast the stage when a change alters much of the remaining output.
Build the base stage fee from the resources that you need to produce the defined output. List tasks and role hours. Add external costs, overhead, and margin. Then test that estimate against similar completed stages.
Your own records matter more than a generic stage split. Track time against the same stage and output structure that you estimate. This makes the guide to tracking time against RIBA stage fees the feedback step in this method.
Keep contingency separate from redesign that has no defined limit. Use contingency only for a named uncertainty within the priced scope. Do not use it as a promise to absorb any client change.
Project Control Academy explains that contingency can be based on a simple percentage or on identified risks. It also explains that a general percentage does not address risks unique to one project. Use its contingency methods as a prompt to name the uncertainty, not as a benchmark for architect fees.
Estimate the possible effect of each named risk. Decide whether to include an allowance, change the pricing method, or resolve the risk first.
Do not hide a likely second option inside contingency. Price that option or exclude it.
Put the six boundary items in the fee schedule as well as your estimating notes. Use the same names in your programme, design review agenda, time codes, and change record.
Before work starts, ask the project architect what closes the stage, which reviews are included, and which event requires a new estimate.
If the answers differ from the proposal, the boundary is not ready.
During delivery, record approvals, client inputs, and review comments against the stage. When a trigger occurs, stop the affected work. Assess the effect and issue a priced change before the extra work starts.
At stage close, compare estimated and actual hours by output. Record which assumptions held and which triggers occurred. Use that evidence to price the next stage and the next project.
Korrel records deliverables, assumptions, risk allowances, and payment milestones in a proposal. It then tracks project time and costs against work and rates carried over from the estimate. Start your trial and build your next stage fee around a clear boundary.
RELATED READING:
Price discovery as a defined engagement. Cost each output, name the unknowns it will test, and charge only for the risk that remains.
Three piles: what you price firm, what you price with a stated allowance, and what you refuse to price until the client answers. And how to write each into the quote.
A revision round becomes billable the moment it stops matching the proposal. Define a round properly, then price the extra ones from what delivering one costs.
Korrel turns briefs into structured proposals, then tracks what each job actually costs.
START FREE TRIAL