Costs and pricing
Part of Budgeting for a branding agency: a worked example and VAT treatment
What branding agency return on investment can mean once the denominator is complete
Assess branding agency return in England by separating observed outcomes, attribution and causal effect, then applying a fully defined cost denominator.
What to take away
- A publishable branding ROI needs a causally supported numerator and a denominator carrying the full economic cost over the same period.
- Revenue rising after launch is an observation; a reporting rule can attribute it, but neither proves the brand work caused it.
- Cost the work in pounds across every component, and keep recoverable VAT and the cash-flow record out of the economic cost.
- Report a range with named sensitivities, and treat the decision as reopenable when scope, price or measurement changes.
Fix the decision before you touch the ratio
A ratio answers nothing until the decision is named. State the intervention, the eligible audience, the locations and the channels. Add the start date, the exposure period and the outcome window.
Record the decision owner and the alternative: no action, the current identity, or another defined programme. The counterfactual asks what would have happened without the selected work.
Build the complete denominator
Total economic cost is not the agency invoice. It is every economic cost the buyer bears over the stated period, in pounds sterling, matched to the outcome window.
Total economic cost components
- External net costagency fees
- Internal labour opportunity costclient time
- Third-party costresearch, rights
- Implementationtraining, governance
- Measurementchange, exit
total economic cost = external net cost + internal labour opportunity cost + third-party cost + implementation + measurement + change + exit
Value client labour at a documented internal rate. ONS earnings and working hours data gives a defensible reference for comparing freelance day rates with agency fees.
UK standard-rate VAT is 20%. Exclude recoverable VAT from the economic cost, yet keep it visible in the cash-flow record. A budget template with a VAT field shows both views side by side, which is why it belongs next to this denominator.
Agency fees are usually quoted as a fixed project fee, a monthly retainer or a day rate. On a project of this size the fee is normally the largest single line.
Example: an English services firm costed in full
Every figure below is illustrative and labelled, not a market benchmark.
| Cost component | Basis | Illustrative amount |
|---|---|---|
| Agency fees | Fixed project fee, three stages | £68,000 |
| Client labour | 220 internal hours at £55 | £12,100 |
| Research | Interviews and survey fieldwork | £9,500 |
| Third-party rights | Font licence and photography | £2,400 |
| Implementation | Print, web build, sign-off | £14,000 |
| Training | Two internal workshops | £3,000 |
| Governance | Legal and brand council review | £2,500 |
| Measurement | Baseline and post-period tracking | £6,000 |
| Change | Internal rollout support | £4,500 |
| Exit | Asset handover and archive | £1,000 |
| Total economic cost | £123,000 |
Suppose the causal estimate is £340,000 of incremental revenue at a 45% contribution margin. Incremental contribution is £153,000. Take £123,000 away and divide by £123,000, and the ratio is 24.4%.
Change the causal estimate to £250,000 and the ratio falls to -8.5%. Raise it to £420,000 and the ratio reaches 53.7%. The denominator did not move.
Keep the three evidence layers apart
Observed outcome records the defined event in the eligible population and period. Attribution assigns credit using a declared rule, and must specify identity resolution, the window, duplicate handling and missing data. Causal effect compares the observed result with a credible counterfactual.
Three evidence layers
Observed outcome
- What it records
- defined event
- Key requirement
- eligible population
- Main risk
- event selection
Attribution
- What it records
- declared credit rule
- Key requirement
- identity resolution
- Main risk
- poor incrementality
Causal effect
- What it records
- counterfactual comparison
- Key requirement
- credible design
- Main risk
- untenable assumptions
Attribution can support reporting while remaining a poor estimate of incrementality. The guide to five evidence layers shows where attribution stops and causal design begins, so read it before you draft any claim.
HM Treasury's Green Book 2026 and Magenta Book set out appraisal and evaluation discipline for their stated public-sector context. They do not validate a private brand study. Their separation of costs, benefits, risk and uncertainty is still worth copying.
Calculate the ratio only after validation
Define contribution as revenue minus the variable cost of fulfilling that revenue. Do not subtract that variable cost twice if a supplied margin measure already includes it.
incremental contribution = estimated causal revenue effect x applicable contribution margin commercial ROI = (incremental contribution minus total economic cost) / total economic cost
- Confirm the causal design survives scrutiny, and write down its identifying assumptions.
- Fix one contribution margin definition and apply it to every cohort.
- Check that the denominator period matches the numerator period.
- Publish the ratio with its currency, units and period attached.
Report uncertainty and reopen the decision
Show observed outcomes, attributed outcomes and the causal estimate in separate rows. State the estimate range and how missing data was treated.
Set sensitivities with named bounds rather than a generic optimism adjustment. A benchmark specification written before you trust any agency figure explains how to set those bounds, so read it before you publish a range.
A positive attributed ROI does not cure a rights, privacy or contract failure. Reopen the decision after a material change in scope, price, audience or measurement.
Common questions
What makes a branding ROI publishable?
The numerator must be causally supported incremental contribution and the denominator must carry the full economic cost for the same period. Revenue after launch is an observation, and attribution is a reporting rule; neither alone proves cause.
Why does the denominator matter more than the ratio?
A denominator holding only the agency invoice overstates the apparent return whenever the buyer supplied substantial time, media or implementation. The worked case above pairs a £123,000 economic cost with a smaller invoice.
When should causal language stop?
Stop when selection, concurrent activity, measurement change or spillover makes the identifying assumptions untenable. If random allocation is unsuitable, pre-specify another design and its assumptions.
How should VAT be treated in England?
UK standard-rate VAT is 20%. Keep recoverable VAT out of the economic cost and inside the cash-flow record, and confirm the treatment with a qualified tax review.
