Costs and pricing
Working out whether a creator media kit paid for itself, without inflating the evidence
Creator media kits return on investment in England: separate observed outputs, allocated value, causal evidence and projections before calculating a return.
What to take away
- No England-wide figure for creator media kit return appears in this research, so the useful output is a bounded calculation with labelled inputs, not a percentage presented as measured impact.
- Keep four categories apartrecorded output, allocated value, causal impact and unverified projection.
- Compare an evidenced incremental cost with an evidenced incremental benefit across one fixed window.
- When a benefit cannot be evidenced, report the ratio as UNKNOWN and show any allocation-based figure as a scenario.
- Attribution needs a credible counterfactual, and timing alone proves nothing.
Decide what change you are testing
Name the change under review: a new hosted portfolio, a different export process, or a revised maintenance arrangement. Keep the calculation tied to that change. A general rise in brand enquiries is not automatically an effect of the kit.
Fix the observation window before you look at results. Write down the previous arrangement and anything else that changed in the same months. Anything you cannot evidence stays as UNKNOWN.
HM Treasury's Magenta Book separates process, impact and value-for-money evaluation. It is UK government methodology rather than a creator-industry benchmark, and it is used here only as a framework.
Keep the four evidence categories apart
Four evidence categories
What the record contains
- Recorded output
- A verified export, correction or logged recipient action
- Allocated value
- An owner-approved valuation rule applied to a record
- Causal impact
- Evidence of what would have happened otherwise
- Unverified projection
- Assumptions about future use
What it cannot establish
- Recorded output
- Money caused by the kit
- Allocated value
- Cash received
- Causal impact
- A result inferred from timing
- Unverified projection
- An observed return
Log the source and the date each time you add an entry. Relabelling a projection does not turn it into an output.
[figure 1]
Evidence categories and limits
Recorded output
- What it contains
- Verified export or action
- What it proves
- Activity only
- Logging rule
- Source and date
Allocated value
- What it contains
- Owner-approved valuation rule
- What it proves
- Not money received
- Logging rule
- Approved purpose
Causal impact
- What it contains
- Counterfactual evidence
- What it proves
- Not timing alone
- Logging rule
- Design support
Unverified projection
- What it contains
- Future assumptions
- What it proves
- Not observed return
- Logging rule
- Separate label
Value labour time only under a stated rule
If the operator logs less time on a comparable maintenance task, keep the task definition and the time record together. Decide whether that is released capacity, an avoided payment, or a different way of recording the work. Each has a different financial consequence.
Any monetary allocation needs an evidenced rate and an approved purpose. If pay is unchanged and the freed time is not used for paid work, do not call it a cash saving. An auditable blank budget for a creator media kit with VAT fields and due dates keeps the cost column tied to documents you can produce.
Check attribution before you calculate
Monitoring a change does not establish causation. The quality in policy impact evaluation guidance explains why a credible counterfactual matters, and its worked examples come from public policy rather than creator work.
Ask what evidence could separate the kit from a new recipient list or a changed service offer. A before-and-after comparison stays descriptive. The comparison of evaluation designs for a media kit question sets out what each design can carry and where none wins outright.
Example: an illustrative year with labelled inputs
Every figure below is an illustrative input for the method, not an observed England market rate and not a real creator case. Replace each one with a document you hold.
- Set the change and the windowa hosted kit replaces a PDF, 1 April to 31 March.
- Add incremental costshosting £96 and a one-off design fee £250, so incremental cost C is £346.
- Record outputsthree qualified briefs arrive and are logged with dates, with no deal value attributed to the kit.
- Apply an allocation rule if the owner approves one£150 per qualified brief gives £450.
- Calculate(£450 minus £346) divided by £346 equals 0.30, a 30 per cent net return under the allocation rule only.
- Report the causal benefit as UNKNOWN, because no counterfactual exists.
[figure 2]
Can the ratio be calculated?
Are C and B evidenced, incremental and comparable?
apply (B minus C) divided by C
report the ratio as UNKNOWN
Illustrative year with labelled inputs
| Line | Amount | Evidence status |
|---|---|---|
| Hosting, 12 months | £96 | Invoice |
| One-off design fee | £250 | Invoice |
| Incremental cost C | £346 | Evidenced |
| Allocated benefit B | £450 | Allocation rule only |
| Causal benefit | UNKNOWN | No counterfactual |
| Net return, (B minus C) divided by C | 30 per cent | Allocation scenario |
If the creator is VAT registered, VAT charged on that design fee is normally recoverable and stays out of the cost column. GOV.UK's VAT registration guidance sets out when registration is required.
Report the ratio with its label attached
Publish the records, the valuation decisions and the attribution limits alongside any number. Reporting on a creator media kit without inventing impact covers how to place a figure next to what it does not prove.
If the incremental benefit is unproven, state the observed activity and stop there. Do not carve a share out of a later contract and call it measured impact.
Common questions
Can I claim a return by counting portfolio views and valuing each one?
No. A view count is recorded output. Putting an invented price on it creates an allocation, not evidence of money received.
What does a missing or zero cost do to the ratio?
A missing benefit or cost leaves the ratio UNKNOWN. A zero incremental cost makes it undefined, because the divisor cannot be zero.
How long should the observation window be?
Long enough to cover one enquiry-to-invoice cycle and any renewal date, which for many creators means 12 months. Set the dates before you review results.
What if the kit cost only my own time?
Count that time as a cost only if it displaced paid work or an invoice you would otherwise have paid. Otherwise keep it as recorded output and leave the ratio UNKNOWN.
