Most creator programmes are judged on discount code redemptions, and that number is wrong in both directions at once. Codes miss the people who watched a creator and bought a week later. They also credit creators for customers who would have bought anyway. Measure in three layers instead: tracked links and codes, paid amplification you can test, and a holdout. Then reduce it all to one figure you can put next to every other channel, cost per incremental customer.
Where the count goes wrong
Creator content sits somewhere between advertising and word of mouth. Someone sees a video, does nothing, meets your brand again in a search result a few days later, and buys. Last-click attribution hands that sale to search. A discount code hands it to the creator, but only if the buyer remembered the code.
Codes also leak. They end up on coupon sites and in browser extensions that apply them at checkout, and existing customers use them because a discount is a discount. The creator then gets credit for sales they didn’t cause.
That is undercounting (no code, no click) and overcounting (leaked codes, existing customers) in the same report. No single tracking method fixes both. Layering does.
Layer 1: links and codes, set up properly
Codes and links are cheap and fast. They give you a floor, and a rough ranking between creators. Set them up like this:
- Give each creator one code and one link, named on a consistent pattern so codes and UTM parameters group by campaign and creator. Our UTM naming convention tactic has a template.
- Restrict codes to new customers where your platform allows it. That removes most of the existing-customer overcount.
- Watch for leakage. A spike of code orders from people who never visited the creator’s link, or redemptions carrying on long after the content stopped, usually means the code has reached a coupon site.
- Ask at checkout. A one-question “Where did you hear about us?” survey with creator names as options picks up some of the people who bought with neither code nor link.
- Track by cohort. Compare the 90-day repeat rate and refund rate of creator-sourced customers with your other channels. A creator who brings cheap first orders and no second orders is expensive.
Read this layer as directional. It tells you which creators to keep testing. It can’t tell you what the programme is worth.
Layer 2: partnership ads and whitelisting
The second layer is paid amplification of creator content, usually called whitelisting or, on Meta, partnership ads.
Meta’s help article on partnership ad permissions explains that branded content ads are now called partnership ads, that you need the creator’s permission to run ads from their handle, and that creators can revoke that permission at any time. Permission comes in two kinds: content-level for a specific post, and account-level for ongoing partnerships.
Once creator content runs as an ad, you can test it the way you test any ad. Put creator content against your brand-made creative in a split test with the same budget and audience. Compare creators with each other on cost per new customer, spend held equal. And use the ad platform’s conversion tracking, which records view-through and later conversions that a code never sees.
The economics change too. Organic reach on a creator’s post happens once; a post that works as an ad can run for weeks. Some brands now pay creators mainly for content and usage rights, then buy the reach with their own media budget. Our UGC creator ads tactic covers briefing and contracting for this.
Write usage rights into the contract: which platforms, how long, whether you can edit, and whether you can run it from the creator’s handle or only from yours.
Layer 3: holdouts and lift
Only an experiment tells you what would have happened without the creators. Three versions are practical for most brands.
Platform lift tests
Meta’s Conversion Lift compares conversions between a test group that sees your ads and a control group that does not. If your account is eligible, run one on the partnership ads campaign. The difference between the groups is the incremental effect of the paid creator content.
Geographic holdouts
You can’t hide an organic creator post from a control group, but you can choose where you work. If you sell across several cities or regions, run creators in some and not others for six to eight weeks, then compare new customer growth in each region against its own previous trend. It works best when creators have local audiences, which in India is common with regional-language creators.
Brand search and direct traffic
This is the cheapest signal of the three. Watch branded search volume and direct traffic in the days after each post goes live. A creator who sparks real interest shows up as a bump in people searching your name. The data is noisy. It is also free, and it catches the “saw it, searched later” buyers that codes miss.
Whichever you pick, write down the expected effect and the decision rule before you start. The hypothesis design lesson explains how.
Cost per incremental customer
The three layers come together in one division:
Cost per incremental customer = total programme cost ÷ incremental new customers
Total programme cost means creator fees, product sent, shipping, paid amplification, agency or platform fees and your team’s time. Incremental new customers come from the holdout or lift test. Code redemptions don’t go in this sum.
The figures below are hypothetical: a D2C brand running a ₹10,00,000 creator programme over two months.
| Line | Amount |
|---|---|
| Creator fees (12 creators) | ₹5,50,000 |
| Product and shipping | ₹50,000 |
| Partnership ads spend | ₹3,00,000 |
| Management time | ₹1,00,000 |
| Total cost | ₹10,00,000 |
Codes and links show 400 new customers, an apparent ₹2,500 each. The geographic holdout tells a different story: regions with creators gained 700 more new customers than comparable regions without them, so the cost per incremental customer is about ₹1,430. And when someone checks the 400 code users, 60 turn out to be existing customers and 50 came through coupon sites. The code count was inflated and still missed most of the effect.
The ₹1,430 is the number to set against your paid social or search cost per incremental new customer when you decide next quarter’s budget.
Disclosure rules in India, the US and the UK
None of this measurement helps if the programme ends in a regulator’s ruling. The three markets overlap but differ in detail. As of October 2026:
| India (ASCI) | US (FTC) | UK (ASA and CMA) | |
|---|---|---|---|
| Trigger | Any material connection, including free products | Any material connection, including free or discounted products | Payment or any other incentive, requested or not, including gifts |
| Preferred labels | Advertisement, Ad, Sponsored, Collaboration, Partnership, Employee, Free gift, Affiliate, or a platform disclosure tool | “Ad”, “advertisement”, “sponsored” | “Ad” upfront; ASA advises against “sponsored”, “gifted”, “spon” |
| Platform tool alone | Listed as acceptable | Not assumed to be enough | “May not be enough on its own” |
| Who is responsible | Advertiser and influencer | Brand and influencer; brands must train and monitor | Everyone involved; brands at least jointly for affiliates |
India
ASCI’s influencer guidelines require disclosure whenever there is a material connection, even if the review itself is unbiased, and the label must be upfront and prominent. Video timing is spelt out. For videos of 15 seconds or less, the label stays on screen for at least 3 seconds; between 15 seconds and 2 minutes, for a third of the length; for anything longer, for the whole section where the brand is discussed.
Virtual influencers must also disclose that they are not human. Influencers giving technical advice on financial or health products need relevant qualifications (SEBI registration for stock and investment advice, for example) and must state them upfront; generic content doesn’t need them. If a post breaks the rules, the advertiser must ask the influencer to edit or delete it. India’s Department of Consumer Affairs also published “Endorsement Know-hows!” guidance for influencers in 2023 under the Consumer Protection Act.
United States
The FTC revised its Endorsement Guides in June 2023, according to its press release. Its Disclosures 101 guide says to put the disclosure with the endorsement, in the video itself (not only the description), repeated during live streams, and warns: “Don’t assume that a platform’s disclosure tool is good enough.” Its FAQ says advertisers need “reasonable programs in place to train and monitor members of their network”. The FTC’s Consumer Reviews and Testimonials Rule is a separate instrument covering fake reviews and testimonials, including misrepresenting a paid influencer’s experience.
United Kingdom
The ASA’s guidance on recognising ads says the ASA and the CMA prefer labels that explicitly say “advertising”, such as “Ad”, and advises against “sponsored”. A discount code on its own doesn’t make the relationship clear, and any content about a product with an affiliate link or code counts as advertising. The CMA’s guidance for content creators adds that the label must be clear from the first interaction, without scrolling, expanding or checking the bio.
For a programme running in all three countries, the safe default is “Ad” at the start of the caption and on screen in the video, the platform’s paid partnership tool on top, and written briefs with a pre-approval step for every paid post.
Before your next creator wave
- Issue one code and one tracked link per creator, codes restricted to new customers.
- Add a “Where did you hear about us?” question to checkout or sign-up, with creator names listed.
- Get partnership ad permission from your two best creators and split-test their content against your own creative.
- Choose the holdout regions for a six-week test and write down the decision rule now, before the posts go live.
- Put disclosure requirements, usage rights and a pre-approval step into your creator contract template. If you’re new to working with creators, start with the influencer seeding tactic.
- When the wave ends, work out cost per incremental customer and compare it with your other channels. Nobody gets renewed before that.
Questions people ask
Are discount codes a good way to measure influencer marketing?
They give you a floor and little else. Codes miss people who saw the post and bought later without the code, and they over-credit creators when codes leak to coupon sites or are used by existing customers. Combine them with tracked links and a holdout or lift test.
What is the difference between paid partnership labels and partnership ads on Meta?
A paid partnership label discloses an organic post a creator made for a brand. Partnership ads are when the brand pays to run that content as an ad from the creator's handle, which needs the creator's permission and can be measured like any other Meta ad.
Is 'Sponsored' an acceptable disclosure label?
It depends on the country. ASCI in India lists 'Sponsored' among permitted labels and the FTC says 'sponsored' is acceptable in the US, but the UK's ASA advises against it and prefers 'Ad'. If you run one campaign across all three, use 'Ad' at the start of the post.
How do you calculate cost per incremental customer for creators?
Divide the total programme cost (fees, product, paid amplification and management time) by the extra new customers measured against a holdout or baseline. Code redemptions are the wrong denominator. The result can be compared directly with your other channels.
Sources
- ASCI: The ASCI Code guidelines (influencer advertising in digital media)
- FTC: Disclosures 101 for social media influencers
- FTC: Endorsement Guides, what people are asking
- FTC: Federal Trade Commission announces updated advertising guides to combat deceptive reviews and endorsements
- FTC: The Consumer Reviews and Testimonials Rule, questions and answers
- ASA: Recognising ads, social media and influencer marketing
- GOV.UK (CMA): Social media endorsements, being transparent with your followers
- Meta Business Help Center: About partnership ad permissions
- Meta for Business: Conversion Lift
- Department of Consumer Affairs (PIB): Endorsement Know-hows! for celebrities, influencers and virtual influencers
