Creator partnerships and affiliate marketing: models, contracts and disclosure
A guide to working with creators: gifting, flat fees, affiliate commission and partnership ads, plus contracts, usage rights and US, UK and EU disclosure rules.
How to set objectives, choose KPIs by funnel stage, track influencer sales with UTMs, codes and affiliate links, test incrementality and check audience quality.
Most influencer campaigns are judged on a screenshot of likes and a discount-code tally, and both tell you less than they appear to. Measuring a creator partnership properly means deciding in advance what it is supposed to change, tracking it through more than one route, and accepting that some of the value only shows up when you compare people who saw the content with people who did not. This guide sets out a measurement plan a small or mid-sized business can run without a data team.
Influencer activity can do three quite different jobs, and each needs its own yardstick.
Write the objective into the brief and the contract. A creator paid for awareness should not later be judged on code redemptions, and a creator on an affiliate deal should not be expected to deliver reach targets. Misaligned expectations are the most common reason partnerships end badly.
| Stage | Primary KPIs | Where the data comes from | Typical reporting window |
|---|---|---|---|
| Awareness | Reach, views, average watch time, audience match | Creator’s platform insights; paid partnership ad reporting | During the posting period |
| Consideration | Link clicks, saves, shares, profile visits, landing-page sessions | Platform insights; your web analytics | Posting period plus one to two weeks |
| Conversion | Orders, revenue, new customers, cost per acquisition | Your ecommerce platform; affiliate network; code reports | Posting period plus your normal purchase cycle |
| Retention | Repeat purchase rate of acquired customers | Your customer database | Three to six months |
Definitions matter here. On Instagram, Meta’s Media Insights reference distinguishes views (the total number of times a piece of media has been played) from reach (the number of unique accounts that saw it), and it deprecated the older impressions metric for media created after 2 July 2024. A creator who reports “views” as though it were reach will overstate their audience. Our explainer on social media engagement metrics sets out the formulas in full.
No single tracking method captures every sale. People see a video on their phone, search for the brand on a laptop later, and buy without clicking anything. Combining methods narrows the gap.
Give each creator a link with UTM parameters so your analytics can attribute sessions and conversions. Google’s guidance on campaign URLs lists the standard parameters: utm_source, utm_medium, utm_campaign, and optionally utm_content to tell creatives apart. A workable convention is:
utm_source: the platform (instagram, tiktok)utm_medium: influencerutm_campaign: the campaign name (spring_launch_2026)utm_content: the creator’s handle and format (janedoe_reel1)Keep parameter values lower case and consistent; analytics tools treat “Instagram” and “instagram” as different sources. Short links are fine as long as they redirect to the full tagged URL. Bear in mind that consent banners can suppress analytics for visitors who decline cookies, so tagged-link data will undercount; our guide to cookie consent and analytics explains why.
Codes catch buyers who never click: they remember the code and type it at checkout. They also leak. Codes get posted on voucher sites, so a code that suddenly produces orders from unrelated traffic sources has escaped. Use one code per creator, set an expiry date, and compare redemptions against the creator’s posting dates.
Affiliate platforms, and the creator tools built into TikTok Shop and some Meta products, track clicks through to purchase with a cookie or platform-level identifier and pay commission automatically. They add a lookback window, which you set, and they make commission-based deals simple to administer. Affiliate links are also a disclosure matter: the US Federal Trade Commission’s Endorsement Guides FAQ says creators must disclose the commission relationship clearly and close to the link, and that labels such as “affiliate link” alone may not be understood by consumers. Our guide to creator partnerships and affiliate marketing covers disclosure rules in the US, UK and EU.
You will end up with two sets of numbers that do not match, and that is expected.
Platform metrics (views, reach, engagement, clicks) are measured inside the app with signed-in users, so they are complete for on-platform activity but say nothing reliable about what happened on your website. Your own analytics measure sessions and purchases on your site, but lose people who decline tracking, switch devices or buy later through another route. Link clicks reported by the platform will almost always exceed sessions recorded by your analytics, because of in-app browser behaviour, slow page loads and consent choices.
Rather than picking one source, keep a simple reconciliation sheet per creator: platform clicks, analytics sessions, orders attributed by UTM, orders with the creator’s code, and affiliate-tracked orders. De-duplicate orders that appear in more than one column. Over several campaigns, the ratios between columns become a useful baseline. If a creator’s clicks-to-sessions ratio is far worse than everyone else’s, the link may be broken or the clicks may not be genuine.
Attribution tells you which sales touched a creator. It does not tell you how many of those people would have bought anyway. Loyal customers who follow a creator you sponsor will happily use a discount code for a purchase they were already going to make.
The way to estimate the causal effect is a holdout: deliberately withhold the activity from a comparable group and compare outcomes. There are three practical versions.
Illustrative example: a skincare brand attributes 400 orders to a creator campaign through codes and links. A geographic holdout suggests that sales in the test regions rose by 250 orders more than in the control regions. The brand’s best estimate of incremental orders is therefore about 250, not 400, and its cost per incremental customer is correspondingly higher than the attributed figure suggests. Neither number is “the truth”, but the gap between them tells you how much of the campaign was harvesting existing demand.
Metrics are only meaningful if the audience is real and relevant. Follower counts can be inflated with purchased followers, engagement pods or bot comments, and creators may have a genuine audience that simply does not match your customers.
Regulators treat fake influence as deceptive. The FTC’s rule on fake reviews and testimonials, announced in August 2024, prohibits selling or buying fake indicators of social media influence, such as bot-generated followers or views, where the buyer knew or should have known they were fake and used them to misrepresent influence for a commercial purpose. In the UK, the CMA’s guidance for content creators, updated in 2025 to reflect the Digital Markets, Competition and Consumers Act 2024, is clear that endorsements must not mislead.
Practical checks before signing a creator:
HowSociable’s guide to spotting real versus fake followers walks through these warning signs in more detail and is a useful checklist to work through with a shortlist.
At the end of each campaign, produce a one-page summary per creator: objective, cost (fee, product, commission, ad spend), delivery (posts, reach, views), response (clicks, sessions, saves), outcome (attributed orders, new customers, incremental estimate where available) and content quality (could the assets be reused in ads or on product pages?). Our comparison of ecommerce analytics tools may help if your current setup cannot produce these figures easily.
In our view, the strongest signal for renewal is not the single best campaign result but consistency: creators whose audience reliably matches your customers, whose tracked results reconcile cleanly, and whose content keeps working when reused as paid partnership ads.
A guide to working with creators: gifting, flat fees, affiliate commission and partnership ads, plus contracts, usage rights and US, UK and EU disclosure rules.
What engagement rate, reach, views, saves, shares, watch time and click-through rate measure, with formulas, worked examples and why benchmarks rarely transfer.
A practical guide to shoppable posts, in-app checkout versus link-out, catalogues, live shopping and DM sales, with platform differences as of November 2026.