Organise links, codes and costs to assess an influencer campaign and honestly present attribution limits.
A view, click and order answer different questions. Choose the objective first, then build a shared table with creator, content, date, cost, link, any code and observation period. Test links before publication and document offer changes.
Read results cautiously
Suppose spend is €1,000 and 40 orders are identified by a code: the cost per identified order is €25. That does not prove the content caused all those orders; some people may buy without the code, while others already knew the brand. Platform clicks and website visits can also differ.
Define what is missing
Report return on investment only when relevant revenues and costs are available under a consistent method. State returns, measurement window and the share of orders actually attributable. Without reliable data, present observed signals and their limits. For personal data, consult CNIL's cookie and tracker guidance and select a method appropriate to your circumstances.
Build a link and tracking table
Associate each creator with a link version and publication date in a spreadsheet. Fictional example: https://example.com/produit?utm_source=instagram&utm_medium=creator&utm_campaign=lancement&utm_content=profil_a. Use the same campaign name and a different content identifier per post. Parameters should not contain names, email addresses or other personal information. Google explains campaign parameters and dimensions.
Review technical tracking alongside the privacy policy and visitor choices. CNIL explains conditions for consent exemptions for certain audience-measurement tools: do not assume a configuration is exempt without checking those conditions.
Checklist
- Agree a window and source for each indicator.
- Test links and codes; record promotions and stockouts.
- Deduplicate orders appearing in multiple reports.
- Document returns, costs and coverage before reporting a ratio.
- Explain what was not measured and what should be tested next.
For example, if the same order appears in both link and code reports, match its order identifier before adding totals. If identifiers are unavailable, disclose possible duplication in the report.
Frequently asked questions
Does a code measure every influenced sale? No. Some buyers do not use it; others might have purchased without the campaign.
Is revenue divided by spend profit? No. It may omit margins and other costs.