Glossary
What is marketing attribution?
Attribution is deciding which marketing channels and touchpoints get credit for a conversion, such as a sign-up or sale. Models include first touch, last touch and multi-touch. Since buyers often see a product in several places before acting, no model is perfect; asking customers where they heard of you fills gaps.
Common models
| Model | Credit goes to |
|---|---|
| First touch | The first channel that brought the person |
| Last touch | The channel right before converting |
| Linear | Every touch equally |
| Data-driven | Weights learned from data (as in GA4) |
What's hard to attribute
- AI assistant recommendations, when the user types your URL later.
- Directory listings seen weeks before a branded search.
- Podcasts, word of mouth and private messages ("dark social").
These often show up as direct or branded search traffic.
Practical approach
Tag links with UTM parameters, watch referral traffic and branded search growth, and add a "How did you hear about us?" field at sign-up. The self-reported answers regularly reveal channels analytics can't see, such as "ChatGPT recommended it".
Example
Analytics credits most sign-ups to direct traffic. A sign-up question reveals a quarter mention a directory or an AI assistant. The founder keeps investing in listings that analytics alone undervalued.
FAQ
Which attribution model is best?
None is perfect; use one consistently and add self-reported data.
How do I attribute AI search traffic?
Look for AI referrers in analytics and ask customers directly.
What is dark social?
Sharing through private channels that analytics can't see.