Self-reported attribution (SRA)
Self-reported attribution means asking customers how they found you, which picks up AI recommendations your analytics can't see.
Self-reported attribution, or SRA, means asking customers how they heard about you and recording what they say. Usually it's one question on a sign-up form, at checkout, or in a welcome email: "How did you hear about us?"
It's become more useful with AI search, because a lot of AI's influence never shows up in analytics. Someone reads your name in an answer, then searches for you or types your address. Your analytics files that visit as search or direct. Only the customer knows the assistant sent them.
Why analytics misses it
Analytics tools work out where a visit came from using the referrer string the browser sends. Plenty of AI-driven visits don't carry one. The person copies a name, switches app, or comes back days later. Some never visit at all and buy in a shop.
For example, Kestrel Roasters' analytics shows 40 orders a month from AI assistants. When it adds "How did you hear about us?" at checkout, 110 customers a month choose "ChatGPT or another AI assistant." The real number is closer to the second.
How to ask well
Keep it to one question, and make it quick to answer:
- Offer a short list of choices, with AI assistants as one of them
- Add a free text box, so people can name the assistant or say more
- Ask at a moment people are already filling in a form
- Don't make it compulsory, or people pick the first option to get past it
Then read the free text now and then. People often say exactly what they asked, which gives you new prompts to track.
Its limits
People forget, guess, or pick the most recent thing. Read the answers as a trend and a rough share, next to your analytics, rather than as an exact count.
In AnswerOrbit
AnswerOrbit can't see what people do after reading an answer, such as visiting your site or buying from you. See What we can't see. The AI traffic tab in Website and AI traffic shows the visits your Google Analytics can tie to an assistant, and asking customers fills in the rest.