Attribution Questions: What Can and Cannot Be Known
Understand the limits of attribution before choosing a model.
The Open buttons send the prompt text as currently shown below to that site only when you click. Fill in the placeholders first.
[VERIFY] is an editorial flag, not a blank to fill. Wherever the prompt shows [VERIFY], check that item against a reliable source before you use it.
Explain attribution for my business. Customer journey: [CUSTOMER_JOURNEY]. Channels used: [CHANNELS]. Tracking available: [TRACKING]. Length of sales cycle: [SALES_CYCLE]. Compare first touch, last touch, linear and position-based models using one invented customer journey with four touches and a sale of 100, showing how credit is split under each. Explain what each model hides. Then recommend a practical approach for my situation: which model to report, how to complement it with simple checks such as asking customers how they heard about us, and which decisions should not rely on attribution at all. Include a one-page note for non-technical colleagues. Tracking and consent limits are [VERIFY] for [REGION]. Show the invented journey: ad click, blog visit, email click, direct visit and sale of 100. First touch gives 100 to the ad; last touch gives 100 to direct; linear gives 25 to each of four touches; position-based gives 40 to the first, 40 to the last and 10 each to the middle two. Explain that all four are simplifications, none shows causation, and each favours different channels. Suggest looking at paths to see which channels often appear together and use a holdout test for important channels where possible. Add a short checklist for deciding whether to invest in more sophisticated attribution: spend level, number of channels and quality of tracking. Add a final practical step: choose one important channel and design a simple holdout or pause test to learn its true contribution, with the duration, the measure, the risks and the decision to be made afterwards. Note that the result applies only to the period and audience tested. Add a short explanation of why different tools report different numbers for the same campaign, covering time windows, cookie loss and conversion definitions. Close with a statement in plain words that can go in a report: what the numbers show, what they cannot show, and what test would reduce the uncertainty.
Fill in before use
Optional: type your details and the prompt above updates. Nothing is saved or sent.
Replace each bracketed placeholder with your own details, then paste into Any AI or another assistant.