What attribution cannot do
A buying journey passes through devices, private browsing, conversations, AI assistants and offline moments that no tool observes. Privacy rules and consent choices remove more. Any model that claims to divide credit for a sale precisely between touchpoints is describing the part of the journey it can see.
Accepting this is useful. It moves the conversation from which channel deserves the credit to which decision we are trying to make.
Questions that change a decision
A measurement question is worth answering if a different answer would lead you to do something different.
- If we stopped this channel, what would we lose?
- Which sources produce customers, not only leads?
- What does a customer cost to acquire, by route?
- Where do qualified buyers first hear about us?
Questions that mostly do not
Arguments about whether the first or last click should receive the credit rarely alter a budget. Nor does a decimal-place comparison between two attribution models. If the two views rank your channels in the same order, pick one and move on.
Use three views together
No single method is reliable alone, and they fail in different directions. Platform and analytics data show what was clicked. A plain question on the enquiry form, asking how the person heard about you, shows what they remember. Controlled changes, such as pausing spend in one region for a period, show what happens when a channel is removed.
Where all three point the same way, you can act with confidence. Where they disagree, you have found something worth investigating.
Get the definitions agreed first
Most reporting disputes are definitional. Before building a dashboard, agree with sales and finance what counts as a lead, a qualified lead, a customer and revenue, and which system is the source of truth for each. A revenue view built on shared definitions and imperfect data is more useful than a precise one that the finance team does not recognise.
