Analytics and attribution

What it is

The tracking, the data model, and the reporting that let you answer "what produced that?" without a debate.

When it is the answer

When most of last year's revenue is attributed to "referral" or "not sure". When you cannot defend a budget decision with anything but instinct. It is usually the first thing to fix, because every other decision depends on it.

When it is not

Rarely. The honest caveat is that attribution is never perfect, and a model that claims to be is hiding its assumptions.

What we do with it

• Instrument every path a customer can take to reach you, including the ones nobody built on purpose
• Join marketing activity, CRM records and closed revenue into one model
• Identify the revenue your current reporting cannot explain — usually the interesting part
• Report on revenue, cost per customer and payback rather than on impressions
• Publish the assumptions in the model, so it can be argued with

What it looks like when it is working

You can trace a conversation, a proposal and a payment to the specific thing that started it — and you move budget because of it.

The mistake we see most often

Believing the dashboard. Every attribution model makes choices about credit, and a model whose assumptions are not written down will quietly flatter whoever built it.