Analytics6 min read

Your attribution model is wrong. Here is how to make it useful anyway.

Perfect attribution does not exist in a six-month, multi-stakeholder buying cycle. Directionally honest attribution does, and it is enough to make good decisions.

Every attribution debate eventually reaches the same impasse: last-touch overcredits the bottom of the funnel, first-touch overcredits the top, and the multi-touch models are opaque enough that nobody in the room believes them.

Stop trying to be right, start being consistent

The goal is not a model that perfectly assigns credit. It is a model that tells you the truth about direction — whether a channel is doing more or less than it was last quarter. A consistently applied imperfect model does that. Three competing models do not.

The three numbers worth reporting

  • Cost per qualified opportunity by channel, using a single agreed model
  • Assisted influence: what fraction of closed deals ever touched the channel
  • Incrementality tests on your largest line items, run once or twice a year

Why incrementality beats modeling

If you genuinely need to know what a channel produces, turn it off in a controlled way and watch what happens. A two-week geographic holdout on branded search answers a question that no amount of modeling will settle, and it costs less than the meetings spent arguing about it.

Get in touch

Let's talk about what growth actually looks like for you.

Tell us what you sell and who buys it. We'll tell you honestly whether we can help — and what we'd do first.