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.