Attribution only counts if it ends at a won deal
Channel reports that stop at leads or clicks will undercount some channels and flatter others. Tie spend to closed revenue.
Attribution models tend to stop where marketing data stops: at the lead, the form, the click. That is where the reporting is easy. It is also where the answer gets distorted.
I have seen a channel that brought in real customers show up as almost nothing, simply because its leads were recorded in a way the model didn't recognize. The spend decisions that followed were built on that gap.
What I have learned
Attribution is useful when two things are true. The channel definitions are consistent from the first touch to the CRM record, and the chain ends at something the business can verify, which is a won deal.
Once advertising spend was tied directly to won deals, the conversation changed. It stopped being "which channel generates the most leads" and became "which channel generates customers, and at what cost."
What I would check first
- Does every lead source value in the CRM map to a real channel, or is "Other" doing a lot of work?
- Can you follow one won deal back to its first touch without asking anyone?
- Is ad spend reported next to closed revenue, or only next to leads?