The attribution bug that made paid look broke

The attribution bug that made paid look broke

A subscription B2B tool was about to cut its paid search budget. The analytics said paid converted at half the rate of direct, and the CFO had noticed.

The clue: the 'direct' channel had ballooned to 38% of conversions, suspiciously high for a brand nobody types in. Digging into the referral report revealed the payment processor's domain was sending users back un-tagged. Every paid buyer who completed checkout got re-labeled 'direct.'

They added the processor domain to the referral exclusion list. One config line.

Paid search's tracked conversion rate jumped from 1.9% to 3.6% overnight — no real behavior changed, the credit just landed correctly. The budget cut was cancelled.

Lesson: before you trust a channel comparison, audit where 'direct' is really coming from.