Why last-touch lies for B2B

Last-touch attribution credits whichever channel happened to be present when the deal closed. For e-commerce with a 48-hour purchase cycle, that's an honest enough approximation of cause. For B2B with a six-month sales cycle, it's a story about who answered the door, not who built the house.

Here's the failure mode in concrete terms. A buyer first encounters your brand through a webinar in February. They follow you on LinkedIn through March. They read three case studies in April. They request a demo in May after a conversation at a conference. By the time they convert in July, they're typing your URL directly into the browser.

Direct gets the credit. Direct didn't do the work.

The fix isn't multi-touch averaging. Linear attribution treats the conference and the throwaway retargeting click as equal contributors, which is also wrong — it just spreads the wrongness around.

The fix is a credit model that asks the right question: how much would conversion have dropped if this channel hadn't been there?That's Data-Driven's removal effect. It re-credits LinkedIn for the months it was quietly moving the deal forward. It punishes the retargeting click that was riding on existing intent. It does this without you tuning weights.

What this looks like in practice

On the seeded demo, last-touch credits Direct with 31% of revenue. Data-Driven says it's 11%. The 20-point gap migrates to LinkedIn, Events, and Partnerships — the channels that did the long work.

That gap is the most expensive number in your marketing report, because it's the gap between where your budget is going and where your budget is working.