The attribution report looks clean — a pie chart, a handful of channels, a percentage on each. It is easy to present and easy to budget against. And in a legal tech sale, the cleaner it looks, the less of the story it is telling.
The math doesn’t work at committee scale
6sense’s global buyer research puts buying groups at more than ten people on larger deals, each logging around sixteen interactions with the winning vendor over a cycle of about ten months. Take the shape rather than the decimals: ten people, sixteen touches each, most of a year. First-touch attribution hands all of that to one interaction. Last-touch hands it to a different one. Neither is describing the purchase.
It gets worse. The same research shows buyers typically do not contact a vendor until they are well past halfway through their decision. Most of what shaped the purchase — the peer conversations, the comparisons, the internal debate — happened before anything reached your CRM.
The touches you can see aren’t the ones that mattered
Refine Labs compared what attribution software said about 620 conversions with what buyers said when asked. The software credited search with 78% of high-intent leads. Buyers named search 12% of the time. It is one agency’s look at its own clients, but the mechanism is not in dispute: the channels a tracking script can see get the credit for demand created where it cannot see.
Podcasts do not leave tracking links. Neither do private communities, a hallway conversation at a legal ops conference, or a partner recommending you to a general counsel. HubSpot’s documentation says so plainly: its attribution reports leave out interactions that happen outside HubSpot without tracking URLs. The report is accurate about what it saw. It is silent about the rest, and silence shows up as zero — which is how channels that work end up with their budgets cut.
The cleaner the attribution chart, the more it left out.
Three changes that help
- Ask buyers directly. Add one open question to your demo form: “How did you hear about us?” The answers are messy, and they surface sources no tracking tool will ever report. Read them next to the platform data and treat the gap as information, not error.
- Count accounts, not contacts. Ten stakeholders are not ten journeys. They are one buying process. Reporting on contacts will keep misreading which accounts are actually moving.
-
Use a model that spreads the credit. HubSpot offers W-shaped and full-path models that distribute credit across the journey instead of pinning it to one end. Contact-level attribution comes with Marketing Hub Professional; deal and revenue attribution need Enterprise.
None of these changes produce a cleaner chart. They produce a more honest one — and in a market where most of the buying happens in rooms you are not in, being clear about what the report cannot see is worth more than precision about what it can.
A fifteen-minute test
Pull your last ten closed deals. For each one, compare what your attribution report credits with what the rep remembers the buyer saying on the first call. Count how often they agree. If it is fewer than half, your report is describing your tracking, not your pipeline. Then run the same comparison on your last ten lost deals. The gaps tend to be bigger there.
So the goal isn’t a perfect model. There isn’t one. It’s a report that shows its own blind spots instead of hiding them — because a number with its limits stated survives the budget meeting, and a clean number that falls apart under one question doesn’t.
If you find this content insightful, consider subscribing to our LinkedIn Newsletter and following our company page for our latest articles!
