The pattern is familiar enough to be boring. Campaigns running across LinkedIn, Meta, search and content; the deal eventually arrives through branded search; and everything else in the mix looks like it didn't contribute. So it gets cut, deprioritised, or at least questioned at the next budget review.
Something put that brand search there, though. Someone saw an ad a week earlier, or a partner mentioned you, or a colleague filled in a form and forgot about it. You don't really know which of those it was.
Search behaviour is making this harder. There are more zero-click results and more instant answers than there were, and people have less reason to browse, so even when someone is actively searching they may not click anything until they're already close to deciding, and by then the one click you do get looks like it did the whole job in a single afternoon.
Paid sits in an awkward place
On LinkedIn you can get very close to the right people. In one account it drove around 60% of qualified leads while total spend came down over time, which is real contribution by any reasonable measure, and the sales team wanted to work those leads. As the client put it: "About 60% of all qualified leads we generate in ANZ come from LinkedIn."
It still doesn't tell the full story, because a good share of the impact surfaces somewhere else entirely. We've seen pipeline grow past twice its previous level in accounts where costs were rising across the board, which shouldn't happen if you read platform metrics literally. Probably a fair bit of that is upstream work landing later, somewhere the attribution model was never going to connect it back to the campaign that caused it.
SEO has the same problem from the other direction. One account had plenty of traffic that simply wasn't converting - broad queries and low intent, producing charts that looked healthy. Shifting toward higher-intent searches meant less traffic and more revenue. One client described the goal as "increasing revenues" rather than "likes or clicks which can look good on a chart, but may have no bottom line impact." Even then, a chunk of those conversions still arrive through brand.
Why the reporting falls apart
This is a modelling problem more than a tooling one. Attribution wants to assign a single clean source to something that was never clean.
Buying journeys get stitched together from things people saw, things they heard, internal conversations, previous experience, and timing. You only ever see part of that, and the part you see is whichever fragment happened to carry a tracking parameter.
The only approach that gets closer
The closest thing we've found is asking people directly, with proper onboarding questions - where did you hear about us, what triggered the search.
It isn't perfect. People forget or guess when you put them on the spot. It's still considerably better than pretending platform data is complete, and the quality of that input matters more than it looks: as one client said, "We rely on data to help us make informed decisions on where to spend our marketing budget."
The aim is attribution good enough to make better decisions on, which is a much lower bar than attribution claiming a precision it doesn't have. Feed the answers back into how campaigns get run and the picture gets closer to what actually happened.
Most SaaS marketing looks efficient until you follow it through to revenue. Brand search collects the credit and rarely does the work on its own, and most teams are still optimising the side of that gap they can see.
