There is a question I heard from a colleague of mine recently who sat with me. We were discussing campaign performance and they asked, with complete sincerity: “Does the client even care about incremental revenue?”
I was surprised. Of course businesses care about incremental revenue. That’s the whole point of starting marketing at all. But the more I answer the question, the more I think it reveals something unsavory about how most companies actually operate. Not whether they care about scaling in theory, but whether the systems they’ve built can scale it. And if the system can’t measure something, the business stops doing anything about it, even if anyone claims to care.
The case that changed the way we think about this
The product brought us to develop paid social work. We have done a full account restructuring. Various creations. Increased video investment, which we know opens up placements and better engagement with audiences. With every measure we could use, things were moving in the right direction.
Then the data came back and the conversions reported dropped significantly.
This is the time when many agencies panic and start to reverse course. We didn’t, because we would have done parallel growth studies and they would have told a very different story: we were generating more business income. Field performance was strong. The mounting evidence was clear.
So we entered. What we found was a change in conversion plans that no one was talking about. Before the redesign, about 80% of conversions attributed were click-throughs, meaning someone clicked on an ad and then made a purchase. After that, it’s closer to 20%. The reason was simple: we were into video. Video influences people by watching, when they see something, they remember it, and come back a few days later, rather than clicking there and then. And the client’s data system worked on click first, so it couldn’t see what the video was doing.
The income was there. The measurement infrastructure could not account for it.
Why is this more important than a single account
If this is not caught, the next discussion would be about why performance has decreased and what to do about it. Reading from that data alone, the answer seems obvious: go back to video, rely on still images, prepare for post-click conversions. Reasonable conclusions, but wrong.
Even a “correct” temporary fix, which expands our portion of still images to capture the re-referenced change, isn’t really a fix. It’s a concession to a rating system that can’t see the full picture. You may get numbers in a client report while at the same time reducing the incremental value you generate. Growth tests done after that type of adjustment usually confirm exactly that.
A measurement model doesn’t just tell you how you’re doing. It determines what you prepare. And if the model is wrong, everything below is also wrong.
The problem of a single source of truth
Many businesses choose a measurement method, build their reporting infrastructure around it, and treat that output as objective reality. The reasons are understandable. Attribution tools are expensive to build, complex to modify, and focused on how functionality communicates internally. Going from first click to last click, or to a data-driven model, is not a technology transition you can make in an afternoon. It’s reshaping the way each channel justifies its budget.
But the result is that companies make important strategic decisions, budget allocations, channel mix choices, innovation strategies, based on a model that was often chosen for historical or organizational reasons rather than ideological ones.
The first click attribute makes sense when search dominates the funnel and every journey starts with a specific question. It’s a remnant in a world where the paid community operates at the top and middle of the funnel, where video influences purchasing decisions days before anyone types a search, where the customer journey goes through six touch points and none of them tell the whole story.
How much better it looks
The answer is not to find the “right” attribute model and switch to that. The assumption of a single model is the problem, not a specific model. The answer is to create a measurement structure that does not rely on any single signal as the ground truth. In doing so, four things work together.
- Run developmental studies often enough to direct, not as an occasional outlet.
- Use it modeling the media mixwhich compares each channel’s contribution to spend and results rather than tracking individual users, so you get views that don’t depend on cookies or pending clicks.
- It heals platform data as the best, most useful and fastest guide, but not the gospel.
- Then reconcile all three: see where they agree, and really pay attention to where they disagree. Disagreement, as was the case with this client, is often where the most useful information is hidden.
There is a trading post buried here as well. If the only language shared between agency and client is “specified change in reporting tool,” the conversation is tied up before it even begins. Any strategy that doesn’t show up cleanly in that tool becomes very difficult to defend, no matter how important it is. Good tricks are discontinued. Strategies that work are exchanged for strategies that look good on the dashboard. The discussion can be weekly discussions, or the work is gradually moving towards improving what is reflected in the reporting tool rather than what is driving growth.
The Bottom Line
My forum partner’s question, “does the client care about incremental revenue?”, was absurd. But the better question underneath, “are we having a conversation about growth?”, is worth asking more often.
Measurement is not a back-office problem. It is a strategic decision that shapes everything else. Businesses that do, that invest in understanding what actually drives growth rather than what appears in their reporting, are the ones that make better decisions. Everyone is preparing a number that may have little to do with the result they really want.
