Your Customers Don't Buy in Channels. Why Do We Measure Them That Way?
Your channel reports can each tell a convincing story yet still leave the key question unanswered: what should the business do next? I learned how far apart those two things can be on a direct-response TV schedule.
Photography by Erin Barr
The client had one station that seemingly outperformed every other on its schedule. That station also carried the client's most memorable toll-free number. Every other station used ordinary digits.
I had built attribution models for vanity numbers before, so I knew how they behave. People remember them. They call days or weeks later, and from places the number never ran. A memorable number works more like a brand than a tracking code, so I suspected the station's results were partly the number's.
I recommended running the memorable number everywhere. The client worried about losing station-level tracking, and they were partly right. Responses spread across a longer tail, and more calls arrived with no clear source. We worked through that with the attribution model, which showed the "best" station wasn't the best after all. We reoptimized the schedule.
Within a few months, the client had its best November in eight years. The most precise report on the schedule had been precisely wrong.
Eventually, that number went on everything, including billboards and radio, where tracking gets even muddier. It became part of the brand. What they gave up in precision, they gained in results, and the number became one of their most valuable assets.
The same problem followed marketing into URL tracking and digital attribution. People remember a brand, return later, and respond somewhere other than where their interest began. The tracking mechanism changes. The need to separate what created the demand from what captured the response remains.
The report measures what it can see
Most channel reports are accurate on their own terms. They record where a response landed: a click, a call, a promo code. But they can't see what created the demand.
Customers build familiarity, compare options, seek reassurance and decide whether to act across many encounters with a business. Advertising, reviews, referrals and sales conversations each play a part, and they build on one another. A report that assigns the sale to a single touchpoint turns that shared work into a contest for credit.
Most channel reports over-credit whatever sits closest to the sale, and most budgets follow them.
Incentives make it worse. When people are paid on a channel's results, the reports learn to argue for the channel. That isn't dishonesty; it's how incentives work. The budget ends up following the best-argued channel rather than the most valuable one.
I've been told a channel produced "zero-quality" leads one week and been asked to buy more of it the next. The leads hadn't changed. The question had.
The person who calls may not be the person who buys
Customer journeys don't follow our reporting calendar. Google's research on purchase behavior describes people looping through exploration and evaluation many times between the first trigger and the purchase.[1]
In high-stakes decisions, the person who responds is often not the person who buys. In healthcare and senior care, a family member or caregiver frequently does the research and makes the first call. A report built on one person, one path and one click will often misread that journey.
Some of the most honest journey data I've seen came from a sales or intake conversation, not a dashboard. Someone asked how the customer found the business and recorded the steps behind the decision. That told us more than a checkbox crediting the most recent channel, and the answers sat in a system marketing had never looked at.
The tools change. The questions don't.
Incrementality testing, triangulation and full-funnel modeling answer an old question better than we ever could. I use all of them, and I've built them into my own platform. But better measurement gives us stronger evidence. It doesn't set the strategy. That takes understanding what customers value, how the brand competes, how marketing interactions reinforce one another, and what the business can profitably deliver.
Those interactions depend on one another. A direct-mail response starts a relationship; a seminar or home appointment helps close the sale weeks later. First-touch attribution credits the mail and hides what closed the sale. Last-touch credits the appointment and hides what started the relationship. Google's own modeling guidance identifies a related problem: careless handling of branded search can understate what the upper funnel contributed.[2]
The same problem shows up when platforms and customer records disagree. I've seen Meta report nearly twice as many conversions as the CRM. Choosing the more flattering number doesn't resolve the difference. Establish what each system counts, reconcile the customer records, align the time periods and attribution windows, and test what the marketing added. The CRM records the business outcome. Attribution assigns credit. Neither, on its own, establishes what caused the sale.
The purpose is to improve the business, not to defend a channel. The evidence may point to wasted media spend, an ineffective offer, or a sales process that loses the demand marketing delivered. Follow it to the decision, wherever it leads.
The question underneath hasn't changed: what caused the customer to act, and what would have happened without it?
Measure what you need to decide
A useful review starts from the decision, not the report:
Can we acquire more valuable customers at an acceptable cost?
Are we creating new demand, or reaching the same ready buyers again?
Why are inquiries rising while completed appointments stay flat?
Where should the next dollar go, and what should we stop?
Platform reports, customer records, models, research and experiments each answer part of those questions. They will disagree. Resolving the disagreement means examining their definitions, time horizons and assumptions, and asking which decision is at stake.
Sometimes the right move gives up some measurement to gain performance. Then you use a model or a test to recover what you can. Precision is a tool. The decision is the goal.
Why I built Elevate Signal
I kept meeting the same hurdles: reports that graded their own channels, budgets nobody could test against the goal, plans never checked against what happened. I built Elevate Signal to turn conflicting evidence into a growth plan grounded in customers, margins and the business's real constraints. The platform does the analytical work. I bring the strategic judgment.
A growth plan needs more than a forecast. It needs a clear case for the investment, a way to test the assumptions, and a course of action when results change. Signal brings that evidence together so we can defend the recommendation, learn from what happens and adjust the plan.
Your customers don't experience your channel reports. They experience your business.
If your reports disagree, what decision is that holding up? I run Business Immersions that turn that disagreement into a plan you can defend. Let's talk.
Making it easier to respond matters. Giving people a reason to respond comes first. Next: "Cultivate and Harvest."
Sources
Google, How people decide what to buy lies in the messy middle of the purchase journey (2020). https://business.google.com/en-all/think/consumer-insights/navigating-purchase-behavior-and-decision-making/
Google Meridian, Full-funnel MMM: unifying upper-funnel and lower-funnel measurement. https://developers.google.com/meridian/docs/advanced-modeling/full-funnel
Erin Barr is the founder of Elevate Marketing and the builder of Elevate Signal. He helps consumer brands turn marketing investment into profitable growth, whether customers buy online, over the phone or through an appointment. His experience spans healthcare, elective medical, insurance, senior markets, ecommerce and consumer goods.
Comments