Ask most marketing organizations to describe their performance, and the answer arrives organized by channel: how did paid search perform, how did email do this quarter, what did the social team ship, how is the website converting. This is not a failure of imagination. It is simply how marketing organizations are structured to observe themselves — by function, by team, by budget line — because that is how the work is planned, staffed, and reported.

No customer experiences marketing this way. A customer does not encounter “the email channel” and then, separately, “the paid search channel,” evaluating each on its own terms. They encounter an ad, then perhaps a search result days later, then an email that assumes context from neither, then a conversation with a salesperson who has seen none of it — a sequence, organized entirely by their own path through it, with no awareness of or interest in which internal team produced which piece.

This is worth distinguishing carefully from a nearby but different problem HMG has examined elsewhere: individual channels performing well while the overall system underperforms. That problem is about aggregate outcomes — whether local wins add up to a system-level win. This is a different and prior question, about frame rather than aggregation: even before asking whether the system is winning, what does the marketing actually look like when it is observed by channel versus observed by customer sequence? These are not the same exercise, because channel view and customer view are not simply two different levels of zoom on the same object. They are organized around different units entirely — one around internal function, one around individual lived experience — and a marketing effort can be legible and well-performing in one frame while being confusing or contradictory in the other, independent of whether any individual channel is winning or losing.

Consider what channel view typically makes visible and invisible. It makes visible: relative efficiency between channels, which one is producing results at the lowest cost, where budget might be reallocated for marginal gain. It makes invisible, almost by construction: what any single customer actually experienced, in what order, with what cumulative impression. A channel-view report can show every channel performing within acceptable bounds while having no way to represent whether a specific customer received a coherent, sensible sequence of contact or a disjointed, repetitive, or contradictory one — because “sequence experienced by one person” is not a unit the channel-view frame is built to observe. It was never designed to answer that question; it was designed to answer a different one.

Customer view inverts this. Following one customer’s actual path — the ad they saw, the page they landed on, the email that arrived, the support interaction that followed — makes visible exactly what channel view cannot: whether the experience, taken as a lived sequence, made sense. It also makes invisible something channel view captures easily: comparative efficiency across the organization’s total spend, which is not something a single customer’s path can reveal, because one path is not a sample large enough to judge aggregate channel performance.

Neither frame is more correct than the other, and neither should replace the other. The mistake is not choosing one frame — every organization has to organize its internal reporting somehow, and channel view is a reasonable and necessary way to manage budget and staffing. The mistake is assuming that because channel view looks healthy, the marketing is therefore healthy in the frame that actually matters to the person receiving it — an inference the channel-view frame was never built to support, because it cannot see the thing the inference is being made about.

This has a specific and testable implication. An organization that wants to know whether its marketing is coherent to the people receiving it cannot answer that question from channel-organized reporting alone, however sophisticated that reporting becomes — better channel-level dashboards do not create customer-sequence visibility, because they are answering a structurally different question. Answering the customer-view question requires deliberately following individual paths — not as a sampling technique to improve channel attribution, but as its own distinct exercise, undertaken because the channel frame, however well built, was never designed to see what a single person actually experienced.

The deeper point is that “how is our marketing doing” is not one question with one answer waiting to be measured more precisely. It is at least two different questions, organized around two different units of observation, and an organization can have a confident, well-supported answer to one while having no answer at all — not a wrong answer, but a genuine absence of visibility — to the other.