A reported frequency of six built on two-second skips is an effective frequency near zero. Manage attentive frequency instead.
There is a conversation that happens in brand reviews everywhere, and it has a missing variable in the middle of it. The dashboard reports strong delivery, reach is on target, frequency is sitting at six across the flight, the plan is working. The brand tracker, meanwhile, has not moved in three waves. Two instruments, both accurate, describing what looks like the same activity. The dashboard is counting served impressions, the tracker is measuring memory. Between those two numbers sits attention, and nobody in the room has a report on it.
The gap matters because of how brand building physically works. Mental availability, the chance your brand surfaces when a buying moment arrives, is built by encoding memories, and durable memory scales steeply with attention. Karen Nelson-Field's research has put numbers on the relationship, low-attention exposure is worth a fraction of attended exposure, there is a threshold of a couple of seconds of active attention below which little durable memory forms, and the longer attention holds beyond it, the longer the memory persists. The qualification matters as much as the headline. A glance at a well-branded ad can refresh memory structures that already exist, the low-attention processing literature is clear on that, and it is part of why distinctive assets earn their keep. What a glance largely cannot do is build structures that do not exist yet, and for the light buyers growth depends on, almost nothing exists yet. For them, the impression is the opportunity for encoding, the attention is the encoding, and a serve that ends in a sub-two-second skip is an opportunity that almost nothing came through.
Consider what a reported frequency of six actually contains on a skippable medium. If each exposure lasted under two seconds before the swipe, the buyer experienced your brand six times for a combined ten seconds, fragmentary, barely attended, mostly below the threshold where new memory forms. Reported frequency, six. Attentive frequency, the number of exposures that cleared the threshold, possibly zero, plausibly one. The tracker is flat because, in the sense that matters most to memory, the campaign barely reached that person.
This is why the traditional frequency conversation, where teams debate caps to avoid wearing out the audience, has become almost charmingly misdirected. Accounts are carefully capping a number that may represent very little attention, protecting buyers from an overexposure most of them never experienced. The discipline is right, the variable is wrong. Stop capping served frequency, start managing attentive frequency, the number of times a buyer actually watched enough for something to encode. One attended exposure outworks many skipped ones, which means an account with lower reported frequency and higher completion can be building more brand than an account with double its delivery numbers.
TV never forced this distinction because it could not measure the absence of attention, so planning treated exposure as attention and mostly got away with it, the ad break was captive enough. The skip made the difference between served and seen visible, measurable and economically enormous. The dashboards just kept reporting the old number.
The practical correction is to re-read your own reporting with completion as the lens. In one account we measured, the brand's existing ad earned 728 completed views in its flight. The ads built as part of a creative system, running on the same budget in the same period, earned between 5,000 and 11,000 completions each. The instructive number is the rate behind those counts, the share of served impressions that converted into watched ads, because on a served-impressions report the ads look broadly similar, comparable reach, comparable frequency, comparable spend. Weighted for completion they are not similar at all, one of them barely happened.
Completion-weighted reach is the closest thing to a real reach figure the dashboard offers. Completion is a proxy, not attention itself, a video can play out under a wandering gaze, which is exactly why attention researchers warn against treating any platform metric as eyes-on measurement. But it is a far harder test than a served impression, it is reported per ad at no extra cost, and re-basing reach on it produces the first honest baseline most accounts have had. The number that comes out is usually a fraction of the reported one. It explains the flat tracker without mystery, it shows where budget is buying delivery but not viewing, and what it claims should then be checked the proper way, against brand lift and tracking, the measures the proxy exists to predict.
The auction economics compound the point, because the platforms price predicted outcomes and ad quality, and work that holds viewers improves both. Attentive creative wins cheaper delivery, so attentive reach builds more brand per impression and costs less per impression. The same creative improvements that move completion from a fraction of a percent to a multiple of it are simultaneously cutting the cost of every exposure. Media creates the opportunity to see your ads. Creative determines whether they work.
One caution belongs in any honest version of this argument. Optimising delivery toward completion has a selection bias of its own, the algorithm goes looking for the people most inclined to watch, who are not automatically the category's light buyers. An account can chase completion into its own hall of mirrors, brilliantly watched by the easily watchable. The corrective is the rest of the system, broad reach settings, a portfolio varied enough to give every segment something worth watching, and brand effects measured on the population that matters, not just the audience the metric finds congenial.
Run the account on the corrected variable and the weekly decisions become clearer.
Plan reach on completions, not serves. When the media plan promises reach, ask what completion-weighted reach it forecasts, and judge channels and formats on the cost of an attended exposure rather than a delivered one.
Watch per-ad completion decay as the fatigue signal. When an ad's completion rate starts falling for the audience it has been reaching, attentive frequency has peaked for that creative, and the next exposure is buying skips. That is the moment to remix or replace, a data-visible event, not a calendar guess.
Treat creative variety as the frequency solution. The way to earn a buyer's attention a fourth and fifth time is not the same ad again, it is a different route in, a different face, moment or entry point from a portfolio built to provide them. Attentive frequency accumulates across distinct executions long after a single execution has worn out its welcome.
And take the brand tracker seriously as the final auditor. When attentive reach rises, brand effects should follow, that is the chain the whole system runs on, and the tracker is where the proxy gets audited. Vaseline's 3x brand lift and purchase intent, measured in lift testing, arrived through exactly this mechanism, the same media spend converted into vastly more watched exposure.
None of this requires abandoning the dashboard, it requires knowing what the dashboard is counting. Served impressions are the opportunity, attention is the event. A reported frequency of six made of two-second skips is a tracker that will not move and a media budget quietly converting itself into very little. The brands that grow on this channel are the ones managing the number in the middle.