The scoreboard: attention, watch time, completion and high-intent engagement. Why "low attention channel" is a story the industry tells itself, and where the brand has to sit in the first three seconds.
Ask a marketer what effective brand advertising looks like on television and you will get a confident answer, reach against the buying audience, sufficient frequency, a campaign distinctive enough to be remembered and attributed. The playbook is seventy years old. Ask the same question about Meta and TikTok and the answer usually borrows from somewhere else, TV's reach and frequency, performance marketing's clicks, organic social's engagement counts. The largest media channel in the world, $306 billion a year, is still being judged with other channels' scorecards, and it deserves its own definition.
Start with the objection that usually ends this conversation, the idea that social is a low attention channel, good for reminders and retargeting but too shallow for brand building. It does not survive contact with the audience. People give these platforms 2 hours and 41 minutes a day, more than any other channel, and the platforms hold that attention because every scroll is tuned to keep it. The attention is there in industrial quantities.
What is low is the tolerance. The feed trains people to dispatch anything irrelevant inside two seconds, and advertising built for captive audiences gets dispatched. Low attention is a story the industry told itself because it is more comfortable than the truth, the audience pays enormous attention, just not to work that assumes they have no choice.
Television never had to answer for attention. An ad break reached the room, and the industry assumed the room watched, because nothing could be skipped and nobody could measure otherwise. Attention was priced in without ever being demonstrated.
Social removed both conditions. Every placement can be skipped, and every second of watching is measured, per person, per exposure. That does not make reach and frequency wrong, you still need to reach the category, light buyers most of all, and you still need repetition for memory. It makes them the entry ticket rather than the scoreboard. A reported frequency of six built on two-second skips is a frequency your brand tracker will never feel. On this channel you no longer have to assume anyone watched, you know, which means effectiveness has to be defined by what happens after the impression is served.
The same assumption hides in the research stack. A focus group or a forced-exposure brand lift study shows people the ad and then measures what moved, which answers whether the message works once watched, never whether anyone would have watched it. That distinction barely mattered on TV. In the feed the choice to watch is the first hurdle, so a result earned in a room where nobody could skip carries little relevance for an ad that loses the viewer at second one. The strongest message in the world scores zero with the audience it never held, and only in-flight measurement can tell you the difference.
Effective brand advertising on Meta and TikTok shows up in a specific set of numbers.
Attention up front. The share of people who stop in the first seconds, because nothing downstream exists without it.
Watch time and drop-off. Where the ad loses people, second by second. A drop-off curve is the most honest creative feedback in advertising, it tells you the exact moment the work stopped being worth watching.
View-through. The share who watch to completion, the difference between opening a story and telling one.
High-intent engagement. Saves, shares and comments, the actions that cost the viewer something. A share is a person putting their own name next to your brand.
And two familiar numbers barely make the list. A like is the cheapest gesture on the platform, and a click measures a different job, demand capture, the harvesting of intent that already exists. Brand advertising creates the demand, so judging it on clicks is measuring the wrong end of the process.
None of this is our invention. Watch behaviour is how the platforms themselves decide what deserves distribution, the discipline is taking the definition the platform already enforces and pointing it at the brand.
These are the leading indicators. Brand outcomes remain the verdict, and the two move together when the work is right. KitKat's portfolio delivered 15x ThruPlay completions and 2x purchase intent. Whiskas saw a 237% higher completed view rate alongside 5x brand favourability. The watch metrics and the brand metrics rose as one, which is the entire argument.
Ask two brand teams about attention and you can get opposite briefs in the same week. One wants pure attention and asks for the branding to be cut back because it slows the hook. The next says attention is overrated, lead with the logo card so the exposure counts even if nobody stays. Each brief is half right, and each one, followed to its end, fails. Chase attention alone and you win watch time that never touches the brand, entertainment with a sponsor. Lead with a logo slate and the feed reads ad in the first frame, the skip arrives before the message, and the exposure the plan counted is two seconds of nothing.
Our position sits where the brand science, the platform evidence and plain common sense all point. Karen Nelson-Field's research says attention is the gate, Jenni Romaniuk's says the brand has to be encoded or the attention deposits nothing, and both are visibly true in any ad account's data. So the brand is present in the first three seconds and throughout, natively. Not an intro tile, the product in hand, in the scene, doing what it does, the colours and codes carried inside content that still belongs in the feed. Attention without branding is entertainment, branding without attention is invisible, the work has to deliver both at once. The test is attention to the brand, not attention near it.
That is also what the AI layer is for. Every ad is pretested before a media dollar touches it, scored for whether the distinctive brand assets are actually present and early, whether the work is consistent with the brand's codes, whether it will pass compliance. The pretest exists to guarantee both halves of that bargain before a single impression is bought.
Creators became the dominant force on these platforms for one reason worth studying. They make content for their audience, not about themselves and not to impress other creators, and they live and die on watch metrics, so they optimise for them relentlessly. That is the whole method, and it transfers. We apply the same optimisation loop the platforms use to judge content, with the brand's campaign message and distinctive assets carried inside work that stays relatable and relevant to the person watching.
It is also why this cannot be produced the way TV is produced. A radio ad running on television would feel strange to everyone watching, and television production running in a feed feels the same, the polish itself signals ad before the message lands. Studio-produced brand advertising stands out on social, in the wrong direction. The craft is real, it is just a different craft, native grammar, real people, real environments, made by people who understand what the feed rewards.
Effective brand advertising on Meta and TikTok is advertising people choose to watch, carrying the brand from its first seconds in the platform's own grammar, measured on attention, watch time, completion and high-intent engagement, and validated by movement in the brand outcomes those signals predict. Every part of it is measurable today, which is something brand advertising has never been able to say before.
Media creates the opportunity to see your ads. Creative determines whether they work. On this channel, "work" finally has numbers attached.