Every marketer believes in wear-out. The idea is intuitive enough that it rarely gets examined: run an advertisement long enough and people stop noticing it, so you retire the thing before it dies of exposure and commission something fresh. Whole budgets are built on that assumption, and so are a good many agency relationships.
The evidence for it is remarkably thin. When Analytic Partners went looking, examining more than 50,000 advertisements that had been taken off air and replaced, it found fourteen where audiences had genuinely tired of the work. Fourteen out of fifty thousand, which means that almost every campaign killed that year was killed while it was still earning.
The test that should have settled it
System1 came at the same question from a cleverer angle. The firm keeps a database of tens of thousands of video advertisements, each scored for its potential to build long-term brand growth, and most of them are tested within days of first airing. Some, though, arrived late. A handful were tested as much as nineteen years after they originally ran, by which point audiences had had two decades to grow sick of them.
If wear-out were real, those late-tested advertisements should have scored badly, because familiarity is precisely what the theory says destroys effectiveness. They did not. Average scores held steady between 2 and 2.4 on System1’s star rating whether an advertisement was tested the week it launched or long after everyone had supposedly stopped looking.
Sky Media reached a similar conclusion from the media side, finding across more than a hundred campaigns that wear-out began somewhere around fourteen or fifteen exposures, far beyond the frequency caps most planners still work to. Les Binet and Sarah Carter have described Knorr campaigns running in Scotland that target consumers had seen close to a hundred times and still responded warmly to. Carter’s summary of it is the whole argument compressed into four words: consistency leads to popularity.
So why does everything get replaced?
Two reasons, and neither of them is the audience.
The first is proximity. A marketing team lives inside a campaign for months before a single frame reaches the public, sitting through the strategy, the scripts, the edits, the pre-tests and the internal presentations, until the work feels worn through long before anybody has seen it. Mark Ritson’s standing correction applies exactly here: you are not your customer. The person you are advertising to is busy, distracted, largely indifferent to your category, and sees a fraction of your media at best. What feels like the hundredth exposure to you is often the third to them.
The second reason is career structure, and it is the more expensive of the two. A new chief marketing officer arrives with a short tenure and an obligation to make a visible mark, and nothing is more visible than a new campaign with a new agency. Killing an existing platform is legible to a board in a way that patiently extending one is not. The incentive runs against the evidence, which is why the evidence keeps losing.
What repetition actually buys
None of this would matter much if running a campaign for longer were merely cheaper. It turns out to be more effective as well, and the clearest evidence sits in a body of work Orlando Wood began at System1 nearly a decade ago.
Wood, who is chief innovation officer at System1 and an honorary fellow of the IPA, coined the term fluent device to describe the recurring characters and scenarios that hold a long-running campaign together. The Geico Gecko is one. So are the Compare the Market meerkats, Colonel Sanders, the Oxo Mum and the particular shape of a Specsavers joke that you can identify as a Specsavers advertisement before the logo appears. Working with the IPA Effectiveness Databank, System1 tagged more than 300 campaigns going back to 1992 and had Peter Field measure them against commercial outcomes.
Campaigns built on a fluent device were 37 percent more likely to grow market share and 30 percent more likely to grow profit than campaigns without one, and on the measure that matters most to a finance director, they were 73 percent more likely to report a large profit gain. Wood’s explanation is that these devices build memory structures that keep working long after a campaign ends, which is why a character can remain instantly recognisable decades later.
And here is the awkward part. Over the same period that the evidence accumulated, the practice collapsed. Fluent devices appeared in 41 percent of IPA Effectiveness Awards entries in 1992 and in around 12 percent by the time System1 measured it. The industry has been walking away from its most reliable technique while publishing the research that proves it works.
The sequel is the proof
This magazine made a version of the same argument in August, writing about brand collaborations, where the surest sign that a partnership succeeded is that it came back for a second run. Two companies agreeing with money to repeat an association is a more honest verdict than any announcement about it. e.l.f. and Liquid Death restocked a lip balm that sold out in 45 minutes. Reese’s and Oreo ran limited editions until fan demand justified a permanent product.
The logic travels beyond collaborations. A campaign that returns is a campaign somebody chose to fund twice, having seen what it did the first time, and that decision carries information no launch ever can. Launches are promises. Sequels are receipts.
The honest counter-case
Repetition is not a virtue in itself, and the research says so plainly. Kantar’s caution is worth taking seriously: a weak advertisement run repeatedly does not become strong, it becomes an expensive irritant, and there are real warning signs to watch for in tracking data when audiences begin to enjoy a campaign less rather than more.
There is a quality threshold underneath all of this. System1’s wider work suggests only a small fraction of global advertising generates enough emotional response to drive growth at all, which means the case for running work longer is really a case for running good work longer. Extending a mediocre platform simply distributes mediocrity more efficiently.
Categories also move. A campaign built around a product benefit that competitors have since matched, or a cultural reference that has curdled, needs retiring regardless of how well it once tested. The question is never whether a campaign is old. It is whether it is still true.
Repeat the idea, refresh the execution
The practical distinction that separates the brands doing this well is not between changing and staying the same. It is between the layers that should hold and the layers that should move.
Specsavers has run one joke for over twenty years and produced hundreds of different advertisements inside it. Compare the Market built a business on a meerkat who has since acquired a backstory, a family and a range of merchandise. Neither brand repeated itself in the sense of showing the same film twice; both repeated the thing that made the first film work, which is a harder discipline and a cheaper one.
The test to apply before scrapping a platform is straightforward enough to run in a meeting. Ask whether the audience has stopped responding, and produce the tracking data that shows it. If the honest answer is that the team is bored, that is not a finding about the market. It is a finding about the team, and it should not cost the company a new campaign.
Sources: Analytic Partners, analysis of replaced advertising creative (2020) · System1 Group, ad wear-out analysis of its video database, reported by Marketing Week · System1 and IPA Effectiveness Databank, fluent devices research with Peter Field, tagging 300-plus campaigns from 1992 · Orlando Wood, Lemon (IPA, 2019) and Look Out (IPA, 2021) · Sky Media, frequency and wear-out study across 100-plus campaigns · Les Binet and Sarah Carter, on Knorr in Scotland, Uncensored CMO · Kantar, on creative quality and in-market tracking · BBF Monthly, “Why Every Brand Suddenly Wants to Wear Another Brand’s Logo” (August 2026).
Author: Tanvir Ahmed
