How Many Impressions Does Brand Recall Take? Frequency and Exposure Science for Creator Campaigns

There is no single universal number of exposures that guarantees brand recall. Anyone who promises one is oversimplifying eighty years of contradictory advertising research. What the research actually gives a brand is something more useful than a magic number: a well-documented shape. Recall climbs with each additional exposure, but climbs fastest early and flattens later. Purchase intent and receptivity don't follow the same curve, and can decline even while recall keeps rising. Where an exposure lands matters as much as how many exposures there are. For a brand running a creator campaign, where each "exposure" is a single creator's video reaching part of that creator's own audience rather than one ad impression served identically to everyone, that shape has a direct, practical translation: spreading a message across multiple creators over a campaign window generally serves recall better than concentrating repetition inside one creator's audience.

This guide walks through where exposure-frequency theory actually comes from, what a rare piece of recent controlled testing found when it measured recall and purchase intent at specific exposure counts, why digital video behaves differently from television on this exact question, and what all of that implies for planning a creator or logo-placement campaign rather than a traditional media buy.

Estimates, not guarantees. Creator earnings and campaign results vary by niche, geography, and platform performance. Any figures cited are planning estimates or third-party data, not promises of income or campaign results.

Key facts at a glance

Table of contents

1. Why frequency matters at all: the forgetting curve
2. Krugman's three-exposure theory: what it actually claims
3. Why "three" never meant what the industry thought
4. The recency challenge: one exposure at the right time vs. many at the wrong time
5. What a rare controlled study found: recall up, purchase intent down
6. Digital video's different frequency ceiling
7. An exposure isn't an ad slot: translating this to creator campaigns
8. Creative variation: the practitioner's answer to wear-out
9. How LogoImpress's campaign structure interacts with exposure
10. Common mistakes brands make with frequency in creator campaigns
11. Setting an exposure plan for a creator campaign
12. FAQ
13. Sources

Why frequency matters at all: the forgetting curve

Before asking how many exposures are enough, it's worth being precise about why exposure count matters at all. The underlying mechanism is memory decay, not persuasion in the abstract. German psychologist Hermann Ebbinghaus's 1885 research on his own memorization of nonsense syllables produced the "forgetting curve": the finding that newly learned information is lost fastest immediately after learning, with the rate of loss slowing over time (Ebbinghaus, 1885). Ebbinghaus's own experiments didn't test spaced review directly, but later memory researchers built directly on his forgetting-curve data to show that this decay can be measurably softened by reviewing the same information again at intervals rather than only once. That combined finding, that repetition at intervals fights decay more effectively than either a single exposure or a single block of repetition, is the psychological foundation underneath every advertising frequency theory that followed, whether the theory argues for three exposures, one exposure at the right time, or six.

The practical implication for a brand is that "frequency" is never really the goal. Memory retention is the goal, and frequency is one lever for achieving it (timing relative to a purchase decision is the other, covered below). That distinction matters because the two levers don't always point the same direction, which is exactly what the rest of this article works through.

Krugman's three-exposure theory: what it actually claims

The most widely cited number in advertising frequency is three, and it comes from Herbert Krugman's analysis published while he worked at General Electric, building on his broader 1965 study "The Impact of Television Advertising: Learning Without Involvement" (Public Opinion Quarterly). Krugman's argument was psychological rather than statistical: he proposed that advertising exposures create only three meaningfully different mental states, not an open-ended series of equally weighted repetitions.

• The first exposure triggers a "what is it?" response. The viewer is processing the ad as a novel stimulus, evaluating what's actually being shown before absorbing any persuasive content.
• The second exposure shifts to "what of it?" Now that the stimulus itself is familiar, the viewer starts to evaluate whether the message is personally relevant.
• The third exposure is, in Krugman's framing, the first one that functions as an actual reminder. The viewer has already processed novelty and relevance, so this exposure is free to do the job of reinforcing memory.

Krugman's own, frequently quoted conclusion was blunt about what happens after that: "there is no such thing as a fourth exposure psychologically; rather fours, fives, etc., are repeats of the third exposure effect" (Krugman, cited via effective frequency literature). In other words, Krugman wasn't arguing that three ad impressions is a magic operational target for a media plan. He was arguing that the psychological work of an exposure sequence completes by the third repetition, after which additional repetitions aren't doing meaningfully different cognitive work. That's a narrower and more specific claim than the "rule of three" shorthand it turned into across the industry, and it's worth holding onto that distinction before applying it to any specific campaign.

Why "three" never meant what the industry thought

The "rule of three" is one of the most misapplied findings in advertising history, and the misapplication happened almost immediately. Krugman's original claim was about the psychology of one viewer's own exposure sequence, a specific, narrow proposition about diminishing novelty, not a universal operating instruction for every campaign. Michael Naples's 1979 book for the Association of National Advertisers, Effective Frequency: The Relationship Between Frequency and Advertising Effectiveness, is frequently credited with popularizing "three-plus" as an industry planning benchmark. But Naples's own conclusion, buried under the number that made him famous, was that each brand should experiment to find its own proper frequency level rather than adopt a universal figure, a caveat the industry largely ignored in favor of the single number (summarized via the effective-frequency literature).

Later critics were blunter about the gap between the finding and its industry use. Media researchers writing in the years after Naples's book warned planners directly that they had taken away an "erroneous impression" that a national advertising body had "decreed three or more exposures to be the magic number" for every product category, a benchmark nobody had actually established with that level of universality. That same account of the debate lands on a correspondingly cautious current position (summarized via the effective-frequency literature): there is no single magic number, and effective frequency depends on the creative itself, the product category, how familiar the audience already is with the brand, how cluttered the surrounding media environment is, and whether the campaign is launching something new or maintaining an established name.

The practical lesson isn't that three is wrong. It's that "three" was never meant to travel as a portable constant across every brand, platform, and audience. Every specific figure in this article, including the ones from the more recent studies below, should be read the same way: as evidence about the shape of the recall/persuasion trade-off under the specific conditions each study tested, not as a number to import unmodified into an unrelated campaign.

The recency challenge: one exposure at the right time vs. many at the wrong time

Krugman's framework describes what happens within a sequence of exposures to one viewer. A different, later strand of research asked a different question entirely: does when an exposure happens matter more than how many exposures there are?

John Philip Jones's research, published as When Ads Work: New Proof That Advertising Triggers Sales, used single-source data, matching individual households' ad exposure to their actual purchases, and found that a single exposure occurring within the seven days before a purchase decision produced a far greater effect than what additional exposures further from that window added (summarized via OAAA's recency-theory coverage). That finding is the empirical seed of the recency-planning discipline Ephron built, described next.

Erwin Ephron built recency theory into a practical media-planning discipline from that finding. Since a marketer can rarely predict which specific week an individual consumer will be in the market to buy, the planning goal shifts from maximizing frequency to maximizing reach: getting an exposure in front of as many different people as possible in as many different weeks as possible, so that whichever week a given consumer happens to be ready to buy, an exposure is there waiting. Ephron's own framing was that media planning's job is to be present at the right moment for the right person, not to repeat a message at the same person as many times as budget allows.

Krugman and recency theory aren't strictly contradictory. Krugman describes the internal psychology of a single viewer's own exposure sequence, while recency theory describes how to allocate a finite budget across an entire audience with unknown, staggered purchase timing. But they do pull in different practical directions: Krugman's framework can be read as license to concentrate exposures on the same audience until the third hit lands, while recency theory argues that spreading exposures across more of the audience, more consistently over time, usually beats concentrating them on people you've already reached. That tension is precisely why later research had to actually measure what happens as exposure count climbs, rather than reason about it from theory alone.

What a rare controlled study found: recall up, purchase intent down

Controlled, single-variable testing of ad frequency is genuinely rare. Most of what the industry knows comes from aggregate marketing-mix modeling rather than a study that holds everything constant except exposure count. A 2023 study from Magna (IPG Mediabrands' media research unit) and the ad-tech company Nexxen, titled "It's All in the Delivery: How Repeating Ads Affect CTV Viewers, Brands and Platforms," did exactly that: 1,246 streaming (connected-TV) viewers were shown the same ad, creative from New Balance or Applebee's, either once, four times, or six times within a single one-hour viewing session, with recall and purchase intent measured afterward (Magna/Nexxen, July 2023, reported via Digiday; MNTN Research summary).

The results split sharply by metric:

• Recall climbed steadily with frequency: reported at roughly 64% after one exposure, 85% after four, and 92% after six.
• Purchase intent moved in the opposite direction: viewers who saw the ad six times showed roughly a 16% decline in purchase intent compared with viewers who saw it only once.
• Viewers exposed to the ad the most frequently were also more likely to describe it as "annoying" or "disruptive," the negative-sentiment finding that tracks with the purchase-intent decline.

This is the clearest empirical demonstration available that recall and persuasion are not the same outcome and don't move together past a certain point. A brand optimizing purely for "did they remember seeing it" can walk straight into a frequency level that's actively working against purchase intent. It's also a direct, measured illustration of the tension between Krugman's framework and recency theory's warning against over-concentrating frequency on the same audience.

What this study doesn't tell you. It's worth being precise about the study's actual scope before generalizing from it. It tested two specific brands (New Balance and Applebee's), compressed all exposures into a single one-hour viewing session rather than spreading them across a realistic campaign timeline, and measured connected-TV viewing rather than short-form social video. None of that invalidates the pattern it found: recall and purchase intent moving in opposite directions past a moderate exposure count is consistent with the broader theoretical tension between Krugman's and Ephron's frameworks described above. But the exact percentages (64%/85%/92% recall, 16% purchase-intent decline) are evidence from one controlled test under one set of conditions, not universal constants that transfer unchanged to a multi-week creator campaign on a different platform.

Digital video's different frequency ceiling

The CTV study above measured exposures compressed into a single hour-long session, a useful controlled test but a much higher concentration than most real campaigns deliver. At the level of an actual media plan spread across weeks, the frequency ceiling looks different by platform. A Google-commissioned Nielsen meta-analysis, pooling results from 15 U.S. consumer-packaged-goods marketing-mix-modeling studies that each measured both YouTube and television results, found that brands could raise their average weekly YouTube frequency from one exposure to three without losing ad effectiveness, a notably higher tolerance for repetition than the same analysis found for linear television, where returns diminished at a lower weekly frequency (Nielsen/Google, 2021).

Two things make this finding directly relevant to short-form creator content specifically, even though the underlying study covered pre-roll and in-stream YouTube ad formats rather than creator-native placements. First, it confirms that digital video audiences tolerate materially more weekly frequency than a traditional broadcast audience before wear-out sets in, likely because digital platforms deliver more varied surrounding content and viewing contexts than a single TV channel's programming block. Second, it reinforces that frequency tolerance is context- and platform-specific rather than a fixed constant across formats. A brand shouldn't import a TV-era frequency cap wholesale into a YouTube Shorts or Instagram Reels campaign plan without adjusting for the format.

The table below lines up what each piece of research actually measured, since conflating them is the single easiest way to misapply this literature to a creator campaign:

Reading across the row, notice that none of these four is actually measuring the same thing: one is a theory, one is purchase-timing data, one is aggregate weekly modeling, and one is a compressed single-session experiment. Treating any single row as "the" frequency answer for a creator campaign means importing a finding from a materially different measurement context.

An exposure isn't an ad slot: translating this to creator campaigns

Every frequency theory above was developed for a world where an "exposure" means the same ad, served the same way, to the same measurable audience: a TV spot, a pre-roll ad, a banner impression. A creator or logo-placement campaign breaks that assumption in a specific, important way. One exposure is one creator's video, reaching whatever portion of that specific creator's audience actually watches it. There is no single, uniform "the ad" being repeated identically. There's a roster of different creators, each with a different audience, each producing a differently framed piece of content that happens to carry the same brand placement.

That structural difference changes how the frequency research above should actually be applied:

• "Frequency" within a single creator's audience works close to the classical model. A viewer who follows one creator and sees that creator's videos repeatedly over a campaign window is accumulating exposures the way Krugman's or the CTV study's framework describes, with the same risk of the recall-up-but-receptivity-down pattern if the same placement runs too many times on one creator's channel.
• "Reach" across many creators' distinct audiences is structurally closer to recency theory's actual prescription. A brand exposing many different people, across many different creators' follower bases, to the placement, spread across a campaign window, rather than repeatedly hitting the same audience. This is the natural shape a multi-creator logo-placement campaign already takes, even without a brand deliberately planning around exposure theory.

The practical takeaway: a brand running a single-creator sponsorship is the context where the classic frequency research applies most directly, and where over-repetition risk is most concrete. The same audience, watching the same creator, seeing the same brand mention on every video, risks the recall-up-persuasion-down pattern within weeks. A brand running a roster campaign across many mid-tier creators is, by the nature of the format, already closer to what recency theory recommends: broad reach spread across many distinct audiences and many different weeks, rather than concentrated repetition on one audience. Neither structure is automatically better; they're suited to different goals, echoing the same memory-versus-persuasion distinction that recurs across placement and disclosure research generally, including the distinction between passive and active placement formats covered in our research-backed look at passive brand visibility vs. active creator integration.

Illustrating the difference. Take two hypothetical campaigns of equal size and budget, purely to illustrate the structural point above rather than as a benchmark either would actually hit. A single-creator campaign concentrates its entire budget on one creator's audience, so a follower who watches most of that creator's output over a 30-day window might accumulate five or six exposures to the same brand mark, squarely inside the range the CTV study found recall still climbing but purchase intent already falling. A ten-creator roster campaign spreads the identical budget across ten distinct audiences; most individual viewers in that combined audience see the placement once, from whichever single creator they happen to follow, with only the (typically small) overlap between those audiences accumulating more than one exposure. The roster structure doesn't guarantee better results (that still depends on the campaign's actual goal), but it mechanically produces a frequency distribution much closer to recency theory's "reach, not repetition" prescription than a single-creator deal does, without a brand having to plan for that outcome deliberately.

Creative variation: the practitioner's answer to wear-out

If Krugman's framework is right that a fresh exposure's psychological value comes from resolving novelty ("what is it?") and relevance ("what of it?"), then the most direct lever for slowing the recall-up-receptivity-down pattern the 2023 CTV study documented is straightforward: change what the viewer is actually looking at. A viewer who has already resolved novelty and relevance for one specific creative execution starts a new novelty-resolution cycle when a genuinely different execution appears carrying the same underlying message. That's the logic behind the common industry practice of rotating multiple creative variations through a campaign rather than running one execution on a fixed loop.

For a creator or logo-placement campaign, that logic translates into concrete, low-effort options a brand already has available inside the format itself. Alternating between a static logo and an animated GIF overlay across different videos from the same creator, varying which video moments carry the placement, or simply relying on the fact that a multi-creator roster campaign inherently varies the surrounding content (a different creator, a different video, a different framing) around an otherwise consistent brand mark. That's a form of creative variation a single repeated TV or pre-roll spot doesn't get for free. None of this is a documented, controlled finding the way the CTV study above is. It's a direct application of Krugman's own psychological mechanism to the practical reality that most brands can't commission an entirely new ad execution every time wear-out risk appears, and creator campaigns happen to have a structural advantage here that traditional media buys don't.

How LogoImpress's campaign structure interacts with exposure

LogoImpress campaigns run over a defined window, typically 30 days, during which a brand's static logo, animated GIF, or pinned-link placement runs across the creators matched to that campaign, rather than being repeated indefinitely on a fixed schedule; our complete guide to logo placement sponsorships covers the campaign-setup mechanics end to end. Two structural features of that model connect directly to the exposure research above:

• Short-form video's view curve is heavily front-loaded. A large share of a given video's lifetime views typically arrive within the first two days of publication, tapering sharply afterward. That means most of a single placement's "exposure" to any one viewer concentrates early, rather than spreading evenly across the full 30-day window the way a scheduled TV flight would. A brand spreading a campaign across a roster of creators publishing on a staggered schedule is, in effect, creating a rolling sequence of front-loaded exposure spikes across different audiences over the window, closer to recency theory's "reach spread over time" than to a single concentrated frequency burst.
• Views and clicks are verified through platform APIs (the YouTube Data API v3, and the Instagram Graph API where applicable) rather than projected from a frequency or reach model. That means a brand evaluating whether its campaign hit a reasonable exposure level can check actual per-creator, per-video view counts rather than relying on an assumed frequency distribution, a materially more precise input than the aggregate marketing-mix modeling most classical frequency research is built on.

Neither of these facts tells a brand a specific "right" number of creators or exposures to buy; that depends on the campaign's actual goal, as the research above establishes. What they do mean is that a brand can plan a roster size and campaign length with the recall-vs-receptivity trade-off in mind, and then verify what frequency was actually delivered per audience segment after the fact, rather than guessing.

Common mistakes brands make with frequency in creator campaigns

1. Treating "three exposures" as a literal target to hit on every viewer. Krugman's framework describes a psychological sequence within one viewer's own exposure history, not an instruction to serve exactly three impressions per person, and even the "three" figure predates digital video's different frequency tolerance entirely.
2. Optimizing purely for recall. The 2023 CTV study is a direct warning: recall can keep climbing well past the point where purchase intent and brand sentiment start declining. A campaign report that only tracks "did people remember seeing it" can miss real receptivity damage.
3. Concentrating a whole campaign's budget on one creator to maximize frequency with one audience, when the goal is broad awareness. Recency theory and the roster-based structure of a logo-placement campaign both argue against that for this particular goal.
4. Assuming TV-era frequency caps apply to short-form digital video. The Nielsen/Google meta-analysis specifically found YouTube tolerating a higher weekly frequency than television before losing effectiveness. Importing a broadcast-era cap under-delivers relative to what digital video audiences will tolerate.
5. Never checking actual delivered frequency per audience segment. Verified, API-sourced view data makes it possible to know what frequency a campaign actually achieved per creator's audience. Skipping that check means flying blind on the exact variable this entire body of research is about.

Setting an exposure plan for a creator campaign

1. Decide whether the goal is memory reinforcement or persuasion, since the research above shows those two outcomes don't move together past a moderate exposure count. A brand introducing a new product leans toward broader reach across more creators; a brand reinforcing an already-familiar name can tolerate more repetition within fewer creators' audiences.
2. Default toward reach across a creator roster over repetition within one creator's audience, consistent with recency theory's core finding, unless the campaign's specific goal calls for concentrated frequency.
3. Watch for the recall-vs-receptivity split within any single creator relationship. If the same brand placement is running repeatedly on one creator's channel over the campaign window, that's the scenario closest to the CTV study's six-exposure condition, and worth checking sentiment or engagement trends for wear-out signs.
4. Use the campaign's front-loaded view curve deliberately. Staggering different creators' publish dates across the campaign window creates a rolling sequence of fresh exposure spikes to different audiences, rather than one concentrated burst.
5. Check actual delivered frequency against verified view data once the campaign runs, rather than assuming a frequency model held. The classical research this article covers was mostly built on aggregate modeling precisely because per-person delivered frequency used to be hard to measure directly; a platform-API-verified campaign doesn't have that excuse. Because campaigns are priced against verified delivery rather than a flat fee (the mechanics of which our guide to pay-per-view creator sponsorships covers in full), a brand can also see directly how its budget is distributing across a roster, not just whether the campaign is broadly "working."

FAQ

How many times does someone need to see an ad to remember the brand?

There's no single universal number. Classical theory (Krugman) argues the psychological work of an exposure sequence largely completes by the third exposure; a 2023 controlled study found recall continuing to climb through six exposures in a single session (64% at one exposure to 92% at six), but purchase intent fell over that same range, so "more exposures for more recall" isn't automatically the right goal.

What is Krugman's three-exposure theory?

Herbert Krugman's framework, from research published in 1965, proposes that an ad exposure sequence moves through three distinct psychological stages: a first exposure asking "what is it?", a second asking "what of it?", and a third that functions as the first true reminder. Krugman argued that exposures beyond the third largely repeat the third exposure's effect rather than doing new psychological work.

What is recency theory and how is it different from frequency theory?

Recency theory, associated with Erwin Ephron and built on John Philip Jones's research, argues that a single ad exposure in the week before a purchase decision has a disproportionately large effect compared with additional exposures further from the purchase, so media planning should prioritize reaching more people across more weeks (reach) over repeating exposures on the same people (frequency).

Does more ad frequency always hurt purchase intent?

Not always, but a 2023 Magna/Nexxen study found a clear pattern in one controlled test: purchase intent declined about 16% between one exposure and six exposures in the same viewing session, even as recall rose sharply over that same range. That's evidence that recall and purchase intent can move in opposite directions at higher frequency.

Does frequency work the same way on YouTube as on TV?

No. A Google-commissioned Nielsen meta-analysis of 15 U.S. CPG marketing-mix-modeling studies found brands could raise weekly YouTube frequency from one to three exposures without losing ad effectiveness, a higher tolerance than the same analysis found for linear television at a comparable weekly frequency.

What counts as one "exposure" in a creator or logo-placement campaign?

One creator's video reaching the portion of that creator's audience that actually watches it, not a uniform ad impression served identically to everyone. A viewer who follows one creator accumulates exposures the way classical frequency research describes; a campaign spread across many creators is closer to reaching many different audiences once each, which is closer to what recency theory recommends for broad awareness goals.

Is it better to run a campaign with one creator repeatedly or many creators once each?

It depends on the goal. Repetition within one creator's audience is closer to the classical frequency model and carries more risk of the recall-up-but-receptivity-down pattern found in the CTV study. Spreading a campaign across many creators' distinct audiences is structurally closer to recency theory's reach-focused recommendation, which tends to suit broad awareness goals better.

How does a campaign's length affect how exposure accumulates?

Short-form video view counts are typically heavily front-loaded, with most of a video's lifetime views arriving in the first couple of days after publication. Staggering different creators' publish dates across a campaign window creates a rolling series of fresh exposure spikes to different audiences rather than one concentrated burst, which changes how frequency actually accumulates compared with a traditional, evenly scheduled media flight.

How can a brand check what frequency its campaign actually delivered?

By reviewing verified, platform-API-sourced view and click data per creator rather than relying on an assumed frequency model. The classical frequency research summarized here was largely built on aggregate marketing-mix modeling because per-person delivered frequency used to be difficult to measure directly; campaigns tracked through official platform APIs don't have that same limitation.

Sources

• Effective frequency — Wikipedia, summarizing Herbert Krugman's 1965 research (Public Opinion Quarterly), accessed September 2026
• The "Magic of Three" Disputed — iMedia Connection, summarizing Michael Naples's 1979 ANA book Effective Frequency and subsequent industry criticism, accessed September 2026 (page unreachable via automated fetch at time of writing; cited from indexed summary)
• Forgetting curve — Wikipedia, summarizing Hermann Ebbinghaus's 1885 memory-decay research, accessed September 2026
• Applying Recency Theory during COVID-19 — Out of Home Advertising Association of America, summarizing Erwin Ephron's recency theory and John Philip Jones's research
• Video advertising ROI drivers on YouTube Ads — Think with Google, citing the Google-commissioned Nielsen marketing-mix-modeling meta-analysis, 2021
• Ad overexposure on CTV hurts streamers as much as brands — Digiday, reporting Magna/Nexxen's "It's All in the Delivery" study, July 2023
• Enough (Ads) Is Enough: Avoiding CTV Advertising Over-Exposure — MNTN Research, summarizing the same Magna/Nexxen study

Tell us your campaign goals — we'll help you plan the right reach and frequency mix