What is frequency capping in programmatic advertising?
To boost reach within their target audience and avoid overexposure, Media Traders mainly use DSP frequency caps, generally setting a cap of two or three so as not to βoverexposeβ their audience.
Here are the basics. Reach estimates how many distinct people a campaign has reached. Average frequency is the number of impressions divided by estimated unique reach. Frequency capping is the buying rule designed to limit how often ads can be served to a user within a period β for instance, twice per day or three times over a campaign.
For example, a campaign with 100,000 impressions and an estimated reach of 50,000 users has an average frequency of 2 (illustrative arithmetic, not campaign performance data). That does not mean every user was shown exactly two ads: some might have seen one, others considerably more.
DSP frequency limits are not perfectly equivalent to human-level exposure. Counting and deduplication depend on identifiers, environments, DSP settings and modelling. See Google's Display & Video 360 documentation on frequency caps.
Why isn't a frequency cap of two or three always the right answer?
But what's the rationale behind a cap of 2?
When you consider that people are exposed to hundreds of advertising messages every day, is that really optimal? (Especially if it doesn't account for the level of Attention π)
Two impressions on a prominent, attention-grabbing placement aren't the same as two fleeting impressions in a crowded page. A short, familiar slogan doesn't have the same job as a complex message for a brand nobody knows yet.
The key question isn't βHow low should we set the cap?β It's βHow much meaningful repetition does this particular message need?β (That depends on objectives, creative, audience, format, campaign duration and how attention is measured.)
We've written about the difference in why viewability alone is not a measure of advertising attention.
What did 1990sβ2000s French TV adverts understand about repetition? πΊπΌ
If you watched TV in France in the 1990sβ2000s, we've got a few counterexamples that should hit home.
Try finishing these slogans from memory:
- βSi Juvabien, c'est [...]β?
- βMercurochrome le pansement des [...]β?
You haven't seen these adverts for over 10 years. And yet you might remember them as if it were yesterday.
Why?
Because these French TV adverts were perfect examples of hammering a message into your head π€―:
- 1οΈβ£ Twice in the creative: in their 10-second spot, they repeated the slogan twice.
- 2οΈβ£ Three times in a row: to make absolutely sure it sank in, they ran the spot three times consecutively in a 30-second advertising slot.
- 3οΈβ£ Twice in the ad break: to make REALLY sure it sank in, that 30-second sequence was played twice in the same ad break.
And back then, there were no smartphones to distract your Attention from TV ads! π«π±ππΊ β β πΊπ
The result: 20 years later, the message is still firmly lodged in your brain. π§
These are deliberately nostalgic examples from the original post, not a controlled experiment or evidence of a universal optimum. And, of course, people could be distracted by plenty of things before smartphones existed.
Reach, frequency and attention: which metric answers which question?
Programmatic buyers often manage the first two and forget the third. Yet all three describe something different:
| Metric | What it tells you | What it cannot tell you alone |
|---|---|---|
| Reach | How many distinct people were exposed to your campaign? | Whether they remember the message. |
| Frequency | How often, on average, were reached people served an impression? | Whether exposure was attentive β or equally distributed. |
| Attention | How prominent or engaging was an opportunity to see the advert, based on the available measurement method? | Whether the campaign caused brand lift or conversions. |
A higher frequency isn't automatically better, and a cap of two isn't automatically more respectful. The goal is to avoid both underexposure and wasted or irritating repetition.
For another French Open Web perspective, see why the Open Web has an attention problem.
How can Media Traders use controlled remessaging? π
If you want to βforceβ repetition of your message, you can use a controlled remessaging tactic. The idea is simple: first maximise reach, then give part of that audience a deliberate additional exposure, with its own budget and pressure rules.
- 1οΈβ£ Cap initial prospecting. Across the lines in your campaign, apply a frequency cap of 1 (to maximise initial reach). Specify the time window and check caps at campaign level too: one impression per line does not necessarily mean one impression per person across all lines.
- 2οΈβ£ Build an exposed audience. Add a retargeting pixel to the creatives where your DSP, creative setup and legally valid consent allow it. Its job is to identify users previously exposed to the message (other consent-compliant mechanisms may be used if supported).
- 3οΈβ£ Set up a dedicated remessaging line. Allocate a separate budget to that audience, targeting it exclusively, with a cap of 1 per day, for example, to maximise planned repetition without simply opening the floodgates.
That creates two distinct jobs: a first wave focused on incremental reach, followed by a second wave focused on reinforcing the message. (It is a proposed activation method, not a guarantee of uplift; retargeting availability and cross-line capping differ between DSPs.)
Crucially, check the overall frequency at the campaign or insertion-order level, not just the line level. Otherwise, your perfectly sensible caps can add up to a less sensible user experience. DV360's documentation explains how its campaign, insertion-order and line-item limits interact.
Which consent and GDPR safeguards matter for remessaging?
Be careful: apply the pixel only to creatives served on inventory where the required consent is valid, and delete the audience pools once the campaign has ended in accordance with your data-retention obligations.
In France, retargeting pixels and other advertising trackers will generally require prior consent. Simply having a campaign running or a contractual relationship with an SSP does not make that consent automatic. Check the purposes covered, your consent management platform (CMP), partner permissions, retention rules and withdrawal handling before activating the audience.
(Even some frequency-measurement cookies may require consent; no-consent traffic must not be silently enrolled in a retargeting pool.)
For the regulatory detail, consult the French CNIL's official guidance on cookies and other trackers. This is practical editorial guidance, not an assessment of a specific DSP's legal compliance.
How do you know whether extra frequency actually works?
Rather than treating β2β or β3β as an unchangeable law, start with a small, clearly measured comparison. Match audiences, creative and inventory as closely as possible, and look at the results across different exposure levels.
- Incremental reach: did the prospecting phase expand your audience?
- Frequency distribution: how many people saw the message once, twice, three times or more β not just the average?
- Attention and creative quality: were the extra impressions meaningful opportunities to notice the ad?
- Outcomes and cost: where measurement permits, did brand recall, consideration or conversions improve enough to justify the incremental spend?
Remember that DSP reach and frequency reports may be modelled and can behave differently across cookies, browsers and devices. Google's reach and frequency measurement guide explains the distinction between average frequency and frequency distribution.
And if you're analysing these patterns in campaign reports, you'll recognise the value of combining your DSP data with an Excel optimisation scorecard.
Bottom line: stop treating the cap as a magic number
Frequency is a central and unfairly underused metric in programmatic advertising. Minimising repeated exposure makes sense when it avoids waste. But sometimes, repetition is precisely what makes the message stick. π
Don't just set a cap of two because βthat's what we always doβ. Test the balance between reach, attention and effective repetition instead.
GreyFox Ads β the smallest programmatic advertising sales house in France. And proud of it. π¦
Sources and further reading
- Original GreyFox Ads LinkedIn post β the TV advert anecdotes and the three-step remessaging proposal.
- Google β Manage frequency with frequency caps (Display & Video 360) β campaign, insertion-order and line-item behaviour.
- Google β Measuring reach and frequency (Display & Video 360) β unique reach, frequency distribution and modelling.
- CNIL β Cookies and other trackers: frequently asked questions (French-language regulator guidance).
- GreyFox Ads: Why viewability alone is not enough and the Open Web's attention problem.