What is frequency capping, and how should a plan set it?
Part of Programmatic media planning
The short answer
Frequency capping is a setting that limits how many times the same user, browser or device is shown a campaign's ads in a set period, such as three impressions per week. The cap only works where the buying platform can recognize the viewer, so its scope, the event it counts and the identifier behind it all belong in the plan.
Key takeaways
- A complete cap names the event counted, the maximum, the time window, the campaign level it applies to and how viewers are recognized.
- Without third-party cookies, platforms fall back on first-party cookies, signed-in users, device IDs or modeling, and each surface documents its own method.
- Caps leak across devices, browsers, platforms and separate campaigns because each of those keeps its own counter.
- On a small site list a cap binds sooner, so plan for it with the list and the flight length before relaxing the cap.
What is frequency capping in programmatic advertising?
Definition
- Frequency cap
- A ceiling on how many times a campaign's ads are shown to one recognized viewer within a time window. Google Ads describes it as a feature that "limits the number of times your Display or Video ads appear to the same person."
A cap is a ceiling, not a target. It stops delivery to a viewer once the count is reached, and it does nothing to make sure each viewer gets that many ads. Platforms that aim for a number of exposures treat it as a separate setting. Google's Target frequency (opens in a new tab) for YouTube awareness campaigns, for example, is "an optimization solution" with a weekly goal of 2 to 7 for multi-format ads and 2 to 4 for non-skippable or skippable in-stream and bumper ads, or a monthly goal.
The definitions of reach and average frequency, and how to calculate them, sit in the reach and frequency guide. This page covers the cap itself: the fields, how it counts, where it leaks and how to choose a starting value.
| Setting | Google Ads Display | Google Ads Video | Display & Video 360 |
|---|---|---|---|
| Event counted | Impressions | Impressions, views or both | Each purchased impression, including YouTube & partners line items |
| Time windows | Day, week or month | Day, week, month or a combination | Minutes to one month; the help page says the maximum period is 30 days after February 28, 2025 |
| Levels | Campaign, ad group or ad | Campaign only | Campaign, insertion order or line item |
| Who is counted | Users, by third-party cookies by default, first-party cookies when those are unavailable | Signed-in users, or devices for signed-out users | Available identifiers plus modeling for mixed inventory |
Sources for the table: the Google Ads frequency capping definition (opens in a new tab), the Display & Video 360 frequency cap help page (opens in a new tab) and the FrequencyCap API reference (opens in a new tab), whose maxImpressions field is "the maximum number of times a user may be shown the same ad during this period" and whose timeUnitCount allows 1 to 6 days, 1 to 4 weeks or 1 month.
How do platforms cap frequency without third-party cookies?
A cap needs a stable way to recognize the same viewer twice. On the open web that used to be a third-party cookie, and a DSP learned it through ID syncing: in OpenRTB 2.6 (opens in a new tab), user.buyeruid is the buyer's ID for the user "expected to be derived from an ID sync", and device.ifa is the advertising ID from the device's operating system. When those fields are empty, the counter has nothing to attach to.
Browsers differ. WebKit's tracking prevention documentation (opens in a new tab) says its Intelligent Tracking Prevention "by default blocks all third-party cookies", partitions third-party storage "per first-party website", and deletes script-written storage "after 7 days of no user interaction." Chrome did not remove third-party cookies: on April 22, 2025 (opens in a new tab) Google said it would keep its current approach to third-party cookie choice, and on October 17, 2025 (opens in a new tab) it said it would retire Privacy Sandbox technologies including Topics, Protected Audience and the Attribution Reporting API. No later Google announcement changed this as of October 2026.
| Environment | Documented method | What happens without an identifier |
|---|---|---|
| Google Ads Display | Third-party cookies by default; first-party cookies "approximate impressions" when third-party cookies are unavailable | The cap is an approximation for that viewer |
| Google Ads Video (YouTube) | Signed-in users, or the device when signed out | A signed-out viewer on two devices is two counters |
| Display & Video 360, mixed inventory | "Modeling and available identifiers"; some line items use machine learning and historical data when cookies or device IDs are unavailable | The cap is modeled |
| Display & Video 360, connected TV | The device's advertising ID, or other supported types such as publisher provided IDs and session IDs | The impression is excluded by line items that use frequency caps |
What frequency cap should a campaign start with?
No industry body publishes a correct cap. The right starting value comes from the brief: how many exposures the message needs, how long the flight runs, how many creatives rotate and how sensitive the client is to repetition. What a planner can do is check whether a proposed cap fits the budget and the list before launch, using the platform's own forecast.
Five steps to a starting cap the client can review
- Write the exposure goal. Say whether the flight introduces a product, reminds existing customers or tells a story across several creatives, and how many exposures per week that implies.
- Estimate the impressions the budget buys. Divide the budget by the expected CPM and multiply by 1,000.
- Get a reach forecast for the list. Use the buying platform's forecast for the approved sites, markets and formats, and note what it counts: users, devices or browsers.
- Compare the implied frequency with the cap. Divide forecast impressions by forecast reach and by the number of weeks. If the result is above the weekly cap, the cap will hold delivery back.
- Record the decision. Write the cap, its level, its window, the identifier behind it, the forecast used and who may change it.
Step four catches a mismatch before launch. A cap of two per week on a list whose forecast audience is small, with a budget sized for a larger one, means the campaign cannot spend on plan. Better to see that in the forecast than in week two.
Why do frequency caps leak across devices and platforms?
Each counter only sees its own viewers. A person reading on a phone, a work laptop and a home tablet without signing in can be three separate counters on a platform that does not link them. A browser that partitions storage per site may look like a new viewer on each publisher. Two DSPs buying for the same brand keep two counters that never meet.
Scope causes leaks inside one platform too. In Display & Video 360 a line item cap applies only to that line item, while a campaign cap covers every insertion order beneath it. Programmatic guaranteed deals treat campaign caps as "best effort", and the system "may continue to serve impressions" when a deal needs more volume to meet its terms. The help page also notes that before a cap is met, several of an advertiser's ads can serve to one user on the same page at the same time. In Google Ads video campaigns, only in-stream and bumper views from other campaigns using the same video count toward a campaign's cap.
Frequency cap checklist before launch
- The cap is written as event, maximum, window and level, for example impressions, three, per week, campaign.
- The identifier behind the cap is named for each environment, and modeled caps are labeled as estimates.
- Every line item, insertion order and campaign that should share one counter sits under the capped level.
- Buys outside that counter, such as a second DSP or a direct deal, are listed with their own caps.
- Guaranteed deals and sponsorships that can exceed the cap are disclosed to the client.
- Connected TV line items with caps have been checked for inventory without a supported device ID.
- The person allowed to change the cap is named, and any change is logged with a reason.
How does a small site list affect frequency?
A short list of specialist sites can have a loyal, returning readership, which may be why the sites were chosen. It also means the same readers make up a large share of the available impressions, so a weekly cap fills faster and delivery slows once the regular readers have reached it. The cap is doing its job; the list and the budget are out of balance.
The fixes, in order: lengthen the flight so the budget spreads across more weeks, add sites that reach a different readership on related topics, then lower the budget or raise the cap with the client's agreement. Raising the cap first buys more repetition from the same readers, which is the opposite of what a capped plan asked for. The inclusion list guide covers sizing a list for a brief.
How can a planner widen a small site list without padding it?
What does setting a cap look like on one brief?
Fictional example
A ceiling fan retailer in Houston and Dallas
A fictional Texas ceiling fan and home lighting retailer plans a four week display and online video flight in the Houston and Dallas areas, in English and Spanish, beside home improvement guides, home design stories and local lifestyle sites. The client is worried about the same person seeing the ad all day.
The planner writes the cap as three impressions per week at campaign level in the DSP, so every line item shares one counter. The budget, divided by the expected CPM, buys a set number of impressions; the DSP's reach forecast for the 40 approved sites, divided into those impressions, implies about two exposures per forecast viewer per week, below the cap, so the cap should not hold delivery back. The plan notes that the forecast counts devices, not people.
A separate direct deal with a Houston news site sits outside the DSP counter, so the client note lists it with its own cap. The connected TV line item is checked for inventory without device IDs. No delivery or performance result is implied.
Frequently asked questions
Does changing the creative reset a frequency cap?
It depends on the level of the cap. A cap set on one ad or creative starts fresh for a new creative, while a cap at the line item, insertion order or campaign level keeps counting across creatives under it. Set the cap at the level that matches what the client wants limited.
Can a frequency cap be set across two DSPs?
Not with the DSPs' own settings, because each keeps its own counter and identifiers. A shared cap would need a common identifier and a system that sees both sets of impressions. In practice, give each DSP its own cap and report combined frequency only from a measurement source that de-duplicates both.
Is a frequency cap the same as target frequency?
No. A cap is a maximum and stops delivery to a viewer once reached. Google's Target frequency for YouTube awareness campaigns is an optimization that aims for a set number of exposures per person each week or month. One limits repetition; the other tries to reach a number.
Do frequency caps apply to guaranteed deals and sponsorships?
Not always. Display & Video 360 fully respects line item and insertion order caps on programmatic guaranteed deals but treats campaign caps as best effort, and it may keep serving when a deal needs volume to meet its terms. YouTube sponsorship and Masthead line items are not restricted by higher level caps.
Sources
- Frequency capping: Definition (opens in a new tab)Google Ads Help
- About Target frequency (opens in a new tab)Google Ads Help
- Manage frequency with frequency caps (opens in a new tab)Display & Video 360 Help
- FrequencyCap, Display & Video 360 API v4 (opens in a new tab)Google for Developers
- OpenRTB 2.6 specification (opens in a new tab)IAB Tech Lab
- Tracking Prevention in WebKit (opens in a new tab)WebKit
- Update on plans for Privacy Sandbox technologies (opens in a new tab)Google Privacy Sandbox
- Privacy Sandbox next steps (opens in a new tab)Google Privacy Sandbox
Researched and written by PlacementGraph Research. Read how we research, source and update these pages. Corrections: diana@placementgraph.com.