What do reach and frequency mean, and how are they calculated?
Part of Programmatic media planning
The short answer
Reach is the number or share of a target audience exposed to an ad at least once in a period. Frequency is how many times, on average, those reached were exposed. Average frequency equals impressions divided by reach, and reach multiplied by average frequency gives gross rating points. Effective frequency is the exposure range a plan treats as enough.
Key takeaways
- Reach counts each member of the audience once, however many times they saw the ad, so it can be shown as a number or as a percentage of the target universe.
- Average frequency is impressions divided by reach, and the MRC defines GRPs as reach multiplied by frequency.
- Effective frequency is a planning assumption about how many exposures are enough, not a measured standard, so write down where the number came from.
- Adding sites raises reach only as far as their readers are new; overlapping readerships add impressions and frequency instead.
What do reach and frequency mean in media planning?
Definition
- Reach
- The unique people, homes or devices in a target audience exposed to an ad at least once during a period, given as a count or as a percentage of the measured population.
Definition
- Frequency
- The number of times those reached were exposed during the same period, reported as an average across everyone reached, and sometimes as a distribution by number of exposures.
The most precise public definitions sit in the MRC's Cross-Media Audience Measurement Standards (opens in a new tab) (Phase I: Video, September 2019). There, reach represents unique users, unduplicated homes or audience "exposed to ads or content (have generated a Viewable Impression) at least once during a time period", expressed as a percentage of the measured population, universe or target. Frequency is the number of times they generated a viewable impression, "expressed as an average among those unique users."
Two details in that definition matter for digital plans. First, the MRC standard builds reach on viewable impressions, not on every rendered impression; the ad impression guide explains the difference. Second, reach is de-duplicated: a person who saw the ad nine times counts once in reach and nine times in impressions. A platform report may count reach on served impressions instead, so check which basis a report uses before comparing two figures.
How do you calculate average frequency?
Divide impressions by reach. The MRC standard writes it as the sum of viewable impressions divided by the sum of unique audience with a viewable impression. The same document gives the related formulas a plan needs, all of which use one measured population as the denominator.
| Metric | Formula | What it tells the planner |
|---|---|---|
| Reach, percent | Unique audience exposed at least once divided by the measured population, times 100 | How much of the target the campaign touched |
| Average frequency | Impressions divided by unique audience exposed | How often, on average, each reached person was exposed |
| Rating | Impressions divided by the measured population, times 100 | Exposure weight relative to the whole target |
| Gross rating points (GRPs) | Reach times average frequency, or the sum of ratings | Total exposure weight, with duplication left in |
Worked arithmetic for one fictional flight
- Set the universe. Assume a target of 1,000,000 adults in the campaign's markets. Every percentage below uses this denominator.
- Take the impressions. Assume the plan forecasts 2,400,000 impressions over four weeks.
- Take the reach. Assume the forecast reaches 400,000 unique people at least once. Reach is 400,000 divided by 1,000,000, or 40 percent.
- Divide for average frequency. 2,400,000 impressions divided by 400,000 reached is an average frequency of 6 over four weeks, or 1.5 per week.
- Multiply for GRPs. Reach of 40 times frequency of 6 is 240 GRPs, which matches 2,400,000 divided by 1,000,000, times 100.
What is effective frequency?
Effective frequency is the range of exposures a plan treats as enough for the message to work: below it, people may not register the ad; above it, extra exposures add cost and irritation. It can be expressed as effective reach, the share of the target exposed at least a set number of times, such as 3 or more in a four week flight.
No standards body sets a correct number. The right range depends on how familiar the brand is, how complex the message is, how many creatives rotate and how cluttered the environment is. Treat any figure as a planning assumption, write down where it came from, such as a past brand study, the client's own research or a platform's documented goal ranges, and check the frequency distribution against it after launch. Google's Target frequency (opens in a new tab) for YouTube awareness campaigns shows one platform's operational range: weekly goals of 2 to 7 for multi-format ads.
| Exposures in four weeks | People reached | Share of reach | Reading against an assumed effective range of 3 to 10 |
|---|---|---|---|
| 1 to 2 | 160,000 | 40 percent | Below the range |
| 3 to 10 | 200,000 | 50 percent | Inside the range: effective reach of 200,000, or 20 percent of the universe |
| 11 or more | 40,000 | 10 percent | Above the range; a cap could move these impressions to new people |
The table shows why effective reach can say more than average frequency. The same 240 GRPs could produce very different effective reach depending on how impressions are spread. Limiting the top row is what a frequency cap is for; how to set one is in the frequency capping guide.
How does the number of sites in a plan change reach?
Reach grows with a site list only when the new sites bring readers the plan does not already have. Two specialist sites in the same niche can share much of their readership, so adding the second adds impressions and frequency more than reach. A site on an adjacent topic with a different readership can add more reach than a larger site in the same niche.
A quick check before adding sites: ask the buying platform's forecast for reach with and without the candidate sites, holding budget constant. If reach barely moves while impressions rise, the new sites overlap the existing list. If reach rises, they bring new readers. Popularity data such as the Chrome UX Report cannot answer this on its own: its rank is a relative measure based on "the total number of navigations on the origin", per Chrome's CrUX methodology (opens in a new tab), not a count of unique readers.
How should reach be reported without personal data?
Reach always rests on some way of recognizing the same viewer twice, and that method decides what the number means. The IAB and MRC Audience Reach Measurement Guidelines (opens in a new tab) (version 2.0, August 2013) separate unique cookies, unique browsers, unique devices and unique users, and note that unique cookies "do not generally represent unduplicated browsers, users or people." They also restrict the word unique to records de-duplicated across the entire reporting period, so a weekly average of daily unique users may not be called a weekly unique count.
Platforms now report reach as modeled, aggregate figures. Google's Measuring reach and frequency (opens in a new tab) help page says Google Ads "uses statistical models that account for user behavior across many browsers and devices", combined with inputs such as census data and probability surveys. Its reach metrics appear only once a campaign passes minimum impression and unique user thresholds, mostly for date ranges of 92 days or less, and its cookie based reach metrics were replaced in August 2020.
Reach reporting checklist for a client report
- The unit is named: people, homes, devices, browsers or cookies.
- The basis is named: viewable impressions or all counted impressions.
- The period is stated, and reach is de-duplicated across that whole period, not averaged from shorter ones.
- Modeled or projected figures are labeled, with the source and method the platform documents.
- Reach from two platforms is shown side by side, never added, unless one source de-duplicated both.
- The report shows aggregate counts only, with no list of individual users or device IDs.
- Frequency is shown as an average and as a distribution.
The MRC cross-media standard asks for the same honesty: inferences, adjustments and projection methods "shall be disclosed with the reported estimates or as part of methodological reference materials." A client report that names its unit, basis and method can be defended without any personal data leaving the platform. For the wider plan, see the programmatic media planning hub.
What does a reach and frequency plan look like on one brief?
Fictional example
A family pizza restaurant chain in Arizona and New Mexico
A fictional family pizza chain with locations in Phoenix, Tucson and Albuquerque plans a four week display and online video flight in English and Spanish, beside local food guides, family weekend listings and local sports coverage, to launch a new menu. The client asks for as many local families as possible to see the launch at least three times.
The planner turns that into a reach goal and an effective frequency assumption of 3 to 10 exposures over the flight, taken from the client's last brand study. The DSP's forecast for the first list of 25 sites shows high frequency and modest reach: most sites are food blogs with a shared readership. Adding eight sites on family activities and youth sports raises forecast reach at the same budget, while the food list is trimmed.
The client report will show reach as modeled people from the platform, the frequency distribution, effective reach at 3 or more, and the method note. No delivery or performance result is implied.
Frequently asked questions
Is reach the same as impressions?
No. Impressions count every exposure, so one person who saw an ad five times adds five impressions. Reach counts that person once. That is why impressions divided by reach gives average frequency, and why a campaign can grow its impressions without growing its reach at all.
Can reach from two platforms be added together?
Only if one measurement source de-duplicated both. People seen by two platforms would otherwise be counted twice, so a simple sum overstates reach. Show each platform's reach separately, and use combined reach only from a source that saw both sets of impressions and documents how it removed duplicates.
Why does reach stop growing as the budget increases?
Once most of the reachable audience on a list has seen the ad, extra budget buys more exposures for the same people. Reach flattens while frequency climbs. Widening the list to sites with different readers, or lengthening the flight, can add more reach than adding budget to the same list.
What are GRPs in digital advertising?
Gross rating points measure total exposure weight against a target. The MRC defines GRPs as reach multiplied by frequency, or impressions divided by the measured population times 100. Two plans with the same GRPs can have very different reach, so report reach and frequency beside any GRP figure.
Sources
- Cross-Media Audience Measurement Standards (Phase I: Video), September 2019 (opens in a new tab)Media Rating Council
- Audience Reach Measurement Guidelines, version 2.0, August 2013 (opens in a new tab)IAB and Media Rating Council
- Measuring reach and frequency (opens in a new tab)Google Ads Help
- About Target frequency (opens in a new tab)Google Ads Help
- CrUX methodology: metrics (opens in a new tab)Chrome for Developers
- Standards and Guidelines (opens in a new tab)Media Rating Council
Researched and written by PlacementGraph Research. Read how we research, source and update these pages. Corrections: diana@placementgraph.com.