Reach and Impressions at a glance
An impression is counted each time an ad is served, rendered or displayed, depending on the system doing the counting. The impression glossary entry sets out the counting rules ad platforms, ad servers and verification vendors use, and why the same campaign can produce several honest totals. The Media Rating Council maintains the industry standard for when a measured impression counts as viewable, which is why buyers increasingly ask which impressions could actually have been seen.
Reach is the deduplicated number of people exposed to an ad at least once during a period. One person who sees the same ad six times contributes six impressions and one unit of reach. What counts as a person depends on the platform's identity system, so most platforms present reach as an estimate rather than a tally. The reach glossary entry explains how it is modelled, and the guide to media objectives and reach goals covers how planners turn reach into a campaign goal.
The two metrics describe the same delivery from different angles. A given number of impressions can mean many people seeing the ad once or a few people seeing it many times, and only reach and frequency tell which. That is the whole difference, and it matters because those two outcomes do very different things for a brand.
Compare table
Reach vs Impressions
| Row label | Reach | Impressions |
|---|---|---|
| Unit counted | People, or the platform's best estimate of them | Showings of the ad, by the counting system's rule |
| Deduplication | Yes: each person counts once in the period | None: every showing adds one |
| How it is estimated | Modelled from logged-in, device and cookie identity, so a platform estimate | Counted when the ad is served or rendered, then filtered for invalid traffic |
| Link to frequency | The denominator: frequency is impressions divided by reach | The numerator: more impressions to the same people raise frequency |
| Pricing role | Rarely priced directly; reach-based buying still bills on impressions | The unit of price: cost per thousand impressions |
| Planning role | The goal: how many of the target audience the plan touches | The input: how much weight the budget buys |
| Cross-platform comparability | Poor: cannot be added across platforms without counting people twice | Better: totals can be summed when counting rules match |
| Common reporting trap | Adding reach across campaigns or platforms | Reporting a large impression total as if it meant many people |
The rows that matter most are deduplication and planning role. Impressions measure how much was delivered; reach measures how widely it was spread. A plan needs both, but it should be judged on the second.
When to choose Reach
Use reach to set and judge the goal of any campaign whose job is to be seen. Awareness, launches and campaigns that build memory over time succeed by putting the message in front of enough of the target audience, and only reach counts that. A reach goal is also the clearest way to explain a brand campaign to people outside marketing: the share of the audience who saw it is easier to discuss than a total of ads served.
Reach becomes most useful when it is tied to frequency. A person who saw the ad once may not have registered it, and the planner's concept of Effective FrequencyThe number of exposures a media plan assumes a person needs before an ad has its intended effect. sets the range of exposures within which the message is judged likely to work. The goal is then written as reach at that frequency or above, which the reach and frequency calculator can model for a given budget and audience.
The cost is uncertainty. Each platform estimates reach from its own identity data, and the estimates cannot be added across platforms. Measuring combined reach needs deduplication or an explicit overlap assumption, which the cross-platform reach and frequency guide covers.
When to choose Impressions
Use impressions for buying, pacing and checking delivery. Media is priced on impressions, so the cost per thousand is the currency of comparison between buys, and the impression calculator turns a budget and a price into the impressions a buy should deliver. Impressions are also the figure an advertiser checks against an invoice, the basis for discrepancy checks between an ad server and a seller, and the denominator of click-through rate and viewability.
Impressions are the right figure for daily pacing. A campaign that is under-delivering shows it first in impressions, and a sudden change usually means a targeting, budget or creative approval problem worth catching the same day. They can also be summed across channels when the counting rules match, which reach cannot.
Where impressions mislead is as a goal. A campaign can meet an impression target by showing the ad again and again to a small group, and a large impression total can hide a narrow reach. On social platforms, frequency rising inside an unchanged audience is also an early sign of creative fatigue, which the frequency and creative fatigue guide describes; Meta's help centre documents frequency controls for reach-optimised ad sets for that reason.
An oat drink brand changes its campaign goal
Flaxmantle Oat Drinks, a challenger brand, runs a four-week social video campaign with a goal of 6,000,000 impressions. It hits the goal in three weeks and reports success. The platform shows reach of 750,000 people at an average frequency of 8.0, against a target audience of about 3,000,000. The campaign had reached only a quarter of the audience, and many of those people had seen the ad more than ten times. For the next flight, Flaxmantle sets the goal as reach of 1,800,000 at a frequency of at least two, adds a weekly frequency cap of three, and keeps impressions for pacing and price checks only. The same budget reaches 1,950,000 people at an average frequency of 3.1.
Decision table
| Situation | Use | Why |
|---|---|---|
| Setting the goal of an awareness or launch campaign | Reach at an effective frequency | The goal is enough different people seeing the message enough times |
| Comparing prices between media buys | Impressions, through cost per thousand | Media is priced on impressions |
| Checking daily pacing and delivery | Impressions | They show under-delivery and errors the same day |
| Reporting results across several platforms | Impressions summed; reach only with deduplication | Reach cannot be added without counting people twice |
| Spotting creative fatigue | Frequency, from both | Rising impressions per person in a steady audience signal wear-out |
| Explaining a brand campaign to the board | Reach as a share of the target audience | People reached is easier to judge than ads served |
Frequency: the number that connects them
Frequency is impressions divided by reach for the same period and level, and it is the bridge every reach and impression conversation eventually crosses. A fixed budget buys a fixed number of impressions; frequency decides whether those impressions are spent finding new people or repeating the message to people already reached. Most media add reach quickly at first and then slowly, as further impressions fall increasingly on people already reached, which is why piling more weight into one platform eventually buys frequency rather than reach.
Frequency is an average over a skewed distribution, so the same average can mean very different things. A campaign with a moderate average may have shown the ad once to most people and many times to a few. Ask for the frequency distribution where the platform offers it, set caps where the platform allows, and read frequency at the same level and period as the reach it was divided by. The CPC versus CPM comparison explains why impression buying gives the advertiser that control over reach and frequency in the first place.
Common mistakes
- Reporting impressions as if they were people, so a large total hides a narrow audience seen many times.
- Adding reach across campaigns or platforms, which counts the same people more than once.
- Setting an impression goal for an awareness campaign instead of a reach goal at an effective frequency.
- Reading average frequency without the distribution, which can hide heavy repetition to a small group.
- Comparing reach across periods of different length, when reach only makes sense for a stated period.
- Treating every impression as seen, when viewability standards exist because many are not.
- Comparing impression totals from systems with different counting rules without reconciling them first.
