What is the difference between reach and impressions?
Reach is the number of accounts that saw your ad at least once. Impressions count how many times your ad was on screen, including repeat views by the same account. Meta estimates reach from sampled data and de-duplicates it, while impressions are counted from all data. Impressions divided by reach gives frequency.
Frequency = impressions ÷ reach
Reach and impressions measure the same delivery from two angles. Reach asks “how many different accounts did this ad get in front of?” Impressions ask “how many times did it appear?” If one account sees your ad three times, that adds 1 to reach and 3 to impressions. When the two disagree in a report, check that difference first, then remember that one is estimated and the other is a full count.
What does Meta count as reach?
Meta defines paid reach as the number of Accounts Center accounts that saw your ads at least once, and labels the metric as estimated. Its reach article adds that the metric is calculated using sampled data, and that reach can be affected by your bid, budget and audience targeting.
Some Meta articles, such as the reservation glossary, describe reach as the number of people who saw your ads. The paid reach definition counts Accounts Center accounts, so treat reach as a count of accounts.
What does Meta count as an impression?
Meta counts an impression the first time an instance of your ad is on screen. Its rules:
- Scrolling away and back to the same ad counts as 1 impression.
- The ad coming on screen for someone at 2 different times in a day counts as 2.
- Outside Meta Audience Network, images and videos are counted the same way, so a video need not start playing.
- Where Meta can’t tell whether an ad is on screen, such as on feature phones, it counts the impression when the ad is delivered to the device.
- Impressions from invalid traffic Meta detects, such as bots, are excluded.
Meta also notes that impressions from certain placements are accredited by the Media Rating Council (MRC).
How do reach and impressions compare?
| Compared on | Reach | Impressions |
|---|---|---|
| What it counts | Accounts that saw the ad at least once | Times the ad was on screen |
| Repeat views | Counted once | Each one counted |
| Measurement | Estimated, uses sampled data | Based on all available data, not sampled |
| De-duplicated across ad sets | Yes, each account counted once | No, every view counted |
| Related cost | Cost per 1,000 Meta accounts reached | CPM (cost per 1,000 impressions) |
Why doesn’t my reach add up across ad sets?
Because reach is unique. Meta’s example: ad set 1 reaches persons A and B, ad set 2 reaches B and C. Each ad set shows a reach of 2, but the campaign shows 3, since B is counted once.
Meta lists two more reasons totals may not match:
- Check whether the report compares sampled metrics, because totals and breakdowns are sampled separately and a total can come out lower than its breakdowns.
- Check whether deleted ad sets or ads are hidden, since the default view excludes them; use the delivery filter to include them.
Where do you see both in Ads Manager?
Meta lists impressions, reach, frequency and cost per 1,000 impressions in the Delivery column preset, so one view shows all four.
FAQ
Does a video have to play to count as an impression?
No, except on Meta Audience Network. Meta counts image and video ads the same way, when the ad first comes on screen, so a video can earn an impression without starting to play. Meta notes this differs from industry standards for video ads.
Should you report on reach or impressions?
Meta notes that on its platforms reach can be more insightful than impressions, because it shows how many accounts your message reached and how efficiently. Impressions are useful for exposure and CPM. Pick the one that matches your goal: unique audience or total exposure.
Why does my reach say “estimated”?
Meta labels paid reach as estimated and says reach and other unique metrics may rely on sampled data, which lets Ads Manager return them faster. Metrics such as impressions and clicks are not sampled and are based on all the data available.
Sources
- Paid reach (Meta, checked 2 Oct 2026)
- Reach (Meta, checked 2 Oct 2026)
- Impressions (Meta, checked 2 Oct 2026)
- About data sampling (Meta, checked 2 Oct 2026)
- Troubleshoot reasons why ad metrics may not add up (Meta, checked 2 Oct 2026)
- Cost per 1,000 Meta accounts reached (Meta, checked 2 Oct 2026)
- Glossary of reservation terms (Meta, checked 2 Oct 2026)
- About columns presets in Meta Ads Manager (Meta, checked 2 Oct 2026)
- CPM (cost per 1,000 impressions) (Meta, checked 2 Oct 2026)
- Frequency (Meta, checked 2 Oct 2026)
Drafted with AI and checked claim by claim against the pages linked under Sources. Found a mistake? Tell us via the corrections policy.