Skip to content
← All articles

Mantra Social ·

Instagram replaced impressions with views: which numbers tell you if your social media is working

If your 2025 report looks better than your 2024 one, check how the numbers are counted first. The catch is a single word that changed.

In brief

  • Since 2024, Meta has replaced impressions with views, which also count the times the same person sees a piece of content again: for photos, carousels and stories, the numbers tend to rise even when nothing has improved.
  • Data from before and after the change can't be compared directly, and each analytics tool handled the switch in its own way, especially for stories and sponsored content.
  • To tell whether a company's social media is working, you need three levels of numbers: how many different people it reaches, how well it holds them and gets them to share, and which concrete inquiries come out of it.

On November 14, 2024, Meta posted a notice on its site for content creators with a very simple example: "if a person views a photo three times in the same day, it counts as three views for the photo instead of one impression." Right after that, the same page warns that, for this reason, the new number may come out higher than the old one.

That example explains a scene that plays out in a lot of companies. The social media report arrives, the chart goes up, and nobody can say whether the profile is doing better or the way of counting has changed. Anyone who pays for a report has the right to know, and the answer starts here.

What exactly changed between impressions and views?

The switch happened in several steps, between 2024 and 2025.

When What happened
August 2024 Instagram announces that views become the main number for reels, stories, photos and carousels
November 14, 2024 Meta extends the choice to Facebook: a single number, calculated the same way on both platforms
January 21, 2025 In version 22.0 of the connections Meta offers to outside tools, impressions become obsolete
April 21, 2025 Deadline set by Meta for updating those connections in all versions

For videos, the change is mostly a new name: what used to be called plays is now called views, and Meta says the calculation shouldn't change. For photos, carousels and stories, the substance changes. An impression counted the times a piece of content appeared on screen; a view also counts every time the same person comes back to it.

There is also a detail that matters to anyone looking at historical data. According to Meta's January 21, 2025 announcement, impressions for individual posts stop being provided only for content created on or after July 2, 2024. Older posts still return them.

Why can't you compare the numbers from before and after?

Because they measure two different things. If a company compares its impressions from fall 2024 with its views from fall 2025, it is holding up two different rulers. An increase could mean more interested people, or the same people coming back more often, or just the new method.

Analytics tools each solved the problem their own way. Metricool, for example, chose a cutoff date: before January 1, 2025 it still shows impressions, and from that day on it shows views. In the post list, it also separates organic views from sponsored impressions, because adding them together, as Instagram did, isn't reliable when there have been promotions, according to Metricool.

The practical consequence is simple. Every report should state the date from which views are counted, and year-over-year comparisons only make sense between periods measured the same way. If the report doesn't say, that's the first question to ask whoever prepares it.

Are stories an exception?

It depends on the tool you use. Meta has stated that, in its connections, stories return views too. Metricool, however, decided to keep showing impressions for stories, explaining that once a story expires, Instagram's connections don't provide all the view data.

This means two reports on the same profile, made with two different tools, can show story numbers that can't be compared. It's worth keeping in mind when you switch agencies or tools as well: your story history may not be the same kind of data.

Which Instagram metrics really matter for a company?

Views on their own say little, and Instagram itself has said so. When it announced them in August 2024, Adam Mosseri, the head of Instagram, urged people to keep looking at other numbers too, in particular how often a piece of content is sent to someone in a private message.

We line them up on three levels, from the easiest to inflate to the hardest.

1. How far you get: people reached. Instagram calls it reach, or accounts reached: the number of different accounts that saw the content at least once. This is the number that truly replaces the old impressions as a measure of distance covered, because each person counts only once. A useful trick: dividing views by people reached tells you how many times, on average, each person saw the content.

2. How interesting the content is: watch time and shares. In January 2025, Mosseri named the three signals that weigh most in Instagram's choice of whom to show a piece of content to: watch time, likes and private shares. He also suggested looking, in the profile data, at average watch time, likes per person reached and shares per person reached. We covered this in more depth in Reels or carousels. Our take: a figure divided by people reached accounts for the size of the audience, while a total does not.

3. What it produces: inquiries. Private messages, taps on the buttons to call or send an email, website visits, quotes requested. Instagram measures some of these; the rest has to be counted inside the company, by asking whoever calls or writes how they found you. This is the level that matters to the person paying, and it's the most neglected one in reports.

A word of caution applies to the official numbers too. In its developer definitions, Meta writes that reach is an estimated figure and that views are a metric still in development. They are tools for spotting trends, not counts accurate to the cent.

How do you read a social media report without being fooled?

With four questions, asked every month.

  • How many different people did we reach? Reach first, then views.
  • How much of this is paid? Organic and sponsored must be kept separate, otherwise a promotion makes the whole profile look better.
  • Which content held people and got them to share the most, relative to who saw it? This is how you figure out what to repeat.
  • What happened next? How many messages, contacts and inquiries came in, and from which content.

Then there's a question that comes before the report. A number is good or bad only against a goal: what specific thing the company wants to be remembered for, and by whom. That is the starting point of Category Design, and it applies to social media too: a thousand views from the wrong people count for less than a hundred from the right ones. This is how we set up reports with the companies we work with through Mantra Social: a few numbers, always the same ones, tied to a business question.

So is a growing report lying?

Not necessarily, but it can show an improvement that isn't there. Take a company profile that mostly posts photos and carousels and that in 2025 reached exactly the same people as in 2024. If the report puts this year's views next to last year's impressions, the chart goes up anyway. Meta put it in writing, in its November 14, 2024 notice: with the new count, the number may be higher than before.

Frequently asked questions

What is the difference between views and impressions on Instagram?

For videos, almost nothing changes: plays were simply renamed views. For photos, carousels and stories, however, views also count the times the same person sees the content again. Meta's example is a photo seen three times in the same day: that is three views, where before it was a single impression.

When did Instagram stop showing impressions?

Instagram announced views as its main number in August 2024, and Meta extended the choice to Facebook on November 14, 2024. For outside tools, version 22.0 of Meta's connections, released on January 21, 2025, made impressions obsolete, with April 21, 2025 as the deadline for all versions.

Can you compare 2024 social media data with 2025 data?

Not directly. Views count repeat viewings, while impressions largely did not. It's best to note in the report the date your analytics tool switched to views and to compare only periods measured the same way.

What is the most important number for a company on Instagram?

No single number. People reached tell you how far you get, watch time and private shares tell you how interesting the content is, and messages and contacts tell you whether social media brings in work. Instagram's head, Adam Mosseri, has urged people not to look at views alone.

Sources

  1. Meta for Creators, Introducing Views to simplify content distribution metrics (November 14, 2024)
  2. Meta for Developers, Introducing Graph API v22.0 and Marketing API v22.0 (January 21, 2025)
  3. Meta for Developers, Instagram Account Insights (definitions of reach, views and impressions)
  4. Metricool, Instagram replaced Impressions with Views: here's what you need to know
  5. Tubefilter, Views, views, views: Instagram standardizes its metrics to promote a multiformat approach (August 7, 2024)
  6. Social Media Today, Instagram shares algorithm insights for 2025 (January 22, 2025)

Keep reading

How is your channel doing?

We look at the channel with you: what you post, how often and what is missing. We tell you what we would change, no strings attached.

Ask for a channel check-up