Impressions are gone. So are Page fans, post_engaged_users and Instagram profile views. Here is what Meta still serves, what it does not, and why a dashboard that shows you an impression count in 2026 is showing you something it invented.
If you have run Facebook Pages or Instagram professional accounts for any length of time, you have a mental model of what the platform tells you. That model is now partly wrong, and the gaps matter more than the remaining numbers.
What was removed
Meta has retired several metrics that reporting tools were built on:
- Impressions. There is no impression count for a post any more. Views and unique viewers are what you get.
- Page fans. The daily series is gone. Follower totals still come from the account object itself, but you cannot read the day-by-day history from the insights endpoints.
- post_engaged_users. The count of distinct people who interacted with a post no longer exists.
- Instagram profile views. Removed entirely.
None of these were deprecated gently. They stopped being served.
What is still there
For a post, Meta serves views and unique viewers — which this product stores as views and reach — plus likes, comments and shares on both platforms, saves on Instagram, and clicks and reactions on Facebook.
At account level you get daily Page views, daily reach, daily post engagements, and the follower total. Instagram additionally reports new followers per day, but only for the last 30 days; the total comes from elsewhere.
The trap in daily reach
Daily reach is the number of unique people reached that day. Add seven of those together and you do not have weekly unique reach — you have a number with no meaning, because someone reached on Monday and Tuesday is counted twice.
Plenty of tools sum it anyway and label the result "reach". If a report gives you a monthly reach figure that is suspiciously close to the sum of its daily values, that is what happened.
A summed daily reach figure is not unique reach, and anything presenting it as such is reporting a number the platform never produced.
The honest handling is to show the sum, label it as the sum of daily reach, and not call it unique.
Why this changes how you report
Two things follow.
First, year-on-year comparisons across the removal are not like-for-like. An impression count from 2024 has no 2026 counterpart. Putting them on the same axis produces a cliff that looks like a collapse in performance and is actually a change in what the API serves.
Second, a gap has to be visible. When a number cannot be collected, the useful thing to show is that it cannot be collected. A blank cell reads as zero, and zero is a result — it says the post reached nobody. "Not reported by this platform" says something completely different, and it is the truth.
That is the rule this product follows everywhere: unmeasured, failing and simulated figures are always named next to the number they affect.