Your impressions number is probably double-counted
Short answer
Most social analytics exports carry the same reach concept under several column names, so summing them double-counts. Take the highest value on each row instead.
Most reach numbers in social decks are too high, and it is almost always the same mistake: someone summed columns that were describing the same thing.
What actually happens in the export
Open a Sprinklr export and you will find several columns that all sound like reach. On a TikTok sheet: TikTok Video Views (SUM) and TikTok Video Reach (SUM). On Instagram: Views (SUM), Reach (SUM), sometimes Impressions. A scraped export gives you plays.
These are not additive. They are different measurements of the same event, and often of overlapping populations. Add them together and you have invented an audience.
The rule is simple: map every view-type column to one impressions metric, then resolve each row to the single largest value. Never the sum.
Reach stays separate. It answers a different question — unique people rather than times served — and it is a different denominator. If you fold it into impressions your engagement rate becomes meaningless.
The engagement version of the same trap
Engagement has the mirror problem. Where the export gives you a platform-confirmed total, that total is the number. Sprinklr's Total Engagements (VML) (SUM) is authoritative. Only sum likes, comments, shares, saves and reposts when no such column exists.
Then check the two against each other, because the gap tells you something:
| Components vs confirmed total | What it means |
|---|---|
| Around 100% | Everything is mapped. Carry on. |
| Under 100% | A component column is unmapped — usually reposts or quotes. |
| Over 100% | Something is double-counted, or the platform's total excludes a component. |
That last row is not hypothetical. Instagram exports routinely land at 102–104% because Reposts Count Trend is a real component that Instagram's own Engagement (SUM) column leaves out. Neither number is wrong. They are answering different questions, and you have to decide which one your report is claiming.
The thing not to do is quietly fix it. Surface the discrepancy and make the call in the open.
The one that catches everybody
Facebook-style exports carry Reactions Love, Wow, Haha, Sad, Angry and Care alongside a Reactions Total. The total already contains them. Map the total, ignore the breakdown, and never let an auto-mapper merge both — otherwise likes get counted twice and your engagement rate doubles.
Same principle as the impressions rule: the export is describing one event several ways, and your job is to pick one description, not to add them up.
Why this matters more than it sounds
An inflated impressions figure deflates every rate built on top of it. Engagement rate drops, and the natural conclusion is that the content is underperforming — so someone changes the content, when the only thing wrong was the spreadsheet.
I have seen a brand rebrief a whole content pillar off the back of it. The fix took ten minutes. The rebrief took a quarter.
Key facts
- A single Sprinklr row can carry impressions, video views, plays and reach as four separate columns describing overlapping things.
- Summing view-type columns inflates reach. The correct rule is take-highest per row.
- Reach and impressions are different denominators and should never be merged.
- Instagram's 'Engagement (SUM)' column excludes reposts, so component columns can add up to more than the platform's own total.
Common questions
- Should I add impressions and video views together?
- No. They describe overlapping measurements of the same event. Take the highest value on each row instead — summing them inflates reach and deflates every rate calculated from it.
- What is the difference between reach and impressions?
- Impressions count times served; reach counts unique people. They are separate denominators and should be kept as separate metrics, never merged into one number.
- Why do my engagement components add up to more than the platform total?
- Usually because the platform's own total excludes a component. Instagram's Engagement (SUM) column omits reposts, so adding likes, comments, saves, shares and reposts lands at roughly 102–104% of it.
Want this run on your own numbers?
The audit tool applies all of it to your export, free, in your browser.