Cross-platform video reporting without lying to yourself

A single slide showing TikTok, Reels and Shorts side by side is what every stakeholder wants. Producing one that is not quietly misleading takes three decisions.

Decision 1 — fix the window before anything else

The most common error in cross-platform decks is not metric definitions. It is one platform reported over the campaign window and another over the video's lifetime, because those exports were easier to get.

Rule: every number in the comparison covers the same dates, computed as the difference between a reading at the start and a reading at the end. If you cannot produce a windowed number for one platform, leave it out or footnote it. Do not silently substitute a lifetime counter.

Decision 2 — know what a "view" includes on each surface

All three major platforms now count a view when playback starts, and all three include replays. That convergence is recent: Instagram consolidated impressions and plays into a single views metric in 2025, and YouTube moved Shorts to counting every playback in March 2025, keeping the stricter engaged views as a separate metric.

What still differs:

• Feed mechanics. How aggressively a feed autoplays and how quickly users swipe changes what a "playback start" is worth.
• Typical length and loop behaviour. Short loops accumulate replays faster.
• Availability of a stricter metric. Only YouTube publishes an engaged-views equivalent, and only in Advanced mode.

The honest conclusion: views are comparable in kind, not in value. Put them next to each other; do not claim one platform "beat" another on views alone.

Decision 3 — choose what to sum and what to keep separate

A combined view total is acceptable as exposure delivered when labelled that way. A combined engagement rate is not: it is an average of ratios with different denominators, dominated by whichever platform contributed most views.

The report structure that holds up

1. 1Headline — one line of delivery: videos live, combined views in window, total cost, blended cost per thousand.
2. 2Per-platform table — views in window, engagement rate, CPM, video count. Same window, stated in the header.
3. 3Trend — daily views for the window, per platform, with the previous equal-length period behind it. Days without data shown as zero, not skipped, otherwise a collection gap reads as a dip in performance.
4. 4Top and bottom performers — ranked by views earned in the window, never by lifetime counters, so an old hit cannot occupy first place forever.
5. 5Method footnote — metric definitions, window, capture times, and any platform where numbers were unavailable.

That footnote is what makes the deck defensible three months later when someone re-runs the numbers and gets something slightly different.

Two habits that eliminate most disputes

Store readings, not summaries. A timestamped snapshot history lets you recompute any window later. A spreadsheet of final numbers cannot be re-cut, and cannot answer "what did this look like on day 3".

Report the freshness of your data. If some videos have not been read since before the window closed, say so and say how many. A report that quietly mixes fresh and stale readings is worse than one that admits a gap.