Skip to content

How to track TikTok, Instagram and Facebook video performance across accounts

By Hayden Barnett · · 4 min read

If you run a report that puts a TikTok, a Reel and a Facebook video side by side, the first question someone asks is “which one did best?” The honest answer is usually “it depends what you mean by views”, because the three platforms don’t count the same thing, and they don’t show the same numbers to a signed-out visitor.

This post is how I’d build a cross-account video tracker on public data: which calls to make, which fields to keep, which numbers you can compare and which you can’t. Every request below is a real Fetchproof request; each one costs 1 credit by default.

Start from the account, not the video

The tracker’s unit of work is an account. Once a day, list each account’s recent videos, store what you see, and only then decide which videos deserve a closer look. One listing call per account gives you stats for a whole page of videos at once, which is far cheaper than looking each video up.

TikTok. Profile videos returns 15 posts per page, newest first, each with a statistics block:

curl 'https://api.fetchproof.com/v3/tiktok/profile/videos?handle=nasa' \
  --header "x-api-key: $FETCHPROOF_API_KEY"

Each item in aweme_list carries statistics.play_count, digg_count (likes), comment_count, share_count and collect_count (saves), plus create_time and url. To go further back in time, pass the max_cursor from the response while has_more is 1.

Instagram. The reels tab returns the first 12 reels on a public profile, each with its play, like and comment counts:

curl 'https://api.fetchproof.com/v1/instagram/user/reels?handle=natgeo' \
  --header "x-api-key: $FETCHPROOF_API_KEY"

Twelve is a hard limit for signed-out reading: paging_info.max_id is always null, and passing max_id returns a 400. For a daily tracker that’s fine, because you only need the newest reels; older ones stop moving.

Facebook. A page’s timeline comes 3 posts at a time, with reactions, comments, shares, the publish time and video URLs:

curl 'https://api.fetchproof.com/v1/facebook/profile/posts?url=https%3A%2F%2Fwww.facebook.com%2Fpacemorby' \
  --header "x-api-key: $FETCHPROOF_API_KEY"

Pass the cursor from each response to get the next 3. Three per page means a busy page can take a few calls a day, so budget for that.

“Views” are three different numbers

This is the part most dashboards get wrong.

  • TikTok returns one number, statistics.play_count, on both the listing and a single video lookup.
  • Instagram has two. A single post lookup returns video_view_count, which is Instagram’s older view count. The number Instagram shows on a reel today is play_count, which the reels tab returns and which a post lookup adds when you pass include_play_count=true. Compare play_count with play_count, and never mix the two in one column.
  • Facebook mostly doesn’t show views to a signed-out visitor. On the timeline, videoViewCount is always null. A single post lookup includes view_count only when the public page shows it, and logged-out reel pages usually don’t.

So a “views” column across all three platforms will be either wrong or mostly empty. My recommendation is to keep views as a per-platform number and use engagement, meaning likes or reactions plus comments plus shares, as the cross-platform comparison. Every platform here exposes those counts publicly, apart from shares on Instagram’s reels tab, which doesn’t include them. Even engagement isn’t quite like for like: Facebook’s count is every reaction, not just likes. That’s worth a footnote on the report, not a reason to skip the comparison.

Normalise once, at ingest

Pick one row shape and map each response into it when you store it. Keep null for anything a platform doesn’t show, rather than 0: a missing number and a zero are different facts.

// One row per video per snapshot. null means "the platform doesn't show this", not zero.
function fromTikTok(item) {
  const s = item.statistics;
  return { platform: 'tiktok', id: item.aweme_id, url: item.url, postedAt: item.create_time,
    views: s.play_count, likes: s.digg_count, comments: s.comment_count, shares: s.share_count };
}

function fromInstagramReel({ media }) {
  const hidden = media.like_and_view_counts_disabled;
  return { platform: 'instagram', id: media.id, url: media.url, postedAt: media.taken_at,
    views: hidden ? null : media.play_count, likes: hidden ? null : media.like_count,
    comments: media.comment_count, shares: null };
}

function fromFacebookPost(post) {
  return { platform: 'facebook', id: post.id, url: post.url, postedAt: post.publishTime,
    views: null, likes: post.reactionCount, comments: post.commentCount, shares: post.shareCount };
}

Two details that save debugging later:

  • Instagram reels carry like_and_view_counts_disabled. When it’s true, the owner has hidden the counts, so store null rather than whatever number comes back.
  • Store a snapshot per video per day, with the time you fetched it, instead of overwriting one row. Performance is a curve, and how much a video gained in its first day or week tells a client more than a single running total.

Refresh the videos that matter

Listing calls cover the newest videos. For a video that’s still climbing, or one that has fallen off the first page, look it up directly:

curl 'https://api.fetchproof.com/v2/tiktok/video?url=https%3A%2F%2Fwww.tiktok.com%2F%40nasa%2Fvideo%2F7708867568920038669' \
  --header "x-api-key: $FETCHPROOF_API_KEY"

curl 'https://api.fetchproof.com/v1/instagram/post?url=https%3A%2F%2Fwww.instagram.com%2Freel%2FDLDXI0fylTC%2F&include_play_count=true' \
  --header "x-api-key: $FETCHPROOF_API_KEY"

curl 'https://api.fetchproof.com/v1/facebook/post?url=https%3A%2F%2Fwww.facebook.com%2Freel%2F1535656380759655' \
  --header "x-api-key: $FETCHPROOF_API_KEY"

All three single-video lookups take cache_max_age (1d, 3d, 7d, 14d or 30d). A cached response inside that window costs 0 credits, which suits a dashboard that many people open during the day. Leave it off for the daily snapshot itself, because the point of the snapshot is a fresh number.

A schedule that keeps the bill flat: refresh a video daily while it’s new, then let your own snapshots decide. Once a video’s daily gain flattens out, drop it to weekly, and stop when it stops moving. There’s no point paying a credit a day to record the same number.

What it costs

Each of these calls costs 1 credit by default, and errors, blocked or timed-out requests, invalid requests and cache hits cost nothing. A worked example:

  • 30 accounts, 10 per platform, one listing call each per day: 900 calls a month.
  • 5 direct video refreshes per account per day: 4,500 calls a month.

That’s 5,400 credits a month. On the Freelance pack ($27 for 25,000 credits) it works out to $5.832 a month, and one pack lasts about 4.6 months. Facebook pages that post a lot need extra listing pages at 3 posts each, so add a little headroom for those.

What this approach won’t give you

Public data shows what a signed-out visitor sees. It doesn’t include reach, impressions, watch time, audience demographics or anything else from an account owner’s insights. If you manage the accounts yourself, the platforms’ own analytics have those numbers. A public-data tracker is for the accounts you don’t own: competitors, creators you’re vetting, or a client’s rivals.

If you’d rather start with one platform, the TikTok video stats guide walks through the same calls step by step, and tracking follower growth covers the account-level numbers.

Try it with 100 free credits

Sign up with an email address, no card. Credits never expire. Errors, invalid requests and cache hits cost 0 credits, and each endpoint page lists exactly what it charges.