1. Where the numbers come from
When you paste a link we read the post from two public, no-login sources:
- The FxTwitter API, which returns the post’s text, views, likes, replies, reposts, quotes, bookmarks, format, and the author’s follower count and checkmark.
- X’s own embed data (the data behind embedded posts on websites), used only when the first source fails. It has likes and replies but no views, reposts or follower count, and the report says so.
Answers are cached for five minutes, so a popular link doesn’t hit those sources again and again. We never use a logged-in X session or scrape logged-in pages.
2. What each number means
- Views per follower: views ÷ the author’s followers. Views count impressions, not people, so this isn’t a count of followers reached; above 1× usually means X showed the post beyond the author’s followers, mostly in the For You feed.
- Engagement rate: likes, replies, reposts, quotes and bookmarks added up, ÷ views.
- Replies per like: how much conversation a post started for each like.
- Shares per like: reposts plus quotes ÷ likes; how often people passed it on.
- Bookmarks per like: how often people saved it to come back to, a sign of useful content.
- Views an hour: views ÷ hours since it went up, an average over the post’s life.
3. How the likely reasons are chosen
Each reason is a rule over the numbers above: for example a post whose views are well above its follower count gets “reached past its followers”, and a post with an outside link gets a note that X tends to show those to fewer people. Every reason says what it’s based on.
The thresholds are our own judgement calls, chosen to flag clear cases. To explain why replies matter, reports quote X’s open-sourced 2023 ranking code, where a reply counted about 27 times as much as a like, a reply the author answered about 150 times, and a repost about 2 times. X has changed its ranking since, so these show direction, not exact scores.
4. Comparing two posts
With two posts we line up the differences that public data can show: account size and checkmark, reach past followers, replies per like, format, links and timing (hours apart in the day, in UTC). When nothing public explains the gap, we say so: the difference usually came from who engaged first.
5. Live tracking
X’s public numbers are totals only. While you keep a tab open on a post with tracking on, the page checks the post every 10 minutes and draws the curve. The history is kept only in your browser.
6. What this tool can’t see
- Who saw a post, and in which feed.
- Whether reach was paid for. For that we link to X’s public ad library.
- The growth history of an older post.
- X’s ranking scores for any post.
Authors see some of this in their own X analytics; an outside tool can only estimate it. Try it on a post.