The Facebook Page Insights reporting dashboard for agencies

Facebook counts a Page like and a Page follow separately, because they are separate acts. Somebody can like a Page and later mute it, and somebody can follow a Page without ever liking it — which is why the two counts drift apart, and why a report that publishes only one of them is answering half the question. The Page growth tab shows both, each with its own pair of flows: new likes against unlikes, new followers against unfollows, drawn as paired bars so the churn underneath a calm net number is visible rather than implied.

Underneath the bars, a daily table walks the most recent days of the selected range and reports each day as gains, losses and the net that came out of them, for both counts at once. That is the view that answers the awkward question in a monthly call — not whether the audience grew, but which day it stopped growing, and whether the audience that left had liked the Page or merely been following it. Unlikes are the number nobody volunteers; putting them in the same table as the wins is what makes the growth story credible when it is good.

The Posts tab lists everything the Page published in the window with its type, the day it went out, reach, the viral share of that reach, engagements, reactions, comments, shares, engagement rate and link clicks — sortable by engagement, by reach or by recency. Each row is a real post, so a spike in the trend line above resolves to a specific piece of content rather than a hypothesis. The type pill on every row is the same vocabulary the filter chips use, which keeps the table and the rest of the report speaking one language.

Comparing formats on Facebook only works if each is measured for what it is meant to do. A link post is built to move people off Facebook and will usually reach fewer of them while earning far more clicks; a Reel is built to travel and often reaches well past the people who follow the Page; a Story is short-lived and small by design. Selecting a content type re-bases the scorecards, the trend, the mix and every tab, and a type with nothing published in the window says so plainly instead of showing an empty panel.

Facebook offers seven reactions, and they do not mean the same thing. The Reactions tab draws each post as a stacked bar across Like, Love, Care, Haha, Wow, Sad and Angry, then totals them in a table where the counts add up to the post reaction total exactly and the percentages add up to one hundred. Sentiment is named beside each row rather than hidden inside a single engagement figure, because a post carried by Love is telling you something different from a post carried by Angry, and both would look like a good month in a report that only publishes engagements.

The same tab reports the four negative-feedback actions Page Insights counts against a post: hiding the post, hiding everything from the Page, unfollowing from the post itself, and reporting it as spam. Each is shown per post and as a rate per thousand people reached, which is the form that makes it comparable across a big post and a small one. Sustained hides on one content type is the clearest early signal that a Page is posting too often, or to people who were never the audience — and it usually appears weeks before the follower count reacts.

A Facebook video view is not one number, and quoting the largest of them is how video reporting loses its credibility. The Video tab reports the whole funnel every video post ran through: impressions, three-second views, one-minute views and the viewers who reached the end, with the length of the video, average watch time in seconds and the share of views that played with sound on. Three seconds is the loosest count Facebook publishes and a minute is the strictest, so seeing both against the runtime turns a headline view count into something an account team can act on rather than something to read aloud.

It is also honest about its own limits. A one-minute view cannot exist on a video shorter than a minute, so short Reels and Stories report none rather than a fabricated number — the column shows a dash and the note beside it says why. Where sound-on share is low, the case for captions and a visual hook writes itself; where average watch time collapses well before the end, the fix is the first few seconds rather than the length. The video posts each get a card with the funnel drawn as bars, and the table underneath ranks them.

Content is only half of what a Facebook Page does. The Page actions tab reports the calls to action people took on the Page itself — website button clicks, phone calls, get-directions taps, Messenger conversations and email clicks — with Page views beside them, each shown as a total, as a rate per thousand people reached, and as a share of all actions. Website button clicks are deliberately kept apart from the link clicks a post earns, because they are different events with different owners, and merging them into one number quietly flatters whichever one is weaker.

The fans-online tab draws a seven-day by twenty-four-hour grid of how many of the Page fans were on Facebook in each hour, marks the busiest hour and day, and can switch to a plain text list of the three busiest hours per day so the same finding is readable without the colour. It is described for what it is: a headcount of who is present, not a promise that a post will reach them. Used that way it is a scheduling argument an account manager can make on a call, from this Page rather than from a general study.