How a brand's posts did over the last days (default 7, up to 90), optionally on one platform: totals (posts, views, impressions, reach, likes, comments, shares, saves, clicks, engagement = likes + comments + shares + saves), top_posts by engagement with each one's numbers, permalink and over_time (its numbers day by day), what_worked (the best posts by engagement rate and how each platform and time of day did, with a plain summary for the person; engagement_rate is engagement per view, or per impression where a platform has no views), counts of posted, failed and needs-action targets, and freshness (when the numbers were last read; posts from the last 2 days are read every 2 hours, older ones once a day). engagement.available is false, with the reason, when numbers cannot be read: analytics is not available for the workspace yet, or it posts through its own apps. best_times has engagement by weekday and hour against the posting times (its grid). insights is what to plan next week from: patterns in the brand's own posts over the last 90 days (format: photo, carousel, video or text; platform; time of day; weekday), each with a plain summary, the sample behind it (best and other post counts), strength ('early' on a few posts, 'steady' on more), a caveat when something else was shared by most of the best posts, and try_next, one thing to do next week; note says why there is none yet and what to post. Say the sample and the strength when you repeat one, never call it a cause. vs_previous compares views, likes and posts with the same number of days before, overall and per platform (change is a share: 0.24 is 24% more; null when there was nothing before or no numbers yet this period; newer posts are still collecting views); vs_previous is null when since is given or days is over 45. since is still accepted instead of days.
