Glossary
What Is an Analytics Dashboard for Social Media?
An analytics dashboard is a view that aggregates performance metrics from one or more social platforms so a creator or team can compare content by format, topic, and cadence, and decide what to make next. The useful version starts with retention, sends, saves, and reach beyond the existing audience, and it treats every number as evidence about a creative choice rather than as a standalone scoreboard.
By Bell Chen, founder of Superdirector — how the analysis works.
Adam Mosseri, who runs Instagram, posted a video on January 21, 2025 (instagram.com) naming watch time, likes and sends as the top three signals. He recommends reviewing average watch time, likes per reach, and sends per reach, with likes slightly more important for connected audiences and sends for unconnected audiences. Put those metrics alongside views and the post's intended business outcome. The dashboard should help the team compare creative tests, rather than turning the listed signals into a universal weight table.
What a dashboard should actually surface
A useful dashboard presents retention, likes, sends, saves, and reach beyond the existing audience alongside views and followers. That is an editorial reporting choice, not a published ordering of actions. Mosseri's Reels caption (instagram.com) names watch time, likes and sends as top signals, with likes slightly more important for connected audiences and sends for unconnected audiences. Choose the metrics that answer the post's goal rather than treating every private share as a stronger growth result than every public like.
On TikTok the same logic runs through the buckets the Newsroom explainer (newsroom.tiktok.com) names: user interactions first, then video information, then device and account settings. The interaction bucket includes watch-through, replays, comments, shares, and the not-interested tap, and it outweighs everything else. A dashboard that aggregates those interaction signals across platforms tells a creator which posts the systems are actually rewarding, which is rarely the same as which posts have the biggest view number.
How to read the numbers without fooling yourself
Sends per reach is a useful derived figure: divide the number of times a post was sent in a direct message by the number of accounts it reached. This ratio normalizes the send count for audience size. Mosseri's caption (instagram.com) recommends reviewing it alongside average watch time and likes per reach; it does not rank it first among engagement signals. A post sent by one in fifty viewers and one liked by one in five describe different actions, so retain both observations.
As an illustrative, hypothetical calibration to make the ratio concrete, and not a benchmark to copy: imagine a clip that reaches 10,000 accounts and is sent 200 times. That is a 2 percent send rate. A second clip reaches 40,000 accounts and is sent 240 times, a 0.6 percent send rate. The first clip has a quarter of the reach but more than triple the per-viewer send rate. That comparison can guide an editorial test of shareable payoffs, but Mosseri's caption (instagram.com) does not let those figures alone predict which clip receives further distribution.
Retention is the other ratio that resists vanity. Read the percentage of viewers still watching at three seconds and at the end, not the average watch duration alone, because a long average can hide a steep early drop. On TikTok the early drop is the swipe-away the Newsroom (newsroom.tiktok.com) counts as a negative user interaction, so a dashboard that shows the retention curve, not just the headline number, is the one that explains why a clip stalled.
How to run a top-posts and low-performers audit
Once a month, pull your strongest posts by goal, not by views. Group them: which drove sends and saves, which drove profile visits and follows, which drove link clicks. For each group catalog the shared traits, hook style, topic category, length, format, and publishing context, so the audit produces a short list of repeatable patterns rather than a feeling. This is the part most teams skip, and it is the part that turns the dashboard from a report into a plan.
Then run the same pass on your weakest posts. Knowing which formats the audience reliably skips is as valuable as knowing which they reward, because it tells the team where to stop spending production time. On TikTok the platform's Traffic Sources view adds context here by showing whether a post mostly reached followers, searchers, profile visitors, or the broader recommendation surface, which is the difference between a post that failed to convert an existing audience and one that never escaped it (newsroom.tiktok.com).
Finally, check the dashboard against the goal of each post before you judge it. A tutorial should be read on saves, a trend reaction on watch-through and shares, a profile-building post on follows, a sales post on clicks. Judging all four by the same number is the most common way a dashboard produces confident, wrong decisions.
Common mistakes
The first mistake is judging a post only by views or followers. Mosseri recommends reviewing average watch time, likes per reach, and sends per reach on Reels (instagram.com); TikTok names user interactions among its recommendation inputs (newsroom.tiktok.com). Keep those observations alongside the post's business goal rather than declaring a universal winning column.
The second mistake is overreacting to the first hour of data. Early numbers are noisy, and the signals that matter accumulate over days, so a same-day verdict trains the team to chase variance instead of patterns.
The third mistake is reviewing posts in isolation rather than as a portfolio. The audit that finds repeatable patterns across the top and bottom posts is what improves the next batch; judging each clip alone produces opinions, not a strategy.
Where Superdirector fits
A dashboard tells you which posts won; it does not tell you why. Superdirector's current reference-analysis results can surface recurring hook structures, pacing patterns, and topic angles as inputs for the next brief. Watch time, sends, and saves remain in native platform analytics, while the operator owns the decision about what to test next. This page compares only the Superdirector.
Disclosure by Bell Chen, founder of Superdirector: the reference-analysis features mentioned here are part of the product I build. The ranking signals and benchmarks in this piece are sourced from the linked platform statements and industry playbooks; treat the tooling note as one input among several, and read every number against your own account baseline.