A social media performance audit is a table with one metric per row and one decision per row. If a number cannot change what you post next week, it does not belong on the scorecard. This one has six rows, scores each from 0 to 2, and takes about 20 minutes a month.
TLDR: Track six metrics, each tied to a question a client or boss will ask. Compare every post to your own median, score each row 0, 1 or 2, and write one decision for every row that scores 0. Most drift shows up as a gap between rows 2 and 5.
Six rows, one question each
Pull the last 30 days of posts and fill in this table. "Median" always means the middle value of your own last 30 posts on that platform.
| # | Question | Metric | Score 2 | Score 0 |
|---|---|---|---|---|
| 1 | Are people seeing us? | Reach per post vs your median | Median reach flat or up | Three posts in a row under 0.8x median |
| 2 | Do they care? | Engagement rate per post vs your median | Most posts inside your usual range | Median engagement down for two months |
| 3 | Is it passed on? | Shares and saves per post | Rising or steady | Near zero while reach is fine |
| 4 | Do they act? | Link clicks or profile visits per post | At least one post a week gets clicks | No clicks for a full month |
| 5 | Is the message landing? | Share of posts carrying this month's priority message, and how those posts perform vs the rest | 30% or more of posts, performing at or above median | Under 10% of posts, or those posts sit below median |
| 6 | Did we do the work? | Posts published vs posts planned | 90% or more shipped | Under 70% shipped |
Score 1 means in between. A perfect month is 12. Anything under 8 gets a meeting.
Row 5 is where drift hides
Drift is not a falling number. It is the gap between what you are trying to say and what you are actually posting. Reach can stay flat while the posts about your new service quietly shrink from half the calendar to a tenth.
To check, tag each post with the one message it carries (a product, an offer, an event, "brand only"). Count them. If the message you care about is under 10% of posts, no metric on rows 1 to 4 will rescue it.
Compare to your own median, then sanity-check with a benchmark
Your own median is the fair baseline. A post at 0.6x is a real miss, a post at 0.9x is noise. For a rough outside line, Mydrop's September 2026 data puts the median engagement rate at 3.1% on Instagram and 1.1% on Facebook, with the middle half of Instagram posts between 1.1% and 6.7%. A 2% Instagram post is normal.
Use that range to catch absurd targets, not to grade yourself. A small account at 7% and a large account at 1% can both be healthy.
Turn every zero into one written decision
A red row without a decision is just a worry. Write the action next to it:
- Row 1 red: check the posting hour first. In Mydrop's data, posts published between midnight and 6 am local time ran 17% below median across platforms.
- Row 2 red: change one format for two weeks. Carousels ran 10% above an account's typical post in the same data.
- Row 3 red: rewrite three posts so someone would send them to a colleague.
- Row 4 red: add one clear call to action per week and track the link.
- Row 5 red: put the priority message on the calendar first, then fill the gaps.
- Row 6 red: cut the plan to what you really ship. Planning 20 posts and publishing 12 is a calendar problem, not a motivation problem.
Keep the audit short enough to repeat
Put the six rows in a sheet, fill them on the first Monday of the month, and keep the old scores so you can see the trend. If you run several accounts, Mydrop's performance view lines every post up against its own account's median, and scheduled reports can send the same six numbers to a client each month.
Next step: pull your last 30 posts today, fill rows 1 and 5 only, and see which one is further from where you thought it was. That takes ten minutes and tells you whether you have a performance problem or a message problem.
Google review
Trustpilot review
Product Hunt comment
