← Back to Blog

How to Audit Your Social Media Accounts: An 8-Pillar Framework

A social media audit is a structured check of one account against a fixed list of questions, scored the same way every time. The most useful version uses eight pillars, proves each one with a metric the platform actually reports, and compares your last 30 days to your own previous 90 days rather than to an industry average. Done properly, it takes about an hour per account and ends with three specific things to fix this month.

Key takeaways

  • Audit one account at a time. The pillars are identical across platforms, but the evidence is different on each.
  • Score against your own baseline: the median of the previous 90 days. Industry averages hide differences in niche, format and account size.
  • Use medians, not averages. One viral post should not make a weak month look healthy.
  • Leave a pillar unscored when the data does not exist. A guessed score is worse than a blank.
  • The output is not the score. It is a 30-day plan aimed at your two weakest pillars.

Before you start: set the window and the baseline

Pick two periods. The audit window is the last 30 days. The baseline is the 90 days before that. Every performance question in the audit becomes “is the audit window better, the same, or worse than the baseline?”

Open a spreadsheet with one row per post or video published in both periods. Record the date, format (long-form, Short, Reel, carousel, photo, text post), and the handful of metrics listed below. You will find them in YouTube Studio, Instagram Insights, Meta Business Suite and TikTok’s analytics.

Then calculate the median of each metric per format for both periods. Format matters: comparing a Reel’s reach to a carousel’s reach tells you about the formats, not about your performance.

The eight pillars and what proves each one

Each pillar answers one question. The table shows where to look for the evidence on each platform.

Pillar Question YouTube Instagram Facebook Page TikTok
Content strategy Do you publish consistently, in formats and topics that earn results? Uploads per week; views by topic and format Posts per week; views by format Posts per week; views by post type Videos per week; views by topic
Discoverability / SEO Can people who do not know you find you? Share of views from YouTube search and suggested videos Share of reach from non-followers; keyword-rich captions and profile Reach beyond followers; complete Page info Share of views from search and For You
Engagement Do viewers respond? Likes and comments per view Interactions (likes, comments, saves, shares) per account reached Reactions, comments and shares per reach Likes, comments and shares per view
Growth Is the audience getting bigger, and at what rate? Net subscribers; subscribers gained per video Follows; follows per profile visit Net followers Net followers; followers per video
Branding Is it obvious who you are and what you offer? Channel name, banner, About text, consistent thumbnails Name field, bio, link, highlights, grid consistency Page name, category, About, cover image Name, bio, link, pinned videos
Retention Do people keep watching? Average view duration; average percentage viewed Limited data; often unscored Limited data; often unscored Average watch time; watched-full-video percentage
Hooks & thumbnails Does the first impression win the click or stop the scroll? Impressions click-through rate; first 30 seconds of the retention graph Views relative to followers on Reels; opening line and cover 3-second video views relative to reach Early drop-off; share of views from For You
Reach / traffic Where does the audience come from, and is it concentrated in one source? Traffic source mix Reach and views; followers vs. non-followers Page reach and views Traffic sources: For You, Following, profile, search, sound

Notes on the tricky pillars

Retention. YouTube gives you the richest retention data of any platform, so it is the easiest place to score this pillar. If you need a refresher on which numbers matter there, see YouTube analytics metrics that matter. On TikTok you can read average watch time and the watched-full-video percentage per video in the app. On Instagram and Facebook the data is thinner and less consistent across formats. If you cannot collect the same number for every post, leave the pillar blank.

Discoverability. This is about strangers. A post that only reaches existing followers is distribution, not discovery. On every platform, the proof is the share of views or reach that comes from people who did not already follow you, plus the basic hygiene that makes search possible: clear titles, descriptive captions, spoken and on-screen keywords, a complete profile.

Branding. This is the only pillar scored by inspection rather than by numbers. Ask someone who does not know your account to look at the profile for five seconds and tell you what you post and who it is for. If they cannot, the score is low regardless of how the profile looks to you.

The 0-4 scoring rubric

Score each pillar from 0 to 4. The thresholds below are starting points, not laws. Adjust them once you know how much your numbers normally move.

Score Performance pillars (audit window vs. baseline median) Inspection pillars (branding, parts of content strategy)
0 No activity or no data collected Missing: no bio, no clear topic, no consistent look
1 More than 20% below baseline Present but confusing or out of date
2 Within 20% below to 10% above baseline Clear, but generic; a stranger gets the topic, not the reason to follow
3 10% to 30% above baseline Clear topic, audience and reason to follow
4 More than 30% above baseline, and you can name the cause All of the above, plus consistent across every post and platform

Two rules keep the scoring honest.

  1. A 4 requires a cause. If the median jumped and you do not know why, score it 3. Unexplained gains tend to disappear.
  2. Unscored is not zero. If a pillar has no reliable data, exclude it from the total instead of entering 0.

The total is: sum of scores ÷ (number of scored pillars × 4) × 100.

The score is only comparable to your own earlier scores on the same account. It tells you where to work, not how you rank.

A worked example

Say you run an Instagram account about home coffee brewing. These numbers are invented for illustration.

In the 90-day baseline you posted 36 Reels with a median of 9,800 views and a median of 410 interactions. In the 30-day audit window you posted 9 Reels with a median of 8,900 views and a median of 455 interactions.

  • Content strategy: Baseline cadence was 3 Reels a week; the audit window is about 2. Topics are consistent. Score 2.
  • Discoverability: The share of reach from non-followers fell from roughly 60% to 45%. That is 25% below baseline. Score 1.
  • Engagement: Interactions per view went from 410 ÷ 9,800 = 4.2% to 455 ÷ 8,900 = 5.1%, about 22% higher. Saves drove it, after you began adding recipe cards. Above baseline with a known cause, but under 30%. Score 3.
  • Growth: 310 follows per 30 days in the baseline, 290 in the window. Within range. Score 2.
  • Branding: The bio says “coffee lover” and nothing about what a follower gets. Score 2.
  • Retention: Not collected consistently. Unscored.
  • Hooks & thumbnails: Median views per Reel are down 9%, and four of nine Reels open with a logo sting. Score 2.
  • Reach / traffic: Median views down 9%, within range, but increasingly dependent on followers. Score 2.

Total: 14 points across 7 scored pillars. 14 ÷ (7 × 4) × 100 = 50.

The number 50 is not the finding. The finding is that engagement among existing followers is strong while discovery is shrinking. This account is getting better at serving people it already has and worse at meeting new ones. That shapes the plan.

Turn the two weakest pillars into a 30-day plan

Pick the two lowest-scoring pillars. If there is a tie, choose the one earlier in the viewer’s journey: discoverability and hooks come before retention and engagement, because nobody can engage with a post they never saw.

For each pillar, write one hypothesis and no more than three tasks. A task must be something you can finish and tick off, not an intention.

For the coffee account:

  • Discoverability (score 1). Hypothesis: captions and on-screen text do not say what the Reel is about, so the platform has little to match against searches and interests. Tasks: rewrite the bio to state topic and audience (the Instagram bio optimizer helps with a first draft); put the specific subject in the first line of every caption and as on-screen text in the first two seconds; publish two Reels a week aimed at beginner questions rather than at existing followers.
  • Hooks (score 2). Hypothesis: the logo sting costs the first second. Tasks: remove it; write three alternative opening lines per Reel before filming (the hook generator is useful for options); start every Reel on the finished result, not the setup.

Lay the tasks out over four weeks. Week 1 is fixes you do once (bio, profile, templates). Weeks 2 and 3 are the new behavior applied to every post. Week 4 is a mini-review: recalculate the two medians you were trying to move. A pre-publish content checklist makes the new habits stick.

What “good” looks like without industry benchmarks

You will be tempted to look up an average engagement rate or click-through rate and compare. Resist it for scoring. Published averages mix niches, formats, account sizes and different formulas. The article on engagement rate formulas by platform shows how much the same account’s rate changes depending on the denominator alone.

Your own history has none of those problems. It is the same niche, the same audience and the same formula. Three patterns are worth looking for:

  • Direction: Is the median rising or falling across three consecutive audits? One period is noise; three is a trend.
  • Spread: Is the gap between your best and worst post narrowing? A tighter spread at the same median means you are getting more repeatable.
  • Balance: Are any pillars more than two points apart? Large gaps point to a bottleneck. Great retention with weak hooks means good videos few people start.

Auditing several platforms at once

Audit each account separately, then put the pillar scores side by side. Reading across the row is often more revealing than reading down the column.

If branding is low everywhere, the problem is your positioning, and a new bio on one platform will not fix it. If hooks score 3 on TikTok and 1 on YouTube Shorts with the same videos, the content is fine and the packaging or the audience differs. If growth is strong on one platform only, that is where your discovery works; study its traffic sources before copying the format elsewhere.

How often to re-audit

A full audit every 90 days is enough for most creators. That gives each 30-day plan time to run and leaves a clean baseline for the next comparison. In between, do a monthly 15-minute check of only the two pillars you are working on.

Audit early when something breaks the pattern: a sharp fall in reach, a format change, a rebrand, or a platform changing how it counts a metric. When definitions change, as they did when Instagram moved from impressions to views, restart the baseline for the affected metric rather than comparing old and new numbers.

Put it into practice

  1. Choose one account. Set the audit window (last 30 days) and the baseline (the 90 days before).
  2. Export or copy per-post data into a spreadsheet with a format column.
  3. Calculate medians per format for both periods.
  4. Score each of the eight pillars 0-4 using the rubric. Leave pillars without reliable data unscored.
  5. Calculate the total: sum ÷ (scored pillars × 4) × 100.
  6. Pick the two weakest pillars. Write one hypothesis and up to three tasks for each.
  7. Schedule the tasks across four weeks and put the review date in your calendar.
  8. Save the sheet. In 90 days, the audit window becomes part of the new baseline.

CreatorsMetrics runs this 8-pillar audit on your connected YouTube, Instagram, Facebook and TikTok accounts using read-only sign-in. You get a written report and a 30-day plan, and a pillar is only scored when the platform provides the data.

Create a free account

If you only want to try it on a YouTube channel first, the free channel audit covers YouTube on its own, and the multi-platform audit page explains what is scored where.

Frequently asked questions

How often should I audit my social media accounts?

Run a full audit every 90 days and a light 15-minute check every month. Re-audit sooner after a big change such as a new format, a rebrand or a sudden drop in reach.

What is a good social media audit score?

There is no universal pass mark, because the score compares you to your own baseline. What matters is which pillars are lowest and whether the total rises between audits.

Can I use the same audit for YouTube, Instagram, Facebook and TikTok?

Yes, the eight pillars are the same, but the metric that proves each pillar differs per platform. Leave a pillar unscored when the platform does not give you reliable data for it.

How many posts do I need before an audit is meaningful?

Aim for at least 20 to 30 posts in the audit window so that medians are stable. With fewer posts you can still check branding and content strategy, but treat the performance pillars as provisional.

Should I compare my audit to competitors?

Use competitors for ideas about formats and topics, not for scoring. You cannot see their reach, retention or traffic sources, so your own previous period is the only fair benchmark.