how to do a social media audit.

most "audits" are a pile of screenshots nobody acts on. a useful audit answers one question: what should we keep, kill, and change? here is how to run one that ends in decisions.

— 01profiles and consistency.

start with the basics — bios, links, visual consistency across platforms. a quick win-list usually appears in minutes: broken links, off-brand grids, missing calls to action.

start with the unglamorous layer, because it is where most of the loss is. correct hours, a working link, the right location, a bio that says what you do, and identical details across every platform and search profile.

a mismatch between profiles is more expensive than it looks — inconsistent details reduce how confidently search engines and answer engines describe you, and they make a business look inattentive.

then check the inbox. how many messages are unanswered, how long the average reply takes, and whether anyone owns it. that single check regularly finds more revenue than the rest of the audit combined.

— 02content and competitors.

find your top and worst posts and ask why. then look at what is working for competitors. patterns emerge fast and point straight at your next content moves.

look at your top ten posts by saves and shares rather than likes. those two indicate intent, and the pattern across ten of them usually contradicts what the team assumed was working.

then look at the bottom ten. the formats that consistently underperform are the ones to cut, and subtraction is the fastest improvement available to most calendars.

for competitors, study gaps rather than formats. what does the category never explain, which customers does everyone ignore, what is nobody willing to talk about. that is where an opening is.

— 03metrics that matter.

review reach, saves, profile actions and conversions — not just followers. the audit should end with three things to change, backed by what the numbers say.

pick one number and record its current value — enquiries a month, bookings, cost per acquisition. an audit that does not end with a baseline has produced a document rather than a decision.

then map how someone actually gets from a post to a purchase, and find where they drop out. usually it is a missing price, a slow reply or a page that answers a different question than the post raised.

and ignore follower count unless you sell to advertisers. it is the metric most likely to be reported and least likely to change anything.

— 04what to do with the findings.

an audit is only worth the changes it produces, so end it with three: one thing to stop, one thing to fix this week, and one thing to test next quarter. more than three and none of them happen.

then re-run it in six months against the number you wrote down. that comparison is the only way to tell whether the changes worked, and it takes an afternoon.

if this is the problem: audit the page before the ads, auditing your brand voice, run your own marketing audit.