How to find your best-performing social media posts

Ask most owners which post did best and they'll name the one with the most likes. That's often the post that reached the most people, which may have more to do with timing or luck than content. And likes are the weakest signal of whether anyone will buy.

Finding your real best performers is worth an afternoon every few months. It tells you what to make more of, which is the most valuable thing your analytics can do. The method below works on any platform and needs nothing more than a spreadsheet.

It also protects you from a common trap: copying whatever did well last week. One post can do well because of a holiday, a share from a bigger account or plain luck. Looking across months of posts, with fair comparisons, separates what you can repeat from what you can't.

Pick one main measure before you look at any data. Different goals produce different winners, and mixing them gives you a list of posts that were good at different things.

Platform dashboards show posts one at a time and forget them after a while. Copy the last three to six months into one sheet so you can sort and compare. On Instagram, the content library lets you view and filter all your content by media type, reach, interactions and time frame, which is a quick way to get the list. Insights timeframes only go back 90 days, so copy the figures before they drop out of view.

If you post to several platforms, keep one tab per platform. Comparing a LinkedIn post with a TikTok video by raw numbers tells you about the platforms, not the posts.

Fill in the columns below for each post. It's tedious the first time, but once the sheet exists, adding each month's posts takes a few minutes. Our social media post tracker has these columns set up.

• Date and time posted
• Platform and format (photo, carousel, short video, text, story)
• Topic or content pillar
• Hook: the first line or first seconds, written out
• Reach or views
• Interactions: likes, comments, saves, shares
• Clicks, profile visits or messages, where available
• Notes: boosted, collaboration, holiday, news moment

Source: Instagram Help Centre: About Instagram insights, checked 15 October 2026

Raw totals favour posts that were lucky with reach and posts that have been up longest. Three adjustments make the comparison fairer. The table below uses invented figures to show why.

First, use rates: divide your main measure by reach and multiply by 100 or 1,000. Saves per thousand reached is more telling than total saves. Second, compare like with like: rank videos against videos and carousels against carousels, because formats are shown differently. Third, compare posts at the same age, such as seven days after posting, or at least don't compare a post from yesterday with one from last quarter.

One great post is an anecdote. Five great posts that share a feature are a pattern. Add a few simple tags to each post, then sort by your main rate and count which tags appear most in the top fifth and the bottom fifth.

Keep the tag list short and decide it before you start, or you'll invent a new tag for every post. The bottom fifth matters as much as the top: a tag that keeps turning up among your weakest posts is a strong hint about what to stop. A hypothetical example: a boutique finds that four of its top six posts show a person wearing the item and five of its bottom six show the item on a hanger. That's a pattern worth testing.

• Pillar: education, behind the scenes, social proof, offer, community
• Format: photo, carousel, short video, text-only, story
• Hook type: question, problem, result first, number, bold claim
• Who's in it: product only, a person, a customer
• Call to action: none, save, comment, message, link
• Length: short or long caption, video under or over your usual length

Some posts top the list for reasons you can't repeat. Mark them in the notes column and set them aside before drawing conclusions.

• Posts you boosted or ran as ads, which reached people organic posts can't
• Posts shared by a much bigger account or a local news page
• Posts tied to a one-off moment: a holiday, a local event, a trending sound that's gone
• Giveaways and contests, which pull engagement from people who may never buy
• Posts that caused a row in the comments: high engagement, not always good news

The point of the exercise is next month's calendar. Turn each pattern into a repeatable post type and give it a regular slot. If how-to carousels about caring for a product are your best saves, make that a fortnightly series. If result-first video hooks beat everything else, use them by default.

Old winners can be reused, too. Repost a top performer from six months ago with a fresh opening, turn a strong carousel into a short video, or answer the questions it drew in the comments. Our content repurposing planner helps map one strong post into several.

Don't abandon everything else. A feed made only of your top post type can tire your audience, and your weaker pillars may still do a job, such as reassuring people who are close to buying. Shift the balance towards what works rather than cutting the rest overnight. Then test the pattern rather than assuming it. Our guide to A/B testing posts shows how to check that the feature you spotted is what made the difference.

Zolyp, which we make, includes Analyze: a plain-language weekly report on reach, engagement and clicks across your connected accounts, with what to do more of and what to drop, and what worked feeds back into your Brand Brain so the next plan reflects it. It can only see what each platform's own API exposes for your connected accounts, and its history starts when you connect, with no backfill. Treat it as a weekly shortlist; a deeper look across a few months, as above, is still worth doing yourself. Zolyp publishes to Instagram, Facebook, LinkedIn, Google Business, X, TikTok, YouTube, Pinterest, and Threads. Plans that create the posts start at $29 a month, Lite is $9 a month per brand for scheduling posts you make yourself, and there's a 3-day free trial.