Straight answer: Likes, followers and impressions do not predict revenue, full stop. The metrics that do are the ones that show buying intent, saves, shares sent by DM, profile-to-website click rate, and DM-to-booked-call conversion. Track those four and you can stop staring at your vanity dashboard every morning.
The post with 40 likes that made more than the post with 4,000
Back when I had well over 300,000 followers across platforms and was regularly named as a top social influencer, I posted something on LinkedIn about pricing mistakes small consultancies make. It got 40 likes. Nothing. I nearly didn’t bother checking the comments because it looked like a flop next to some of my other posts that week, one of which had over 4,000 likes and dozens of shares.
That quiet 40-like post brought in a £12,000 consulting contract. Someone commented, I replied, we moved to DM, then to a call, then to a contract, all within eleven days. The 4,000-like post brought in exactly zero enquiries. Same account, same week, same audience. The only real difference was who was reading and what they needed at that moment.
I’ve told that story on stage more times than I can count because people never quite believe it until it happens to them. High engagement feels like proof something is working. Sometimes it’s proof of the opposite, that you’ve reached a lot of people who were never going to buy anything from you.
The metrics that don’t predict revenue (and why marketers still cling to them)
- Follower count. Vanity, plain and simple. I’ve seen accounts with 2,000 followers outsell accounts with 200,000. Followers tell you reach potential, not intent.
- Impressions and reach. Useful for brand awareness reporting to a board that likes big numbers. Almost no correlation to a purchase decision.
- Likes. The lowest-effort action a person can take. A like costs nothing and means almost nothing.
- Engagement rate as a single blended figure. Most engagement rate formulas lump likes, comments and shares together and divide by followers. That masks which specific action matters, which is the uncomfortable bit nobody wants to admit when they’ve built a whole reporting deck around it.
Agencies keep reporting these because they’re easy to pull, easy to chart, and easy to make look good in a monthly PDF. Nobody gets fired for showing a graph that goes up and to the right, even if it has nothing to do with the sales pipeline. I wrote about this exact backwards approach in why most businesses treat social media as marketing the wrong way round, and metrics are where the mistake shows up most clearly.
The four metrics that correlate with revenue
1. Saves
A save means someone intends to come back to this. On Instagram and LinkedIn, saves consistently correlate with people who later become customers, because saving is a reference behaviour, not a social one. Nobody saves a post to impress their followers, they save it because they need the information.
2. Shares sent privately (not public reposts)
A public repost is social proof for the sharer. A DM share, someone sending your post to a colleague or a friend with “look at this”, is a recommendation, and recommendations close deals. Most platforms won’t show you this number directly, which is exactly why most people never track it and why it’s so undervalued.
3. Profile-visit to website-click ratio
Not raw clicks, the ratio. If 500 people visit your profile and only 3 click the link, something in your bio or your last three posts isn’t earning trust. I aim for at least 8 to 10% on my own profiles. When mine dropped to around 4% for six weeks, I changed my pinned post and bio line and it climbed back to 11% within a fortnight, no other changes.
4. DM-to-booked-call conversion
This is the one that sits closest to your bank account. Count how many DM conversations you have in a month, then count how many turn into a discovery call or a booked demo. My own average sits around 1 in 6. If yours is below 1 in 15, the problem isn’t your reach, it’s your reply.
Building a system to track this (step by step)
Most business owners have no system at all, just gut feel and a screenshot folder. Here’s the version I use with clients, it takes about 45 minutes to set up:
- Step 1. Create a simple spreadsheet with columns for date, platform, post topic, saves, DMs started, calls booked, revenue closed.
- Step 2. Add UTM parameters to every link you post, even in your bio. Without this, “traffic from social” is a guess dressed up as data.
- Step 3. Every Friday, spend 15 minutes filling in the numbers by hand. Yes, by hand. Automated dashboards are handy later but hand-entry forces you to notice patterns.
- Step 4. After four weeks, look for the post types that generated DMs, not the ones that generated likes. There’s almost always a gap between the two.
- Step 5. Kill anything that gets likes but no DMs. Double down on anything that gets DMs even with modest likes.
If you’re not technical and the idea of setting up UTM tracking makes you want to close the laptop, I go through this exact process in plain English in how I use AI every day without being technical, including how I get an AI tool to pull the weekly numbers together for me so I’m not doing it manually forever.
Audience quality matters more than audience size
One thing I do differently now than I did five years ago is check who is in an audience before I trust any engagement number from it. A post that gets engagement from bots, students, or people in the wrong country is worthless no matter how good the number looks. I’ve used audience analysis tools like StatSocial to check whether a following is made up of decision makers or padded out with low-value accounts, and it’s changed which platforms I bother posting on for client work entirely.
This matters because the same metric can mean completely different things depending on who’s behind it. A 5% engagement rate from 500 senior operations managers is worth more than a 5% engagement rate from 50,000 teenagers, even though the percentage on the chart looks identical.
When to bring in outside help
If you’ve built the spreadsheet, run it for eight weeks, and you still can’t join the dots between social activity and closed revenue, that’s usually a sign the gap isn’t in your posting, it’s in your handoff, sales process, or offer clarity, which sits outside social media entirely. This is where a lot of business owners bring in help, either through a fractional CMO who understands both marketing and AI-driven measurement or through direct hands-on support setting up the tracking the first time via an AI implementation coach. I’d only recommend paying for this once you’ve tried the basic version yourself for a couple of months, otherwise you’re paying someone to tell you what your own spreadsheet would have shown you anyway.
If you’re weighing up whether you need a traditional marketing consultant or someone who understands AI-driven attribution specifically, I laid out the actual differences in AI consultant versus traditional marketing consultant, because the two roles solve different problems and hiring the wrong one wastes months.
The uncomfortable bit: most of your revenue never shows up in any metric
Here’s the part that doesn’t fit neatly into a dashboard. A huge chunk of the people who eventually buy from you saw your content, didn’t like it, didn’t comment, didn’t click anything, and then three weeks later Googled your name directly or asked a colleague about you. This is called dark social, and it’s estimated to account for well over half of all referral traffic that platforms and analytics tools simply cannot attribute back to the original post. Ofcom and various media research bodies have flagged this attribution gap for years, and it hasn’t gone away, if anything it’s got worse as more browsing moves to private messaging apps and closed group chats.
What that means practically is this: you will never build a perfect attribution model. Some of your best clients will tell you they’ve “followed you for ages” and you’ll have no record of them ever engaging with a single post. That’s not a failure of your tracking, it’s just how humans behave before they buy something that costs real money. The goal isn’t perfect attribution, it’s directional confidence, enough signal to know which content types are more likely to be doing the quiet work in the background even when nobody’s clicking anything.
What to check every week
- Which post generated the most saves this week, and what topic was it
- How many DM conversations started organically (not from your own outreach)
- Your profile-visit to click ratio, is it moving up or down
- How many of last month’s DM conversations turned into a call
- Whether any client mentioned “I’ve been following you for a while” without you having a record of them engaging
Do that consistently for three months and you’ll have a far more honest picture than any dashboard full of impressions and reach will ever give you. If you want a second pair of eyes to sanity-check the numbers or help you find someone who can run this for you, my guide on hiring an AI marketing consultant who knows what they’re doing covers the exact questions to ask before you hand anyone your budget.
Frequently asked questions
Do follower count and revenue correlate at all?
Barely, and often not in the direction people assume. I’ve seen small accounts with under 3,000 followers outsell accounts with over 100,000 because the smaller audience was made up almost entirely of people who could buy. Follower count tells you reach, not readiness to purchase.
What is the single best social media metric to track for revenue?
DM-to-booked-call conversion rate. It sits closest to actual sales in the funnel because it measures a real human decision to talk to you, not a passive tap of a like button. Everything upstream of that number matters, but this is the one that most directly tracks with revenue.
How long does it take to see revenue from social media metrics?
For most small businesses, expect a genuine lag of six to twelve weeks between consistent posting and a measurable rise in DM conversations and booked calls, and longer still before that shows up as closed revenue. Anyone promising faster is usually counting vanity metrics, not sales.
Should I ignore likes completely?
Not completely, they’re a rough signal that a topic resonated, which is useful for content planning. Just stop reporting likes to anyone who’s asking about revenue, and never let a low-like post convince you a topic didn’t work before you’ve checked saves, DMs and clicks first.
Related reading: How to Plan a Month of Social Content Without a Big Team and Why Your LinkedIn Account Got Restricted (and How to Fix It).
