Social media metrics that sit closest to revenue

Likes are the most visible number attached to a post and the least useful, and the reason they dominate is structural rather than anyone's mistake. They are displayed publicly, they update immediately, and they require nothing from the business to collect.

The problem is what a like represents. It costs a viewer nothing, commits them to nothing, and ends the interaction — a person who liked a post has typically finished with it. That makes it a weak signal of agreement and close to no signal about buying.

The consequence of using likes as the measure is not merely inaccuracy. It changes what gets made. Content optimised for easy approval drifts towards the agreeable and away from the specific, and specificity is what makes someone enquire. A business measuring likes will, over months, produce a more likeable and less useful account.

What follows is the small set of numbers that sit closer to a customer doing something, and how to actually collect them.

Each of these requires a decision from the viewer, which is what makes them meaningful.

Profile visits. Someone saw a post and wanted to know who made it. This is the first moment of genuine interest in the business rather than in a piece of content, and for most small accounts it is the single most informative number available.

Link or website taps. Someone chose to leave the app. Given the friction involved, a small count here carries more weight than a large number of likes.

DM enquiries. Closest to a customer, and usually the number a business actually cares about. It is also the one most often uncounted, because it arrives as a conversation rather than as a figure on a dashboard.

Saves. Not a purchase signal, but the best available indicator that content had durable value, since a save means someone expects to need it again.

Four numbers is a deliberately short list. A dashboard with fifteen metrics gets read as a whole and acted on in none.

The most valuable of those four is the one no platform reports, and this gap is worth closing manually because it is where social media meets revenue.

The method is unglamorous: keep a count. A note, a spreadsheet, a tally — anything that records how many enquiries arrived through social channels each week, and where possible what prompted each one. Asking is legitimate and cheap: someone messaging about availability will usually say what they saw if asked directly.

That last detail is what makes the count actionable rather than merely reassuring. Ten enquiries in a month tells you social is producing something. Ten enquiries of which seven followed the same kind of post tells you what to make more of, which is a different and much more useful fact.

Two cautions. Attribution from asking is imperfect, because people misremember and often see several things before acting. And enquiries are not sales — tracking which enquiries converted is a further step, and worth taking if the numbers are small enough to track by hand, which for most small businesses they are.

Traffic arriving at a website from social media is invisible as a distinct source unless the link carries something that identifies it.

The standard mechanism is a set of tracking parameters added to the end of a URL — conventionally recording the source, the medium and a campaign name. Any analytics tool reads them, and they turn an anonymous arrival into an identifiable one. Building them is a matter of appending text to the link; no integration is required.

The part that makes this worth doing properly is consistency. Parameters spelled differently across posts produce several categories for the same thing and a report nobody trusts. Decide the naming once, write it down, and reuse it.

Where to apply it: the profile link, links in Stories, links inside a link-in-bio page, and anywhere else a social post sends someone. Without this, traffic from the profile link is mixed in with everything else and the contribution of the whole account is unmeasurable.

One caveat worth knowing: some apps strip or alter parameters, so treat the resulting numbers as a reliable comparison over time rather than an exact count.

Having better metrics creates a temptation to draw firmer conclusions than they support, and the honest limits are worth stating.

None of this establishes that social media caused a sale. A customer typically encounters a business several times through different routes before buying, and attributing the purchase to the last identifiable touch overstates that touch's role while understating everything earlier. This is a genuine limitation, not a gap to be closed with a better tool.

What these numbers do support is comparison: this month against last month, this kind of post against that kind, this network against that one. Comparison over time is where the value is, and it does not require solving attribution.

What they cannot support is a claim about a percentage of revenue coming from social media, and any figure of that kind produced from this data is manufactured. The defensible version is narrower and still useful: enquiries from social channels rose or fell, and the posts they followed had these characteristics in common.

The practice that makes any of this work is a short regular review, and it needs to be monthly rather than daily.

Once a month, record four figures: reach as a proportion of followers, profile visits, saves, and enquiries counted by hand. Put them in a file outside the platform, because Insights windows are limited and the trend is the entire point. Then look at the three posts that produced the most profile visits and write one sentence about what they had in common.

That sentence, accumulated over several months, becomes the only genuinely reliable guide to what works for your specific account — more reliable than any general advice, including this article, because it is derived from your audience rather than an average of unrelated ones.

What to avoid is daily checking, which reads variation as signal and produces constant change without accumulated understanding.

Social management tools including Socie can surface what a platform reports through one interface, which reduces the collection effort. The enquiry count and the one-sentence conclusion are the parts that produce the value, and both are done by a person.