Likes Aren't a Metric: What to Actually Track on Social Media
You open the monthly Instagram report. Likes are up from last month, reach climbed too — every number has a green arrow next to it. Then you check with sales: no new leads, no extra orders. Both stories can't be true at once. One of them isn't describing what actually happened.
The problem isn't the like itself, it's the weight the report gives it. A tap on a screen promises nothing, yet it sits at the top of the page because it's the easiest number to pull and it grows on its own as you post more. The question worth asking isn't who saw it — it's who did something next.
Why a like tells you nothing
The like button records where a thumb paused for a second while scrolling — that's the whole mechanism. Someone can watch a reel, tap like, and forget the account existed five seconds later: it's a reaction, not an intention. That's why you won't find a clean line between like counts and DM volume, or between likes and orders — one measures a feeling, the other a decision.
The same trap applies to reach. By the platform's own definition, "reach" counts unique accounts that saw a post, while "impressions" count how many times it was shown in total — if one person saw the same reel three times, reach stays at 1 while impressions read 3 (source). Neither number tells you who cared, only who was shown the post. Giveaways and contests inflate both easily — the entry rule is often "share to enter," but the entrant usually isn't planning to buy anything.
Do this week: re-rank your last ten posts, not by likes, but by the four metrics below. The order will change, because you're measuring something else.
Saves — who is planning to come back
Meta's own reference defines the "saved" metric as the number of times a user has saved that specific piece of content, as documented in Instagram's metrics reference. A save is different from a like: the person is filing the post away to act on later, not reacting to it now. Price lists, delivery terms and step-by-step guides tend to get saved; an entertaining clip tends to get liked and forgotten.
The split runs differently by business type. An online store tends to see product-catalogue posts get saved, because people return to compare before buying; a service business tends to see pricing and process explainers get saved, because the decision takes longer and needs comparing.
Do this week: sort last month's posts by save count. Take the three most-saved topics and revisit them — the interest is already proven, you don't need to test it again.
Shares — someone else is advertising you for free
The same reference defines "shares" as the number of times a post got sent to someone else. A shared post reaches new people without spending ad budget, and it carries a recommendation effect a paid ad can't buy — a friend's forward is trusted differently than a banner.
Instagram also splits a public repost to your own followers from a private share sent to one person — both get counted separately in the same reference. Either way, the effect is the same: "someone I know showed me this", not "a brand paid to show me this" — a kind of trust a banner ad can't buy.
Do this week: split last month's posts into two groups — informational (how it works, why it matters) and promotional (discount, campaign). See which group gets shared more, and plan next month's content around that format.
Profile visits — the gap between curiosity and a step
The same documentation defines "profile visits" as how many times your profile was opened, and "profile activity" as what a person did once there — tapping a link, tapping the contact button. Someone can watch a video and scroll past it, but visiting your profile already means they've started looking into you. That behaviour is much closer to a website visitor than to someone who just liked a post.
The most common reason for a low ratio is the bio itself: if it isn't obvious what you do, who it's for, and what to do next, people open the profile and leave without tapping anything. The contact button needs to sit where it's impossible to miss, not buried on the third line.
Do this week: put profile visits for the week next to that week's DM count. A low ratio points at the profile itself — the bio, the contact button — not at the content: traffic is arriving but not taking the next step.
DMs — the conversation that gets lost between the numbers
The platform's main dashboard doesn't count direct messages; they just sit in someone's phone. But opening a DM isn't a random act — the person is already asking a question, already asking about price. That is the actual start of a sales conversation, it just never shows up in a report, because the platform doesn't surface it as a public metric.
Once the volume grows, tracking it from memory stops working — most people can reliably hold ten or fifteen conversations in their head before questions and posts start blurring together. That's the point where a plain spreadsheet stops being enough and a lightweight CRM entry earns its keep.
Do this week: log which post or story each DM came from, even in a plain spreadsheet. By month's end you'll know which content type brings in real enquiries and which one just collects likes.
A like is a thumb pausing for a second; a save is a promise that someone is coming back.
Why a like-led report always looks good
Likes and reach show up in the platform's main dashboard at a glance, climb on their own as you post more, and need no explaining. Saves, shares, profile visits and DMs have to be pulled from separate places and compared by hand — few people have time for that, so the report keeps the easiest number to find, not the most accurate one.
A good monthly report should carry these five lines:
- Likes and reach — for context, not for decisions
- Save count, and the three most-saved posts
- Share/repost count, and which format worked
- Profile visits and what happened after the visit
- DM count, and which post each one came from
That's not a failure on anyone's part — pulling four different metrics together by hand every month takes real time, and someone has to actually do it. Our reporting system that pulls every source into one place takes that job over: saves, shares and profile visits arrive ready each week instead of getting assembled by hand before every report.
