Creator, Studio and Agency plans from $19/mo, each with a 7-day free trial. Start your trial
by Phronesis AI Lab
← All posts

Measure saves, not likes

Likes are the metric everyone still glances at first, and the one that tells you the least. A like is cheap — a thumb-tap on the way past. A save is a decision: this is worth coming back to. Across formats, saves and shares track future reach far more closely than likes do, because they’re the signals platforms use to decide what to show next.

What we watch instead

  • Saves per post. The clearest signal that content has lasting value. Rising saves usually precede rising reach.
  • Shares. Distribution you didn’t pay for. One share reaches a new audience the algorithm hadn’t given you.
  • Retention (video). How far people watch. A three-second hook that holds to the end beats a clever one that loses everyone at second four.
  • Best slot, per channel. Not a generic “best time to post” — your audience, your account, your data. It moves, so it’s worth re-checking.

How do you read saves properly?

Raw counts mislead, because reach changes from post to post. Use a rate instead: saves divided by reach. Do the same for shares.

A worked example. Say post A reaches 2,000 people and earns 40 saves. That’s a 2% save rate. Post B reaches 8,000 people and earns 60 saves — more saves, but a 0.75% rate. Post B won the headline number; post A earned its saves at nearly three times the rate. A is the one to learn from. It probably wasn’t a weaker post, just a less-distributed one.

Compare like with like

Formats are built for different jobs. A how-to carousel is made to be saved. A joke or a hot take is made to be shared. A video lives or dies on retention. Comparing a meme’s saves to a checklist’s saves tells you nothing.

So compare within a format, and over a rolling four weeks rather than a single post. One outlier is luck. Three in a row is a pattern.

Do likes still matter at all?

A little. Likes are a quick, early read on whether a hook landed, and a sudden drop can flag a problem with reach. They just aren’t worth optimising for. A post with plenty of likes and almost no saves or shares was pleasant to scroll past and easy to forget. That’s fine now and then. It shouldn’t be the backbone of your plan.

Turning the number into an action

Data only matters if it changes next week. When a format over-performs, make more of it and cut what’s flat. When a slot consistently wins, move your important posts into it. When saves climb on a topic, that topic is a content pillar asking to exist.

A fifteen-minute weekly review

  1. Open the last seven days of posts with reach, saves, shares and, for video, average watch time.
  2. Work out save rate and share rate for each post.
  3. Rank them within each format. Note the top two and the bottom two.
  4. For each winner, write one line: the topic, the format, the hook.
  5. Pick one change for next week — more of a topic, less of a format, or a move into a better slot.
  6. Put that change into the calendar before you close the tab.

Quick checklist before you move on:

  • Rates, not raw counts.
  • Formats compared only with themselves.
  • Patterns over four weeks, not one post.
  • Every insight turned into one scheduled action.

The point of analytics isn’t a dashboard you admire on Friday. It’s a short list of decisions for Monday: make more of this, less of that, at this time. Everything else is decoration.

Run your brand's social like a studio

Start a 7-day trial — schedule a month of posts, get approvals, and publish everywhere from one calendar.