Instagram analytics: the eight numbers worth checking weekly

Why native insights mislead

The native app shows totals for a rolling window, mixes formats together, and resets your context every time you open it. That makes it good for a quick glance and poor for deciding what to make next.

Analysis needs comparison: this reel against your last ten reels, this week against the same week last month, this format against that one. Any tool worth paying for exists to make those comparisons trivial.

The eight numbers

Check these weekly. Anything else is monthly at most.

• Non-follower reach % — is the content escaping your existing audience?
• Saves per 1,000 reached — perceived usefulness
• Shares per 1,000 reached — social currency
• 3-second retention on reels — hook quality
• Average watch time — whether the middle holds
• Profile visits per post — curiosity converted
• Profile-visit-to-DM or link rate — intent converted
• Follower growth rate % — the lagging confirmation

Reading reels versus carousels

Judging every format on the same metric produces bad decisions. A carousel that reaches few people but is saved constantly is doing its job. A reel with huge reach and no saves is entertainment, not marketing.

• Reels — retention first, then shares, then profile visits
• Carousels — saves first, then reach, then DMs
• Stories — replies and link taps, never views
• Single images — comments and profile visits

Finding the pattern behind your winners

Once a month, line up your five best posts and five worst by saves per 1,000 reached, then look for what the winners share. It is almost never the posting time — it is the topic, the specificity, or the first line.

• Same topic cluster appearing repeatedly at the top
• Hook style: number, contradiction, or named audience
• Length: short and dense versus long and slow
• Whether the post asked for a comment keyword

Connecting analytics to money

Reach reports stop being useful the moment you sell something. The question changes from 'who saw this' to 'which post produced an enquiry'.

You get that by tagging: a unique keyword per offer in comment-to-DM flows, and tracked links per campaign. Then a post carries a number of conversations and, where checkout is tracked, a revenue figure.

• One keyword per offer so DMs are attributable
• Tracked links per campaign, not one link for everything
• Review conversations-per-post alongside reach every month

A ten-minute weekly review

Analytics fails when it becomes a research project. Keep it short and repeatable.

• 2 min — non-follower reach and follower growth versus last week
• 3 min — top three posts by saves per 1,000 reached
• 3 min — worst performer and one hypothesis for why
• 2 min — write next week's single experiment