Why has my Instagram reach dropped?

Reach is the number of separate accounts that saw a post at least once. Views is every time a post was shown or watched, including the same person coming back for another look. So the two move differently, and a headline number that has gone down can mean two quite different things.

What you compare against matters too. One post that travels lifts a whole month, and the month after it can look like a collapse when it is really the ordinary level coming back. Before you decide anything is wrong, put a longer window on the screen and check whether you are comparing this fortnight to your normal or to your best.

The content itself. In what I see, this is nearly always the opening. Posts that used to open on something a stranger needed resolved started opening on a greeting, a logo, or a slow run-up. Nothing about the account changed; the first second did.

The cadence. You posted less, or you posted in lumps. Fewer posts is fewer chances to be shown to anyone, and an irregular pattern also robs you of the comparison you need to work out what is landing.

The format mix. Instagram is clear that it does not have one algorithm: Feed, Stories, Explore, Reels and Search each rank with their own, tailored to how people use that part of the app. Swapping your carousels for reels, or your reels for photos, changes which system is deciding how far your work goes. That often shows up as a reach change with no drop in what you made.

Who your posts were reaching. Follower share of reach is the slice of the people you reached who already follow you. The rest are strangers meeting you. When reach falls but your followers are still watching, what has usually shrunk is the stranger half, and that is a distribution question rather than a "my audience left me" one.

A quiet stretch. This one is not in any platform document, it is just what I see across the accounts I coach. School holidays, a long weekend, the end of the financial year, a week where everyone in your niche is on the road. Audiences go quiet in patterns that have nothing to do with you, and they usually come back on their own.

A run of posts that did not earn a stop. Not bad posts. Fine posts, three or four in a row, that nobody saved and nobody sent on. Instagram says the predictions it leans on most for reels are whether you will reshare it, watch it all the way through, like it, and go to the audio page. Reshare is the first one on that list. A stretch of posts that nobody passed on is a real thing, and it is something you can change in the next post.

Two things from Instagram's own explainers are worth knowing.

The first is mechanical. Instagram names some of the reasons it makes a reel less visible, and the ones it names are mechanical rather than personal: low resolution, a visible watermark, muted audio, borders around the frame, a video that is mostly text on screen, or something already posted on Instagram. Its creator guidance adds that it is less likely to recommend reposts, watermarked content, or accounts that mostly collect and reshare other people's work, and that a reel has to be three minutes or less to be recommended at all. If you switched to editing in another app and started exporting with a watermark, that is a cause with a fix, not a mystery.

The second is the shadowban question. Instagram's stated position is that it does not suppress content to push people into buying ads, and that Account Status in the app is where you check whether your account and your posts are eligible to be recommended at all. Nobody outside the platform knows the exact weights, and anyone who tells you what a save is worth next to a like is guessing. But a rumour in a comment section and what the company publishes about its own systems are different things, and you may as well start with the second.

Read these in order. Each one rules something out.

Open your insights and widen the time frame so you are comparing like periods. Then break your reach down by followers and non-followers, which Instagram points creators at in its own guidance on reach, and watch how each half moves. If the follower half held and the non-follower half fell, your posts stopped travelling to strangers and you are looking at content and format. If both fell, look at what you published and how often first.

Next, count what actually went out, week by week and format by format. A good share of the "sudden" drops I look at have a gap sitting in the middle of them that nobody remembers taking.

Then look at watch time and retention on your videos. Watch time is how long people stayed, added up; retention is the shape of the same thing, and the curve shows you the second they started leaving. If they go in the opening moments, it is your hook. If they stay and then drift, it is the middle.

Then saves and sends. A save says the post was worth keeping. A send says somebody put their own name to it. If those two have gone flat while your likes are steady, you have a "nice post, no reason to act" problem.

Last, check Account Status, so you can stop wondering.

Change one thing, not five. If you rebuild your openings, keep your posting and your formats steady for a few weeks so the result means something. If you go back to a cadence you dropped, hold it long enough to read it, and pick the one you can keep for months rather than the heroic one you can keep for a fortnight.

Leave your slow posts up. Deleting them takes away the record you need. Answer your comments. Build a reason to save or send into the thing itself, whether that is a checklist, a straight answer or a line that makes someone think of a particular friend. The making side of all this, hooks, pillars, batching and the weekly checklist, is in my free guide Scroll Less, Grow More, and it will keep you busy for a while.

There is a point where you have looked at your own insights so often that you stop seeing them. That is what the audit is for: one payment of A$300, no membership, and within five business days of your details reaching me you get a written read on your account as a private page, with the honest answer up front, every finding tied to the post or the number it came from, and a fix list in order. There is a full example report on the site, free to read, so you can see the depth before you spend a cent.