The Instagram algorithm for shops: three signals Meta named, and the rest is rumour

The Instagram algorithm has three signals Adam Mosseri, the head of Instagram, named in January of last year: watch time, likes per reach and sends per reach. Everything else you read, the golden posting hour, the hashtag count, the claim that a send is worth five likes, is a guess or a copy of another blog.

The three signals the Instagram algorithm actually counts

Mosseri's wording was short: the top three ranking signals are watch time, likes and sends, likes counting slightly more for your own followers and sends for people who do not follow you.

• Watch time: seconds of your video really watched, replays included.
• Likes per reach: of every hundred who saw it, how many liked it. The raw count is not the signal.
• Sends per reach: of every hundred who saw it, how many sent it on.

Watch time: seconds are counted now, not views

Instagram made views the primary metric across every format two summers ago and retired impressions and plays. So a fifteen second reel watched three times beats a sixty second reel abandoned halfway.

Metricool measured close to 24 million posts from about 375 thousand accounts and found the average reel watch time is 8.5 seconds, more than double the year before. What you fight for is second three to second nine, and reels for an Instagram shop covers that.

Sends: the signal that brings strangers

Sending a post to a friend tells Instagram it is worth another person's time. Here is the rumour: blogs claim a send is worth three to five likes. That ratio has circulated since last year and nobody names a source. Mosseri said slightly more important, not five times.

Either way the behaviour is the same, and it has a cost: more sends is more DMs, and thirty open chats by evening is what running an Instagram shop is about.

It is not one algorithm, it is several rankings

That last row is the one most shops abandon, and Instagram SEO is only about it.

Hashtags no longer buy reach

Mosseri has said plainly that hashtags do not help reach, and following a hashtag was removed two Decembers ago. Metricool measured posts carrying hashtags at about 31.7 percent fewer views and 33.9 percent fewer interactions. That is correlation, not cause: newer accounts probably use more. Either way, thirty tags do not do what you think.

What pulls reach down

• Another app's watermark; Instagram says those reels are shown lower.
• Reposting others untransformed; aggregator accounts are recommended less.
• Several posts within minutes; Instagram holds some back to balance feeds.
• Flat single images; Socialinsider, across 35 million posts, measured engagement down 17 percent year over year.

Whether the answer is a carousel or a reel depends on your size, and the numbers are in carousel or reels.

Trial reels: the one official way to test without risk

1. 1Turn the trial option on as you publish; it stays off your grid and reels tab.
2. 2About a day later, read its views, likes, comments and shares.
3. 3If it performs, Instagram shares it to your followers, or you tap share to everyone.
4. 4If not, it stays in the trials tab and no follower watched it fail.

Instagram adds its own caveat: a follower can still meet a trial reel, sent to them or reached from that audio page. Hidden is not absolute.

Which number a shop should read

Engagement is not money. Socialinsider puts the platform average engagement rate at 0.48 percent: fewer than one viewer in two hundred likes or comments. The weekly number is how many conversations arrived and how many became orders, which is what the online shop management guide and registering orders from Instagram DMs are for.

Two places the ranking never touches

Nothing ranks your profile or stories. A stranger arriving from a reel decides in three seconds from your bio, and stories were never built to find new people. Both are covered in the Instagram shop bio and Instagram stories for selling, and selling on Instagram with no website says what a site would have done for you.

What to do tomorrow morning

Open your last ten posts in insights and write reach and sends beside each. Divide one by the other. One or two stand well above the rest. Rebuild that subject this week as a trial reel.