Instagram head Adam Mosseri described four ranking principles in September 2026: originality, recency, showing public posts to non-followers to test them, and larger AI models that weigh more signals. The practical plan is to publish original work often, test formats with Trial Reels, and judge success by engagement rate and reach to non-followers rather than raw likes. This is based on press coverage; I could not open Mosseri's original Reel.
Key Takeaways
- The four principles, as reported: originality, recency, "breaking" content shown to non-followers, and bigger models weighing more factors.
- Reposting without adding value is a losing strategy; original creators are favoured.
- Fresh posts get a boost, so schedule around when you can publish consistently, not around a mythical perfect time.
- Non-follower testing means every public post is a small audition. Hooks matter more than follower count.
- Track engagement rate and non-follower reach, and do not read causation into platform-reported Trial Reels data.
What Mosseri actually said (and where I'm unsure)
Social Media Today's 24 September 2026 summary lists the four principles and three posting tips. Social Media Today reports:
- Originality: Instagram favours original content and its original creator over reposts.
- Recency: newer posts get more exposure.
- Breaking content: posts from small, public creators are shown to non-followers to gauge response.
- Model size: expanded AI models weigh more factors across a broader range of posts.
The tips: track engagement rate instead of total views or likes, experiment (Trial Reels help), and post as often as you can without burning out.
A True Future Media breakdown agrees on the four principles, adding that Mosseri said the share of daily impressions from posts under 24 hours old has increased, with no figures given. It dates itself "Sep 9" with the year not shown, while Social Media Today is dated 24 September, so the original statement's date is unclear. I would say "September 2026" and leave it there.
Principle 1: Originality
Instagram says it prefers the account that made something. For brands that means:
- Stop reposting memes and aggregated clips as a core strategy.
- If you share a customer post or a trend, add a real contribution: your data, your example, a different angle.
- Keep source files and watermark-free exports so your version is the original.
This is not new policy so much as a restatement. Instagram announced on 30 April 2024 that aggregator accounts that repeatedly repost without material changes would be removed from recommendations, that originals would replace identical reposts, and that smaller creators would get wider initial distribution, with exceptions for licensed publishers (per Social Media Today).
I covered the related demotion of low-value content in AI slop and platform demotion.
Principle 2: Recency
New content is favoured, so frequency and timing matter more than polishing one hero post per fortnight. A realistic cadence for a small team is three to four Reels or carousels a week. Do not copy a number from an influencer; pick one you can hold for twelve weeks. My own approach is outlined in my Q4 social growth plan.
Principle 3: Non-follower testing
If public posts are tested on people who do not follow you, then your first two seconds and your caption must work for strangers. Practical moves:
- Open on the payoff, not the logo.
- Write the on-screen text as if the viewer has never heard of you.
- Use specific topics, so Instagram can match your post to the right audience.
- Look at the share of reach from non-followers in Insights for each post.
Trial Reels
Per Meta's announcement (published 10 December 2024, updated 26 June 2025) and the Instagram creators blog, Trial Reels show a Reel to non-followers first, hide it from your grid and followers' feeds, give you views, likes, comments and shares about 24 hours later, and let you share it with everyone manually or automatically if it performs well within 72 hours. Meta said in June 2025, as relayed by True Future Media, that among over 400,000 creators, 40% posted Reels more often after trying Trials and, of those, 80% increased non-follower reach. That is platform-reported, a before-and-after comparison, and True Future Media itself notes it does not show posting more causes better outcomes. Use it as a reason to test, not as a promise.
Principle 4: Larger models, more signals
I cannot tell you which signals a larger model weighs; Instagram has not published a full list. Sprout Social's explainer (published January 2026, modified 28 September 2026) names sends per reach, views and an aggregator penalty as top signals, but it is a third-party summary, so treat it as a hypothesis. Instagram's own Help Center is the place to check official wording. What you can do is stop optimising for a single metric. Aim for content that holds attention, gets sent to friends, and earns saves.
A 4-week content plan
| Week | Focus | Output | What I'd check |
|---|---|---|---|
| 1 | Baseline | 3 original Reels on one topic, 1 carousel | Non-follower reach share, engagement rate |
| 2 | Hook test | Same ideas, 3 different openings, via Trial Reels | Which opening keeps viewers longer |
| 3 | Format test | Carousel vs Reel vs Story-led post | Saves and sends per reach |
| 4 | Double down | Repeat the winner, retire the loser | Trend across weeks, not single posts |
Change one variable at a time or you will not know what worked.
How to calculate engagement rate
Mosseri's tip is to track engagement rate. True Future Media's advice is to show the count alongside the rate and state the denominator. I use engagement divided by reach for posts, because followers-based rates punish accounts that win non-follower reach. Whatever you choose, keep it constant for the month.
Hooks for strangers: a quick template
Because non-followers see your post first, I draft three openings for every Reel and pick the one a stranger would understand without context:
- The result first: "This one caption change doubled our saves." (Only say it if true and documented.)
- The question: "Why does your Reel die at 300 views?"
- The mistake: "Stop doing this in your first second."
Then keep the middle tight and end with a reason to send it to someone, such as a checklist or a template. Sends are not confirmed by Mosseri as a ranking principle in these reports, but Sprout Social's summary treats them as a strong signal, and a post people forward is rarely a bad post.
What not to do
- Do not buy the idea that a posting time unlocks reach.
- Do not read one viral post as a strategy.
- Do not post daily if quality drops. Mosseri's own advice carries a burnout warning.
- Do not treat platform-reported studies as evidence for your account.
How I'd adapt this for an edtech account
At Masai School, the work I was part of grew Instagram from about 26K to 117K, a team effort I would not promise to repeat elsewhere. What carried over was simple: one audience, repeatable original formats and constant testing. The four principles describe the same loop.
Reporting to a client
Give them three numbers monthly: non-follower reach share, engagement rate on reach, and number of original posts published. Add one sentence on what you will change. If the client asks for a "guaranteed reach", decline.
FAQ
What are Instagram's four ranking principles?
As reported, they are originality, recency, testing public posts with non-followers, and larger models weighing more signals. They come from Mosseri's explanation covered in September 2026.
When did Mosseri say this?
Social Media Today dated its coverage 24 September 2026; another outlet shows "Sep 9". I could not open the original Reel, so I cannot pin the date.
Does posting more often increase reach?
Mosseri advises posting as often as you can while avoiding burnout. That is advice, not a guarantee, and quality still matters.
What is a Trial Reel?
It is a Reel shown first to non-followers so you can test it. Meta's June 2025 numbers on it are platform-reported.
What engagement rate should I aim for?
There is no benchmark in the sources I opened. Compare your own posts over time.
Do reposts get penalised?
Instagram says it favours original creators over reposts, so reposting without added value is risky.
Does this change the best time to post?
Recency matters, but I found no evidence for a universal best time. Use your own Insights.
Is the algorithm one system?
Not necessarily. Feed, Reels and Explore may behave differently, and Instagram has not published every detail.
Should brands pay for reach instead?
That is a separate decision. See my post on Meta One subscriptions.
Work with me
I am Younus Fardeen, an India-based organic growth, SEO and AEO strategist with 4+ years of marketing experience in edtech and startups. If you want a content system built around how platforms really rank, see my work and reach me through the contact form at younusfardeen.in.