Half of What Instagram Recommends Is Under a Day Old.

On Meta’s second quarter 2026 earnings call on 29 July, CFO Susan Li said something that should land harder with operations leads than with creative teams: “On Instagram Feed, over half of all recommended content is now less than one day old, more than double from a year ago.”

That is a distribution number, but it is an approval problem. If most of what Instagram recommends was made in the last twenty four hours, the practical window for a reactive post has narrowed to about a day. A team whose sign-off loop takes four days cannot reach that window, and no improvement in the creative will change the arithmetic.

What Meta actually said

Precision matters here, because the figure is narrower than it first sounds.

Li was describing recommended content inside Instagram Feed, meaning the posts the ranking system inserts from accounts a person does not follow. She was not describing the whole feed, and she did not say older content stops being distributed. Just under half of recommended Feed content is still more than a day old.

The mechanism she gave alongside it is the part worth reading twice. Meta has put real time infrastructure behind its ranking models so they can, in her words, “identify high-quality new Reels at creation”, and every public Reels and Feed post on Instagram is now processed through a large language model and analysed on dimensions from topic to tone.

Put plainly, Instagram can form a view on a new post faster than it used to. So it has less reason to wait before recommending one.

Why this is a workflow number, not a creative one

The obvious reading is a content brief: be more reactive, make things about today, post faster.

That reading is correct and useless. Almost every brand already wants to be more reactive. What stops them is not taste or ideas, it is the days between having the idea and pressing publish, which is why a good approval system is a growth asset rather than an admin one.

Freshness is the one input that better creative cannot buy back. A post that is excellent on day six still enters the pool as day-six content.

The arithmetic of a four-day sign-off

Run the clock on an ordinary approval chain. The idea lands Monday afternoon. Copy and edit on Tuesday. Internal review Wednesday. A stakeholder is travelling, so comments arrive Friday morning. Brand or legal looks at it Friday afternoon. It publishes Monday.

Nothing there is broken. That is a completely normal process, and it is six days.

Every hour a finished post spends waiting is an hour outside the window that now holds most of Instagram’s recommended Feed inventory.

Sort your content by risk, not by format

Most approval chains apply identical gates to everything, because they were built around a content calendar rather than around exposure.

Do the sort on paper. Take last month’s posts and mark each one against two lanes:

• Low risk: reactive posts inside categories already agreed, community replies, formats already signed off, evergreen library posts, footage from the shoot you already approved, reposts of your own content.
• High risk: anything that makes a claim, names a third party, involves money, people, safety or legal exposure, or touches a live issue.

The argument for a fast lane gets much easier once the low-risk pile is sitting on the table in front of the people who currently approve it.

What a standing permission actually looks like

A standing permission is a written, bounded licence to publish without a second sign-off. It is not a request to trust the social team more, and it is not a verbal “just go ahead”.

One page, five things:

1. Who holds it, by name, not by job title.
2. Which surfaces it covers, listed individually.
3. Which content classes are in scope, taken from the low-risk lane.
4. What is explicitly out of scope, with an escalation route and a named person at the end of it.
5. When it is reviewed. Quarterly is a sensible cadence.

Then have the person who would have been the second approver sign it. That signature is the whole point of the exercise: it moves the decision from per-post to per-category, which is the only way approval stops costing days. It is also the first document we tend to build in workflow and ops consulting, because it converts a bottleneck into a rule.

Who can post without a second sign-off

One accountable publisher per surface, named. Not a rota, not a role, not “the social team”.

Shared accountability is slow accountability. When three people can publish and none of them owns the surface, every post gets checked by somebody as a courtesy, and courtesy checks are what a four-day loop is made of.

A second pair of eyes on copy is still worth having. Make it a peer read measured in minutes, and be explicit that it cannot block a post in the fast lane.

What genuinely still needs review

Speed is not an argument for removing scrutiny from the things that deserve it. Keep a hard gate on:

• Anything carrying a factual, performance or pricing claim.
• Anything naming or picturing a person who has not agreed to it.
• Anything involving a partner, a rights holder or licensed footage.
• Regulated categories: financial, medical, alcohol, gambling, anything with a statutory line.
• Anything published during a live incident or a complaint that has gone public.
• Anything that commits the company to a date, a launch, a refund or a policy.

Those are the posts where a day late costs nothing and wrong costs a lot. Defending that list properly is what earns the fast lane its credibility internally.

What good looks like

A team that can move inside a day has three things, and none of them are software.

• A ceiling rather than a target: a written maximum time from idea to live for the fast lane, and someone whose job it is to notice when it slips.
• A pre-approved shelf: hooks, formats, templates, cleared music and cleared footage, so a reactive post is an assembly job instead of a fresh commission.
• A visible queue: everyone can see what is waiting on whom. Most approval delay is not disagreement, it is a post sitting unread in somebody’s inbox.

Then measure the only number that matters here: median hours from idea to live, split by lane. If nobody on your team knows that figure, that is the finding.

How NBK thinks about approval speed

Platform changes like this one get read as content news when they are almost always operations news. Instagram has not told brands to make better posts. It has described what its system is currently surfacing, and the constraint on responding sits inside the company, in a sign-off chain most teams have not audited since the day it was built.

When distribution tightens, the reflex in a lot of boardrooms is to reach for budget. Spending does not shorten a sign-off loop. Organic reach is decided by the part of the system you actually control: what you publish, how fast it gets out, and how consistently.

Next step

If a genuinely good idea takes four days to reach a feed, the idea is not the problem.

Write down the current chain, post by post, and mark where each one waited and who it waited on. Our free social media approval process template is a reasonable starting point for that map.

If the chain turns out to be slower than the window your content is competing in, NBK can help rebuild the workflow behind the content.