New Instagram Accounts Play by Different Limits, and Every Guide You Read Was Written for Old Ones

You read a guide that said 500 follows a day is safe. You did 200 on a two-week-old account and got action-blocked.

The guide was not lying to you. It was describing a mature account, which is what almost every limits article silently assumes. And the gap between those two situations is the single most consequential thing nobody puts in the headline.

First: nobody actually knows the numbers

Instagram does not publish hourly or daily caps for follows, likes, comments, unfollows or DMs. Everything you can find is reverse-engineered, and the reverse-engineering disagrees with itself.

Here is the honest state of the third-party estimates for new accounts:

• Follows: around 100/day is the most commonly cited new-account figure. Hourly guidance clusters around 5/hour for new accounts versus 20/hour for older trusted ones.
• Likes: roughly 250/day for new accounts, up to about 1,000/day for established ones. Hourly guidance around 20/hour when new.
• Combined actions: a frequently repeated figure is a 500-action ceiling per 24 hours covering follows, unfollows and likes together during roughly the first 12-20 days.
• Total daily actions: active mature accounts described at up to 1,000/day; new accounts at roughly 100-500.

We have previously documented that third-party figures for mature accounts disagree with each other by a factor of three or more. That has not improved. So do not treat any of the above as a rule you can operate against.

Do not memorize these numbers. Memorize the shape: new accounts get roughly a third of the allowance, and the allowance is not the thing being enforced anyway.

The part that is reliable: account age is its own risk tier

Across sources that agree on nothing else, one structural claim is consistent: accounts in their first 0-3 months are treated as a distinctly stricter category, with thresholds that widen as the account accumulates age, history and genuine activity.

This is the finding worth acting on, and it reframes a common assumption. Operators running multiple accounts tend to think of scale as the risk — more accounts, more exposure. The evidence points somewhere else:

The dangerous moment is not when you have many accounts. It is when you have new ones.

Which is exactly backwards from how most people onboard. A new batch of accounts is when you are most eager to see progress, so it is when you push hardest — on precisely the accounts with the least tolerance.

It also explains a pattern that otherwise looks random: an operator runs twenty mature accounts without incident for months, adds five new ones, and takes restrictions on the new five while the original twenty carry on untouched. Nothing about the setup changed. What changed is that five accounts with no history got handed a schedule calibrated on accounts with plenty of it. The infrastructure was fine. The pacing was inherited from the wrong tier.

Why "pattern, not total" is the only rule that holds up

Every source that stops quoting numbers long enough to describe mechanics says a version of the same thing: enforcement responds to patterns rather than totals. Fifty follows inside five minutes gets flagged even if your daily count is nowhere near any published estimate.

This lines up exactly with Meta's own written policy. The spam standard describes acting "at very high frequencies" without ever giving a number, and adds that restrictions may apply at lower frequencies when other signals — repetitive content, inauthenticity indicators — are present. We covered that wording in detail in Meta's high-frequency rule.

Put the official text and the third-party mess together and you get a clean conclusion:

There is no number to stay under. There is a shape to avoid. Bursts, perfect regularity, identical intervals, and identical behavior across accounts are the shape. Spreading the same volume across the day is not a trick — it is the only thing the policy language actually describes.

This reframing is also why chasing the "real" numbers is a waste of effort. Even if someone handed you Instagram's actual thresholds, they reportedly vary by account age, history and activity level, which means the number that applies to you is not the number that applies to anyone writing a guide. Meanwhile the pattern rule is stable, cheap to implement, and does not expire when the thresholds get retuned. One of those two things is worth building a workflow around.

A warm-up approach that follows from the evidence

Third-party guidance converges on something like: publish 2-3 pieces of content and let the account settle for 2-3 weeks before any high-frequency activity. That is reasonable, but the reasoning matters more than the schedule.

A new account has no history for the platform to weigh. Everything it does is evaluated against nothing. So the goal of a warm-up period is not to "age" the account — it is to give the system a baseline that looks like a person.

Practically:

• Content first, actions second. An account with no posts doing hundreds of follows is a shape with no legitimate explanation. Post before you engage.
• Ramp, do not jump. Increase weekly rather than starting at your target. If mature-account guidance is 20 follows/hour and new-account guidance is 5, the useful information is the ratio, not either endpoint.
• Randomize intervals, not just volume. Same total, different spacing every day. This is where automation should be helping rather than hurting: any tool worth using should be doing actions in randomized ranges with jitter, not firing a fixed count on the hour. It is how NoobClaw schedules engagement by default, and it is the one design choice that maps directly onto what the policy text describes.
• Do not run identical patterns across accounts. Ten accounts doing the same thing at the same time is a stronger signal than any one of them exceeding a limit. This is the same underlying issue as device and network isolation — the behavioral layer is the one most people leave unsolved.

And one honest boundary: pacing reduces the risk of tripping frequency-based enforcement. It does not constitute permission. Meta's scripting carve-out applies to "authorized routes," and driving your own logged-in session is not covered by that phrase. Lower risk is real; approval is a different thing, and anyone conflating the two is selling you something.

FAQ

How long until my account counts as "established"?

No published answer exists. Third-party sources describe 0-3 months as the strict tier and treat thresholds as widening with age, history and activity — so it is gradual rather than a switch that flips on a specific day. The practical read: assume you are in the strict tier for at least the first month and ramp from there, watching for action blocks as your actual signal rather than a calendar date.

I got action-blocked. Did I permanently damage the account?

A temporary action block is generally a rate-limiting response, not an account-level penalty, and it typically clears on its own. The thing to avoid is the loop where you wait for it to lift and immediately resume the same pace — repeated blocks are a pattern, and patterns are what enforcement reads. If you want to know whether anything more serious is attached to the account, Instagram's Account Status panel is the place to look.

Does using a different device or IP per account solve this?

It solves part of it. Device and network separation address two of the three layers; the third is behavior, and it is the one that persists across any amount of infrastructure. Reddit's approach has been described as including keystroke rhythm and voting habits — that is what the behavioral layer looks like when a platform invests in it. If ten accounts on ten clean fingerprints all act identically, the fingerprints did not help.