How to warm up a new social media account

What warming up actually is

A brand-new account has no history. Every platform's spam and integrity systems work partly on history: how long the account has existed, whether its behaviour looks like a person using an app or a script hitting an API, and whether other users engage with it or report it. An account with no history sits in the widest part of the uncertainty band, which is why new accounts get rate-limited, hit verification challenges, and see their first posts reach almost nobody.

Warming up is the deliberate accumulation of that history before you have anything to lose. You spend a week on cheap actions — scrolling, liking, following a handful of accounts, leaving comments a human would write — so that by the time you post content you care about, the account is no longer a blank slate. If something goes wrong during that week, you have lost a week and nothing else.

Notice what warming up is not. It is not a trick that unlocks reach, it does not guarantee your first video performs, and it does not make you immune from restriction — accounts that follow every ritual perfectly still get actioned, and accounts that post on day one sometimes do fine. It removes one avoidable category of problem at a stage where fixing it costs nothing.

What the platforms actually document

Very little of the warm-up folklore appears in official policy. Here is what does — the parts you can rely on rather than guess at.

Platforms publish spam and platform-manipulation policies that explicitly prohibit automated or bulk engagement: mass following, mass unfollowing, repetitive identical comments, coordinated engagement. These are enforceable rules, and they are the actual reason "don't mass follow" is good advice — not because a mysterious trust score drops, but because it is a documented violation.

Platforms also apply rate limits to following, commenting and posting. The limits themselves are generally not published, vary by account, and change. What is documented is that they exist and that exceeding them triggers temporary action blocks.

Reddit is the outlier — the one platform where the rules are genuinely written down, just not by Reddit. Individual subreddits publish their own requirements in their sidebar and wiki: minimum account age, minimum karma, restrictions on link posts, and automod rules that silently remove posts failing those checks. Those thresholds are specific and verifiable, because you can read them yourself before you post.

X documents its sensitive-media setting, which lets you mark your own posts so they sit behind an interstitial. X and several other platforms also have labelling expectations around synthetic and AI-generated media, which matter directly if your character is AI-generated. Beyond that list, you are in the territory of inference.

The shared method

The sequence below is community practice, not platform policy. It is widely followed, internally sensible, and costs you a week. That is the honest framing — not that it has been measured, but that the downside is small and the behaviour it produces is indistinguishable from a normal user.

The logic behind its shape is simple. Start with purely consumptive actions that carry no risk (scrolling, watching). Add low-volume reciprocal actions (likes, follows). Add actions that produce content other people see (comments, replies). Only then post your own material. Each step adds a slightly stronger signal than the last, and you never make a jump a real new user would not make.

Two habits matter more than the exact day count. Spacing: leaving minutes rather than seconds between actions is the easiest way to not look scripted, and it is free. Variation: identical comment text repeated across posts is one of the few behaviours named explicitly in published spam policies. Do the whole warm-up on the device you will actually use, from one location, one account per session — rapid switching between accounts, devices and geographies is a well-known trigger for verification challenges and is entirely avoidable.

Signals worth building that are not folklore

Some of what happens during a warm-up week is worth doing regardless of whether the warm-up theory holds, because it improves the account on its own terms.

Your feed is trained by what you consume. If you want your content shown to a particular audience, spending the first week engaging with content from and about that audience is the cheapest way to tell the recommendation system what neighbourhood your account belongs in. That follows from how recommendation systems work — they cluster accounts by behavioural similarity — and does not depend on any shadowban theory being true.

Filling out the profile properly (name, photo, a bio that reads like a person) is a conversion improvement, not a compliance one. People who land on a half-built profile do not follow it. The common advice to add the photo and bio around day seven is arbitrary — there is nothing wrong with day one. What matters is that it is done before your first post travels, because that is when strangers start arriving. Likewise, following a small number of relevant accounts gives you a feed worth scrolling; following hundreds you do not care about gives you noise and a ratio that looks bought.

What we could not verify

Several widely repeated warm-up claims have no source we could find. We have either dropped them or labelled them here, so this page does not date the moment a platform changes something.

The idea that platforms maintain a per-account "trust score" that warm-up behaviour raises is not documented anywhere. Something functionally similar almost certainly exists inside their integrity systems, but its inputs, weighting and response to a week of scrolling are unknown outside those teams. Treat any specific claim about it as speculation. Likewise, no platform states a required number of days before posting — seven to ten is a convention, not a threshold, and nobody has published a study.

Specific action ceilings — so many comments per hour, so many follows per day — circulate with false precision. Real rate limits exist, but they are unpublished, differ per account, and any figure you see quoted is somebody's guess from their own experience. Read them directionally: low volume is safer than high, spaced is safer than bursty. Not "three per hour is the limit."

Finally, the claim that platforms punish accounts specifically for linking out gets examined properly in the next lesson, because it deserves more than a bullet. The short version: the honest argument for waiting is about reach, not punishment.

The shared warm-up sequence, day by day

This is community practice rather than published policy, and the day numbers are a convention. Compress or stretch it to fit your schedule — the sequence matters more than the calendar.

1. Day 1 — consume only — Open the app and scroll for fifteen to twenty minutes. Watch things to the end rather than swiping instantly. Do not like, comment, follow or post. The only goal is that the account has a session history and a feed that has started to adapt to your niche.
2. Day 2 — light reciprocal engagement — Scroll again, then like a handful of posts and follow five to eight accounts that genuinely relate to your character's niche. Space actions out by minutes, not seconds. Still no posting and no comments.
3. Day 3 — start commenting — Add three to five comments or replies across the session, each on a different post, each with different text. Write something a human would write — a reaction, a question, a joke. Copy-pasted or generic comments are the specific behaviour named in published spam policies.
4. Day 4 — complete the profile and post once — Add the profile photo, display name and bio. Then publish one piece of content: a single image, a short clip, or a text post. Keep the caption simple. No link anywhere on the profile yet — the reasoning is in the next lesson.
5. Days 5 to 7 — hold a steady, unremarkable pace — One to two posts a day, plus the same level of commenting and liking as day 3. Vary your captions; do not reuse the same text. Post when your target audience is awake rather than when you happen to be free — for a US audience that means US evening hours, whatever timezone you live in.
6. Days 8 onward — move to your real cadence — From here the warm-up is over and you are simply running the account. Increase to whatever daily volume you can genuinely sustain. Consistency beats volume: an account that posts once a day for three months outperforms one that posts eight times a day for two weeks and then stops.
7. Whenever you post — check the platform's own rules first — On Reddit specifically, read the sidebar and wiki of every subreddit before posting to it. Its stated karma minimum, account-age minimum and link rules are enforced by automod and are the only warm-up numbers on any platform that are genuinely published.

What actually differs per platform

Everything above is shared. These are the genuine per-platform differences worth adjusting for. Where a number appears, it is community practice unless the row says otherwise.

Frequently asked questions

How long does warming up a social media account really take?

There is no published minimum, and no platform states one. The seven-to-ten-day convention you see everywhere is community habit, not policy. What the week buys you is behavioural history accumulated at a point where losing the account costs you nothing, which is a sensible trade rather than a proven mechanism. If you compress it into three days you are probably fine; if you stretch it over three weeks you have lost nothing but time. The sequence matters more than the calendar: consume before you engage, engage before you post, and post before you ask anyone to leave the platform. Where a real, checkable number exists — a subreddit stating a minimum account age or karma level in its rules — follow that number precisely, because it is enforced automatically and it is the only warm-up threshold anyone actually publishes.

Is shadowbanning real, or is it a myth?

Both parts of that question have honest answers. Platforms do reduce distribution of content and accounts without notifying them — this is documented in policies about reduced or ineligible-for-recommendation content, and it is a real mechanism. What is not documented is the folk model built on top of it: the idea of a specific per-account trust score that particular actions raise or lower by known amounts. Nobody outside those integrity teams knows the inputs or weightings, and anyone quoting exact figures is guessing. The practical consequence is that you should follow published rules carefully and treat everything else as risk management rather than physics. Avoid behaviour that published spam policies explicitly name — bulk following, repeated identical text, automated engagement — because those are enforceable. Do not reorganise your whole launch around unverified mechanics.

Do I need a separate email and phone number for each account?

A separate email per account is genuinely worth doing, and it has nothing to do with warm-up theory. It is basic operational hygiene: recovery flows, security notifications and support tickets all run through the email address, and sharing one inbox across several accounts turns a single compromise into several. Enable two-factor authentication on each while you are there. Phone numbers are harder — several platforms allow one number across accounts and some do not, and the rules change. Where a platform allows account creation without a phone number, you can defer it. What matters more than the credentials is the access pattern: signing into the same account from one device in one location looks unremarkable, while rapid switching between accounts, devices and geographies is a well-known trigger for verification challenges and is entirely within your control to avoid.

Can I warm up several accounts at once?

You can, and most creators running more than one character do. The constraint is not a rule but a practical one: warm-up depends on behaviour that looks like a person using an app, and one person cannot plausibly run twelve accounts through a genuine daily session each. Two or three is comfortable. Beyond that the sessions get shorter and more mechanical, which is exactly the pattern you were trying to avoid. Keep each account on its own email, and avoid switching between them repeatedly within a single session on the same device — the pattern of one login handling many accounts in rapid succession is one of the clearer automation signals available to a platform. If you genuinely need scale, stagger the launches rather than warming all of them the same week, so no single week's activity looks like a batch operation.

What should I actually post during the warm-up week?

Post the content you intend to post afterwards, just less of it. There is no benefit to filler — posting deliberately weak content to "season" the account wastes the only asset a new account has, which is that its early posts are shown to a small test audience whose response shapes later distribution. Your day-four post should be something you would be happy to have pinned. Keep captions short and varied, avoid stacking hashtags, and do not reuse identical text across posts, which is one of the few behaviours named directly in published spam policies. One consideration specific to AI characters: several platforms now expect synthetic or AI-generated media to be labelled, and doing it from your first post is easier than retrofitting disclosure onto an account with hundreds of posts.

Related reading

• Where to post AI influencer content
• Building an AI influencer for Instagram
• Building an AI influencer for TikTok