What it means
Account warm-up is a ramp. A new mailbox, a new domain, a fresh Telegram number or a new X account starts with no history, and history is the only thing a platform can use to decide whether you are a person or a problem. Warm-up manufactures that history deliberately: a small amount of activity on day one, a bit more each day, and enough real conversation along the way that the account accumulates evidence of being wanted.
The part most guides skip is the reason it works. Anti-spam systems are much better at detecting a step function than a high plateau. An account that sends 5 messages a day for a week and then 200 on a Tuesday is trivially detectable. An account that reaches 200 a day over three weeks, collecting replies as it goes, looks like a busy person. Both accounts sent 200 messages. Only one of them changed suddenly.
The second reason is the ratio. Every messaging platform weights the balance between conversations you start and conversations people start with you. Warm-up is partly about volume and mostly about giving the account a stock of two-way threads before it ever writes to a stranger.
The Telegram ramp, day by day
The schedule further down this page is the one we actually run. A few notes on why it is shaped the way it is.
Day 0 is deliberately empty. Registration plus a photo, a full name, a bio, a username and two-step verification, then nothing. Profile completeness is the cheapest trust signal there is, and an account that registers and messages strangers within the hour is the pattern the system is built to catch first.
Days 1 and 2 have no first contact at all. Join a handful of public groups you have a genuine reason to be in, read them, reply where you have something to add, and save real contacts to the phone book. Being in other people's contact lists is a trust signal you cannot buy.
From day 3 the account starts writing, first to people who will answer, then from day 6 to strangers. The cold column of the schedule matters more than the total. If the reply rate on the cold slice sits below roughly one in ten, hold the current level rather than advancing, because silence is what precedes reports, and reports are what precede a restriction.
The end state is 50 messages a day at no more than 20 an hour with randomised spacing, which is exactly the default our rate limiter ships with. Above that, add accounts. A single account pushed to 300 a day is worth less than six accounts at 50, because the six survive.
Expect flood waits during days 13 to 21. A short flood wait is feedback, not failure: the platform is telling you the exact number of seconds it wants. Obey it, do not retry early, and treat two in one day at the same level as a signal to stop climbing.
The email ramp, day by day
Email warm-up has a pre-flight phase that messaging does not, and skipping it wastes the entire ramp. Before the first send: a separate sending domain rather than the one your invoices come from, SPF, DKIM and DMARC published and verified, a custom tracking domain, a reply-to mailbox that a human reads, and one-click list-unsubscribe wired up. A domain registered yesterday, however carefully you ramp it, still looks like a domain registered yesterday.
Then the ramp itself, which raises volume by roughly a quarter per day rather than doubling, and which does not touch a cold list at all in week one. The audience for the first week is people who will answer. A reply is the strongest positive signal a mailbox provider can observe, and it is worth more than any number of delivered messages that nobody touched.
Cold prospects enter from around day 8, capped as a share of the day's volume rather than as an absolute, so the ratio of engaged to unknown recipients stays sane. By day 30 a mailbox sits at 40 to 50 a day of mostly cold traffic and stays there. Practitioner consensus puts the practical ceiling for a cold sending mailbox around that level, and scaling happens by adding mailboxes and domains. See deliverability for what to measure while the ramp is running.
The other channels
Be honest about what is knowable here. Telegram documents its restriction mechanism and will tell you through its own spam bot whether an account is limited. Almost nobody else does.
- X: new accounts face tighter limits than established ones, and daily DM allowances vary by account age and subscription tier. The published figures move often, so treat any number you read as a starting hypothesis and watch your own error responses instead.
- Instagram: action limits on follows, likes and DMs exist and are not published. The observable behaviour is a temporary action block, which arrives without warning and is the only reliable measurement you will get. Ramp conservatively and keep the DM copy varied.
- WhatsApp Cloud API: this one is genuinely documented. A new number starts on a limited messaging tier and moves up as it sends quality traffic, with a per-number quality rating visible in WhatsApp Manager. That rating is your warm-up dashboard, and it is the closest thing to an official warm-up metric anywhere in messaging.
What warm-up does not fix
Warm-up removes exactly one input from the risk model. It does nothing about the rest, and most restricted accounts we see were warmed up properly and then wrecked by something else: a scraped list of people with no reason to care, a link to a domain that has already been reported by other senders, an identical message body sent 400 times, or an offer that violates the platform terms outright.
The full stack is warm accounts, a list with a real reason to hear from you, varied copy through spintax, paced sending, honest opt-out handling, and content within the rules. Warm-up is one of six.
Why it matters
Account replacement is the hidden cost of outbound. A restricted Telegram number takes weeks to replace and takes its conversation history with it. A burned domain can poison a brand's email for months, and it is the domain, not the IP, that you cannot rent your way out of. Set against that, three weeks of ramp is cheap.
There is a second, less obvious payoff. Because the ramp forces you to run small volumes for two weeks, it doubles as a copy test. If your message cannot earn replies at 10 sends a day to a warm audience, scaling it to 500 a day will not fix it, it will only find out faster and louder.
Common mistakes
- Ramping volume but not conversation. A schedule followed to the letter with a zero percent reply rate produces a restricted account on time, just later.
- Doubling daily. Doubling is a step function with extra steps. Add a quarter.
- Batching the daily allowance into one hour. Fifty messages in twenty minutes is not fifty messages a day.
- Warming up on the primary domain. Cold sending belongs on a separate domain so a mistake never touches transactional or billing mail.
- Skipping authentication first. A perfectly paced ramp on a domain with no DKIM is three weeks spent teaching providers to distrust you.
- Ignoring the first flood wait. It is the cheapest warning you will ever get, and retrying early converts a short wait into a long one.
Related concepts
- Flood wait: the feedback signal that tells you the ramp moved too fast.
- Shadowban: what happens when you skip the ramp entirely.
- Deliverability: the instruments to watch while an email domain warms.
- Spintax: removes the duplicate-content input that a warm account cannot compensate for.
- Cold outreach: the activity the warm account exists to support.
- Double opt-in: the cleanest way to keep an email list from undoing the warm-up.
How Pinlyx handles it
Pinlyx enforces the ramp rather than trusting anyone to remember it. Per-account limits are configurable and ship at 20 messages an hour, 50 a day, with a minimum gap between sends and randomised spacing on top. Flood waits are caught, respected for the exact duration the platform asked for, and retried with backoff instead of hammered. Accounts in rotation carry a health score built from reply rate and platform errors, so a sender that starts sliding is pulled from the pool before the platform makes the decision for you. Newly connected accounts can be started on a reduced cap and stepped up on a schedule, which is the ramp above, expressed as configuration.