Your warm-up tool is lying to you

Silent churn is costing you money. Merlin rewards your best readers with points, perks, and tiers. Apply for the beta at getmerlin.app

When you run a warm-up sequence, the tool sends emails between seed accounts. These are mailboxes the vendor controls. They open every email, click every link, pull messages out of spam, and generate replies on a schedule.

The mailbox providers see consistent sending volume and consistent engagement. No complaints. Clean authentication. The domain looks healthy.

But those seed accounts aren't real people. They have no browsing history, no purchase behaviour, and no prior relationship with your brand or anyone else's. They exist for one purpose: to simulate engagement. And the major providers have been aware of this for years.

Google and Microsoft don't just count opens. They look at who is opening, how quickly, from what devices, and whether those recipients interact with anything else in their inbox. Seed accounts fail every one of those checks.

What you built over those 30 days is a reputation with a network of managed mailboxes. That reputation does not transfer to your actual subscribers.

Most tools give you a deliverability score at the end of the warm-up. Something like 94/100. Green across the board. Ready to send.

That score reflects how your domain performs with seed accounts, which are mailboxes the vendor controls. It says nothing about how you will perform with real Gmail users, real Outlook users, or the actual people on your list.

The problem has a name. Warmup tools grade their own homework. Global inbox placement averages 84%. Gmail sits at 87.2%. Microsoft drops to 75.6%. If your tool is showing you 95%, it is measuring delivery to its own network, not yours.

The moment you switch to real sends, you are starting from scratch on the signals that actually matter: engagement rate from genuine recipients, complaint rate, read time, and unsubscribe behaviour. If those first real sends go badly, with low engagement and a few spam clicks, you have a damaged reputation and no buffer to absorb it.

I have seen this play out more than once. Domain scores green on every tool. First real campaign hits 60% spam placement. The warm-up worked exactly as designed. It just solved a different problem than the one the sender had.

The clearest evidence of where this is heading came from GMass.

GMass ran the most popular email warm-up system on the internet. At its peak, over 80,000 active accounts. Over two years, it sent 1.3 billion warm-up emails across 236,000 accounts.

In January 2023, they shut it down. Not because the product failed commercially, but because Google told them to shut it down or lose Gmail API access entirely.

The GMass founder wrote about it publicly. His framing: warm-up systems put strain on Google's infrastructure in service of fooling Google's infrastructure. Google's position, as he described it, is that they do not want warm-up happening at all and now consider it a violation of their terms of service.

The tools that survived switched from the Gmail API to IMAP. The founder's assessment of that workaround: Google knows you are doing it and will hold it against you.

That is not a theoretical risk. That is the company that built the biggest warmup product in the space telling you, in writing, that Google has drawn a line.

Then in November 2025, Gmail escalated enforcement further. Previously, non-compliant email got routed to spam. Now Gmail rejects it at the SMTP level, bounced before it even reaches a spam folder. Most warmup tools have not updated their guidance to reflect this. The scores they are showing you are based on a deliverability environment that no longer exists.

GMass is not the only signal worth paying attention to.

Apollo, one of the biggest platforms in cold email outreach, dropped warmup entirely in 2024. They replaced it with volume pacing: controlled ramp-up with no fake engagement. No seed accounts, no simulated opens, just a gradual increase in real sends.

That is a significant move from a major player. Apollo did not quietly deprecate the feature. They replaced it with something philosophically different, an acknowledgment that artificial engagement signals are not worth building on.

Postbox Services went further. They tested nearly every major warmup tool on the market and found no measurable improvement in open rates and no significant lift in Google Postmaster reputation scores. Their explanation: automated openers running on cloud infrastructure create detectable patterns. The engagement is repetitive, the templates are formulaic, and Google's machine learning is good enough to distinguish bot opens from human behaviour.

The warm-up industry responded by adding features. Read emulation, scroll simulation, time-on-message tracking, and topic matching between warm-up emails and your real content. None of these have been shown in independent testing to move the needle. They are marketing differentiators, not deliverability improvements.

They are not useless, and it is worth being clear about that.

A warm-up sequence gets your domain aged, confirms your authentication is set up correctly, and keeps you from triggering new-domain filters on day one. These are real problems and the tools do help with them.

The issue is that senders treat the end of the warm-up as clearance to send at volume to a cold list. It is not. The tool gets you to zero. Getting above zero requires real engagement from real people.

There are two scenarios where some form of warm-up still makes sense: a brand-new domain with zero sending history, and reputation recovery after a bounce spike or complaint surge. In those cases, a gradual volume increase with careful monitoring can help stabilise things. But that is a floor, not a ceiling, and it should be done with real sends to real engaged recipients, not seed accounts.

Sending to people who want your emails. That is the whole answer. There is no infrastructure workaround that replaces it.

If you are starting a new domain or recovering from a damaged one, start with your most engaged segment. If you have to re-permission people, do it. Volume is not the goal yet.

Watch Google Postmaster Tools from day one. That is where your actual domain reputation lives, not your warm-up tool's dashboard.

Keep a close eye on complaint rate. One bad batch can wipe out weeks of clean sending. Gmail's published threshold is 0.3%. Experienced senders target under 0.1%.

Ramp volume manually. Warm-up tools automate the schedule because automation is easier to sell. Manual ramping, where you are actually watching the signals and adjusting based on what you see, is more effective and does not risk triggering the detection patterns that automated tools now commonly produce.

They give senders permission to skip the hard part. Building a list of people who genuinely want to hear from you takes time. Warming up a domain takes 30 days and a monthly subscription. One of those things feels like progress. The other one is progress.

The tool gives you a green score and you feel like the work is done. It is not.

There is also a $200 million industry with a strong financial interest in maintaining that feeling. The tools will keep adding features, smarter AI, better emulation, more convincing dashboards, because that is what the business model requires. What the business model does not require is telling you that Google has explicitly called this practice a terms of service violation.

Reputation is built send by send, with real recipients, over time. A dashboard number does not change that.

Warm-up tools have a narrow, legitimate use case: getting a new domain off the ground without triggering day-one filters. They do not build the kind of reputation that survives contact with a real list. They do not transfer credibility from seed networks to Gmail users. And they are operating in an increasingly difficult environment, with Google having moved from tolerating them to rejecting non-compliant senders at the infrastructure level.

If your warm-up finished and you are still seeing placement problems, the tool is not going to fix it. Reply and tell me what you are seeing. That conversation usually makes the diagnosis clear within a few minutes.

Before You Go: Here's How I Can Help

— Pavel