Cold email without warm-up: what to do instead

Email warm-up is the practice of sending automated mail between a pool of mailboxes before real outreach, so a new mailbox gets a track record. Google’s published guidance asks for something simpler: start with low volume and increase it slowly. You can follow that with a low per-mailbox cap and stepwise increases on real campaigns, plus rules that pause a mailbox when bounces or complaints rise. Warm-up does not guarantee inbox placement, and neither does this approach.

This is a position article. Dooxout does not run warm-up, by design, and this page explains the reasoning, what is known and what is not.

What warm-up tools do

Vendors describe warm-up in similar terms. As one example, Instantly’s warm-up page says its system sends, receives, opens and replies to emails between a customer’s connected accounts and a network of other inboxes, with volume rising on a slow ramp that starts at 2 emails on day one. The page also makes a performance claim about inbox placement.

Two points to hold on to:

• This describes a product. It is not a statement from Google, Yahoo or Microsoft.
• The performance figures on vendor pages are the vendor’s own. We have not verified them, and we do not repeat them here.

Warm-up tools are a legitimate commercial category with a real search audience, and many teams use them. The point of this article is not that they are useless. It is that you do not have to take their promise on trust, and that they do not cover the things that most often damage a cold sender.

What the mailbox providers say

I checked the primary sender pages for Google, Yahoo and Microsoft on 2026-10-05 for anything about warm-up.

What follows from this:

1. The one concrete ramp instruction from a provider is Google’s: low first, then slowly up. It is about volume, not about fake engagement.
2. Google speaks of “engaged users”. Cold recipients are not engaged yet, so this guidance cannot be read as permission. It tells you the shape of a safe ramp, not that cold volume is welcome.
3. No provider page we found says that automated opens and replies between pooled mailboxes help or hurt. We do not claim either. If you read an article claiming certainty in either direction, ask for the source.

What actually decides whether a new mailbox works

Reputation for a sender comes from what recipients and receivers observe: authentication results, bounces, spam reports, replies, and how the volume behaves. The providers publish limits on the first three and guidance on the last. A new mailbox has no history, and the way to produce a good one is to send mail people do not complain about, at a rate that grows gradually.

Things warm-up cannot fix:

• A poor list. Sending to invalid addresses produces bounces, and sending to the wrong people produces complaints.
• Missing authentication. A mailbox without SPF, DKIM and DMARC is a candidate for rejection whatever its activity history. See SPF, DKIM and DMARC for cold email.
• No unsubscribe path. People who cannot leave report the message instead.
• A sudden jump. A warmed-up mailbox that goes from a handful of messages to hundreds in a day is exactly the spike Google warns about.

What to do instead: a capped, stepwise ramp

The method has five parts. It uses your real campaign volume, so every message sent has a purpose.

1. Start with a low cap on every new mailbox

Set a daily cap on each mailbox that is low enough that one bad batch would not matter. We do not give a number here. No provider publishes one, and a figure from us would be a guess presented as a rule. Choose a value you would be comfortable losing the mailbox over.

2. Send the strongest targets first

Early messages should go to the best-fit, verified contacts. A small number of well-matched messages produces replies and few complaints, and it gives you a clean signal on whether the mailbox and domain work.

3. Step up on evidence

Raise the cap only after the previous period ran clean: low bounces, no complaints, normal authentication results. Do not double volume in one step; Google’s example names that as a risk. Increase in smaller steps than you think you need.

4. Set a ceiling far below the provider limit

Provider limits are not targets. Google Workspace documents 2,000 messages per user per day and Exchange Online 10,000 recipients per day, as covered in cold email infrastructure guide. A cold mailbox sending anywhere near those numbers is not cautious.

5. Pause automatically

A rule that freezes a mailbox when bounce or complaint signals cross a threshold you set is more valuable than any warm-up schedule. Without it, a problem stays hidden until the domain is damaged. Spread volume across several mailboxes and domains so a pause costs part of your capacity, not all of it.

Comparing the two approaches

The last row is the important one. Neither approach is a guarantee, and a method that promises one is a warning sign.

Where warm-up might still make sense

If you already use a warm-up tool and see no problems, nothing here says you must stop. We have not run a controlled comparison, and we do not claim that warm-up has no effect. Our claim is narrower: a documented, capped ramp covers what the providers actually describe, and it removes a dependency on a third-party mailbox pool.

Two cases deserve extra care whichever approach you choose:

• Irregular sending. A few messages one week and a large batch the next is a spike by Google’s definition. Smooth the schedule instead of adding background traffic.
• A domain with history. If a domain was used before, review what it sent and whether it was ever blocked before you trust it with new volume.

How Dooxout handles this

Dooxout does not run warm-up and has no warm-up pool. A mailbox the platform provisions is ready to use at once. The risk of a young mailbox is limited by the server instead: each mailbox has a cap that grows as it gets older, there are caps per domain and per channel, and sending is frozen automatically when the bounce or complaint thresholds you set are crossed. These limits are enforced by the server, not by a prompt, so an agent running campaigns cannot raise them by itself.

The platform warns and recommends; the client decides. Unsubscribed and complaining addresses are suppressed and cannot be sent to, and that one rule cannot be switched off. More at guardrails and security and cold email infrastructure.