Warm a new cold email domain in 21 days by ramping sends from 5 on day one to 100 on day 21, starting a warmup-to-cold ratio of 5:1 and tapering to 1:5 by week four, and never stopping warmup entirely. That sequence is calibrated for Gmail's post-November 2025 enforcement window, which moved spam-rate violations from temporary throttles to permanent 5.7.x rejection codes (Google Workspace Admin Help). Skip warmup or jump volume too fast and a new domain burns inside a week. This guide gives you the day-by-day schedule, the three signals Gmail actually measures, and the troubleshooting matrix when warmup goes wrong.
How long does email warmup take in 2026?
A new cold email domain needs 21 days of structured warmup minimum, and 28 days for high-stakes campaigns. Below 14 days, Gmail's filters do not register enough engagement history to classify the domain as legitimate. Above 30 days, you hit diminishing returns -- engagement signals plateau once Postmaster reputation reaches 'High.'
The 2026 timeline is longer than the 2-week schedules that worked in 2023. Three things changed:
- Gmail's bulk sender enforcement (October 2025 update) added permanent 5.7.x rejection codes for senders crossing 0.3% spam complaint rate (Coldreach Gmail Spam Rules).
- Domain reputation now outweighs IP reputation in Gmail's filtering decision (Bluecore Postmaster Guide). A clean Google Workspace IP does not save a cold domain.
- Postmaster Tools needs 100+ sends per day to display data, which means you can't measure reputation until late in the warmup window.
For reference, the global average inbox placement rate is 84%, but Gmail leads at 87.2% while Microsoft Outlook bottoms out at 75.6% (Truly Inbox 2026 Statistics). Warmup tools optimize for Gmail signals first because that's where the volume and the measurement live.
What is the right daily send ramp for a new cold email domain?
Use this 21-day ramp: Day 1 = 5 sends, Day 7 = 20 sends, Day 14 = 50 sends, Day 21 = 100 sends. Never grow daily volume by more than 20% in a 24-hour window, even if engagement signals look strong. Gate every step on three checks before increasing.
Day-by-day warmup schedule
| Day | Total Sends | Warmup | Cold | Required Signal Before Advancing |
|---|---|---|---|---|
| 1 | 5 | 5 | 0 | DNS records (SPF/DKIM/DMARC) verified |
| 3 | 8 | 8 | 0 | 80%+ inbox placement on seed test |
| 5 | 12 | 12 | 0 | Warmup reply rate above 25% |
| 7 | 20 | 18 | 2 | Zero spam-folder landings in 48 hours |
| 10 | 30 | 25 | 5 | First reply from cold send |
| 14 | 50 | 35 | 15 | Postmaster reputation visible (Medium+) |
| 17 | 70 | 35 | 35 | Spam rate below 0.05% |
| 21 | 100 | 25 | 75 | Postmaster reputation 'High' |
| 30 | 100 | 20 | 80 | Maintenance -- hold this ratio indefinitely |
The critical inflection is Day 14. That's when Postmaster Tools first surfaces meaningful data and when Gmail's filtering algorithm starts treating the domain as established rather than provisional. Hitting 50 sends on Day 14 without seeing a 'Medium' or 'High' reputation tier means something upstream is broken -- usually DMARC alignment or a warmup network Google has fingerprinted (MailReach Postmaster Guide).
What is the correct warmup-to-cold email ratio?
Start at 5:1 (five warmup sends for every cold send) in week one and taper to 1:5 by week four. The warmup-to-cold ratio matters more than absolute volume because it controls the reply-rate signal Gmail watches.
The taper curve:
- Week 1 (Days 1-7): 5:1 ratio. Warmup creates virtually all the engagement signal. Cold sends are rounding error.
- Week 2 (Days 8-14): 2:1 ratio. Warmup is still dominant but cold replies start contributing to reputation.
- Week 3 (Days 15-21): 1:2 ratio. Cold volume overtakes warmup. Warmup acts as a stabilizer when cold replies dip.
- Week 4+ (Day 22 onward): 1:5 ratio. Maintenance mode. 15-25 warmup sends per inbox per day, indefinitely.
Why the taper instead of a clean cutover? Cold email reply rates average 3.4% in 2026, with top-decile performers at 10%+ (Cleanlist 2026 Cold Email Data). Warmup networks deliver 30-35% reply rates by design. Cutting warmup before the domain has Postmaster 'High' status means your aggregate reply rate collapses from 30%+ to under 5% in a single day. That collapse looks identical to spam behavior in Gmail's classifier.
Should you keep warmup running once you're at full volume?
Yes. Cut warmup to 30-50% of peak volume but never turn it off. Postmaster reputation drifts from 'High' to 'Medium' within 8-12 weeks of stopped warmup, even on domains with strong cold reply rates.
The maintenance setting that holds reputation:
- 10-15 warmup sends per inbox per day
- 30-35% reply rate on warmup messages
- 15% spam-rescue rate (warmup network deliberately marks some as spam, then rescues them)
- Send window matching cold campaign hours (don't run warmup at 3am if cold sends go out at 9am)
The mechanism: every cold campaign generates negative signal -- ignored opens, deletes-without-reading, occasional spam complaints. Without offsetting positive signal from warmup, the negative accumulates and reputation slides. Once 'Medium' is locked in, recovery takes weeks (Bluecore).
This is also why teams that scale to multiple inboxes need warmup running on every mailbox, not just primaries. Inboxes on the same domain share reputation. One cold-only inbox without warmup drags down the rest.
What warmup signals does Gmail actually measure?
Gmail measures three signals during warmup, in this order of weight: manual move-from-spam, reply rate from previously-engaged inboxes, and time-in-inbox before archive or delete. Opens are weakly weighted after Apple Mail Privacy Protection (MPP). Clicks help but aren't necessary.
1. Manual move-from-spam (highest weight)
When a recipient drags an email from the Spam folder to Inbox, Gmail registers an explicit override of its own classifier. This is the strongest positive signal a recipient can send (Suped Email Deliverability KB). Warmup networks fake this deliberately -- a 15% spam-rescue rate is the industry standard.
2. Reply rate from warmed inboxes
Replies are Gmail's most reliable engagement signal because they require active intent and produce a measurable conversation thread. Warmup tools generate replies from real human inboxes (MailReach, Warmy, Instantly all run inbox networks of 100K to 1M+ accounts). Target: 25-35% reply rate during the 21-day ramp.
3. Time-in-inbox
Gmail tracks how long an email sits in the inbox before action (open, archive, delete, reply). Emails archived within 2 seconds register as spam-like. Emails opened and then left in inbox for hours register as valuable (Demand Gen Report on Gmail's Smarter Inbox). Warmup networks simulate this by leaving messages unopened for 30-90 minutes before triggering reads.
What Gmail does not measure heavily during warmup: open rate (pixel-tracked opens are MPP-poisoned), click rate (most cold sends should be link-free in the first touch), or bounce rate from valid scrubbed lists.
Why isn't my warmed domain getting good deliverability?
The three most common failure modes are DMARC alignment errors, spam triggers in cold copy that warmup never tested, and warmup networks Google has fingerprinted. Run through this troubleshooting matrix before assuming the warmup didn't work.
Check Postmaster Tools first. If domain reputation is 'High' but cold sends still spam-fold, the issue is your campaign, not your warmup. If reputation is 'Low' or 'Bad' after 21 days, the issue is upstream of cold sends -- authentication, warmup network quality, or send pattern.
Key diagnostic principle: warmup engagement and cold engagement should both contribute to reputation. If Postmaster shows 'High' during pure warmup but drops to 'Medium' the day you start cold sends at meaningful volume, the warmup signal was a Potemkin village -- the network reply rate masked an unhealthy underlying domain.
For a deeper diagnostic flow including SPF flattening, DMARC reporting setup, and BIMI configuration, see our cold email deliverability checklist for 2026.
Which email warmup tools should you use in 2026?
For Gmail-first campaigns, use MailReach, Warmy, or Instantly. For Outlook-heavy lists, add Mailivery or Lemwarm. Avoid any tool that uses AI-generated replies instead of real human inboxes. Google's classifier increasingly fingerprints templated reply patterns.
Quick selection criteria:
- MailReach -- strongest on inbox placement testing and Gmail-specific signal modeling. Best for Workspace domains pre-Postmaster visibility.
- Warmy -- 1M+ inbox network, 24/7 deliverability monitoring, strong on Outlook signal.
- Instantly -- bundled with sequencing, fits if you want one platform for warmup and cold sends.
- Mailivery -- transparent reply samples (you can see the actual warmup conversations), good for compliance-sensitive teams.
What to avoid: any tool advertising 'GPT-generated warmup replies' or 'AI conversation simulation.' These produce reply patterns that statistical classifiers detect, and Gmail has been adding pattern filters since late 2025.
See our full email warmup tools comparison for feature-by-feature pricing and inbox-network sizes.
Spam test: what good looks like after Day 21
After completing the 21-day ramp, run a Spam Test against the major mailbox providers before scaling cold sends. A clean result looks like 95%+ inbox placement on Gmail and Yahoo, 80%+ on Outlook, with all authentication checks (SPF, DKIM, DMARC) passing alignment.
What a clean post-warmup test report contains:
- Inbox placement breakdown: Gmail, Yahoo, Outlook, Apple Mail, ProtonMail percentages
- Authentication results: SPF pass + aligned, DKIM pass + aligned, DMARC pass
- Spam score: SpamAssassin score below 3.0 (lower is better)
- Blacklist check: clean across Spamhaus, Barracuda, SURBL
- Content analysis: no major spam triggers in subject or body
[Embed Spam Test screenshot here: showing Day 21 result with 96% Gmail inbox placement, all auth green, SpamAssassin 1.2]
A passing test is necessary but not sufficient. The test measures whether one email lands in the inbox at one point in time. Scaling to 75-100 cold sends per day per inbox stresses the domain in ways a single seed test cannot detect. The first 5 days of full-volume cold sends are still a probation period -- watch Postmaster spam rate hourly during this window.
| Symptom | Likely Cause | What Postmaster Shows | Fix |
|---|---|---|---|
| Warmup opens fine, cold sends land in spam | Cold copy contains spam triggers (links, money words, attachments) | Domain reputation: Medium, spam rate <0.1% | Strip links from first touch, swap HTML for plain text, A/B test new copy |
| Warmup reply rate above 30% but Gmail still filters | DMARC misalignment or missing BIMI/MTA-STS | Authentication: pass on SPF/DKIM, fail on alignment | Set DMARC to p=quarantine with sp=quarantine; align From: header domain with SPF return-path |
| Domain reputation stuck at Low after 21 days | Sent volume before warmup completed, or warmup network is detected and discounted | Reputation: Low, no spam rate data | Pause cold sends 14 days, run warmup-only at 25/day, switch to a network that mimics human behavior |
| Bounce rate above 4% | List rot or unverified scrape data | Delivery errors: 5.1.1 'user unknown' spikes | Verify list with NeverBounce or ZeroBounce; cap risk-tier sends at 10/day |
| Spam rate spikes above 0.3% | Bad targeting or unsubscribe friction | Spam rate: red, persistent 5.7.x rejections | Stop cold sending immediately, run warmup-only 7 days, add one-click unsubscribe header |
| Replies dry up after week 3 despite clean reputation | Send pattern too uniform; algorithmic flagging on cadence | Reputation: High, replies trending down | Add jitter to send times, vary subject templates, drop daily volume by 30% for 5 days |