Reputation, built
not faked.

The schedule is built around three principles. Engaged segments first, expanding outward as reputation builds. Volume increases by no more than 2× per day, never on a Friday or before a public holiday in the recipient cohort. ISPs warmed in parallel rather than sequentially — same daily volume distributed across Gmail plus Microsoft plus Yahoo plus Apple proportional to the eventual production mix. Warming up Gmail first and adding Microsoft in week three is one of the most common mistakes; the result is great Gmail placement and terrible Outlook placement when full volume eventually lands.

# Engaged cohort = opened or clicked in last 30 / 60 / 90 days
# Volumes are TOTAL across all receiving domains, distributed
# proportional to your production traffic mix.

Week 1  # 30-day engaged only. Hard floor.
  Day 1:    500      Day 2:   1,000     Day 3:   2,000
  Day 4:    4,000    Day 5:   8,000     Day 6:  12,000
  Day 7:   16,000    # Pause and review. Fri/Sat/Sun if needed.

Week 2  # 30-day engaged only. Continue.
  Day 8:   25,000    Day 9:  35,000     Day 10: 50,000
  Day 11:  70,000    Day 12: 90,000     Day 13:110,000
  Day 14: 130,000    # End of critical window. Reputation now established.

Week 3  # Expand to 60-day engaged. Add in 15% chunks.
  Day 15: 160,000    Day 16: 200,000    Day 17: 240,000
  Day 18: 280,000    Day 19: 320,000    Day 20: 360,000
  Day 21: 400,000

Week 4  # Continue expansion. 60-day engaged, full pool.
  Day 22: 450,000    Day 23: 500,000    Day 24: 560,000
  Day 25: 620,000    Day 26: 680,000    Day 27: 740,000
  Day 28: 800,000

Week 5-6  # 90-day engaged becomes safe. Full target volume.
  Days 29-42: ramp to 1,000,000/day at +5% per day,
              hold for 7 consecutive days at target before
              calling the warmup complete.

The schedule is the public version. The private version, which we run against live data, has 22 additional decision points where we either hold steady or accelerate or fall back depending on what we see in Gmail Postmaster Tools, Microsoft SNDS, and the per-IP placement results from the seed-list testing. The decisions look something like this.

# Read each morning at 09:00 Europe/Ljubljana, applied to the day's send.

if complaint_rate > 0.30% over_24h:
    action: pause_24h, investigate_segment, halve_volume_on_resume

elif complaint_rate > 0.15% over_24h:
    action: hold_volume_24h, monitor_closely

if bounce_rate > 2.0% over_24h:
    action: pause, re-verify_remaining_list, do_not_resume_until_clean

if gmail_postmaster_reputation == "Bad":
    action: pause_72h, investigate_engagement_drop

if microsoft_snds_filter_result == "Yellow":
    action: hold_microsoft_volume, increase_engaged_segment_share

if all_metrics_green for 3 consecutive_days:
    action: proceed_to_next_step_in_schedule

The thing the schedule does not show is the cohort work. The 30-day engaged segment in week one is not just "people who opened in the last 30 days" — it is the most active subset of that group, prioritised by recency-frequency-monetary scoring where it is available, by recency-clicks where it is not. Apple Mail Privacy Protection has made open rates an unreliable signal since 2021, so we lean on click events plus reply events plus conversion events as the primary engagement indicators. Subject lines in week one are deliberately plain — no aggressive promotional copy, no urgency triggers, no spam-shaped patterns that would amplify any borderline content filtering.

The schedule above is the plan. The chart below is what actually happens. We pulled three real warmups from the last quarter — anonymised by client, identified only by send profile — and overlaid the planned daily volume against the actual achieved volume. Two of the three tracked the plan within 5 percent. The third hit a wall on day 11 when Yahoo started deferring 30 percent of injected messages, and the curve shows the recovery — three days at reduced volume, then resumption on the planned trajectory by day 16. The shape of the recovery is what most warmup vendors do not publish, because it requires admitting that warmup is not deterministic.

Three datasets on the chart. The dotted teal line is the planned curve from our standard 42-day playbook — starts at 50 messages on day one, doubles every two days through day 14, transitions to multiplicative scaling through day 28, then linear ramp to full volume by day 42. The solid lines are three actual client warmups. Client A tracked tightly. Client B underran the plan slightly because their list quality was better than typical and we accelerated cautiously. Client C hit the Yahoo wall on day 11. The dip on the Client C curve between days 11 and 14 is the recovery period — we cut volume by 60 percent across all receivers for 72 hours while reputation re-stabilised, then resumed. The interesting comparison is between Client B's smooth ramp and Client C's recovery — both ended up at full volume on schedule, but the path to get there was substantially different. Warmup is not a fixed sequence. It is a planned trajectory with feedback corrections.

Two takeaways for buyers comparing warmup services. First, the only honest warmup vendor is one that shows you what failure looks like and how it gets corrected. Anyone who publishes only smooth curves is either hiding their bad ones or has not run enough warmups to have any. Second, the recovery from a Yahoo hit on day 11 is not a magic technique. It is conservative volume reduction, patient observation, and resumption only after aggregate reports confirm the issue cleared, applied with the discipline of an operator who has watched the same shape replay across many clients. Three to five days lost. That is normal. Operators who claim faster recovery are doing one of two things wrong, and we have audited both kinds. What you are buying with a warmup engagement is the patience and the feedback discipline far more than the schedule itself.

Warmup-network tools work by adding your sending account to a pool of mailboxes that automatically open or click or reply to each other's mail. The pitch is that you build "engagement signals" that mailbox providers reward without needing real recipients. The reality is that the mailbox providers — Gmail and Microsoft especially — have been catching this pattern reliably since at least 2021. Validity's published analysis is direct on the subject: manufactured engagement lacks the normal negative signals (deletes, archives without reading, occasional spam marks) that real recipient behaviour produces. The traffic profile of a warmup-network mailbox is detectable, and traffic from senders associated with detectable warmup networks gets penalised.

The senders who use these tools and who do not get caught are typically running cold outreach at very low volumes from individual mailboxes — and the warmup network is propping up reputation that would not exist otherwise. The senders who try to use warmup networks for bulk marketing or transactional infrastructure burn the reputation faster than they would have without the tool. We have done two emergency recoveries this year for clients who tried to shortcut a warmup with a network tool — both took 60+ days of remediation work and one resulted in the IP being permanently retired because the reputation damage was not recoverable.

The reason real warmup works is not because mailbox providers are stupid. It is because they are evaluating a specific signal: does this sender have an audience that genuinely wants their mail? The proof of that is engagement from real recipients in their actual inbox patterns — opens followed by clicks followed by replies followed by occasional unsubscribes followed by re-engagement weeks later when something in the subject line catches them. That pattern cannot be faked at scale. It can only be earned by sending genuinely useful mail to people who asked for it, slowly, until the reputation builds.

The corollary is that we will not warm an IP for a sender who does not have a real engaged segment to send to in week one. If your list is purchased or scraped or so old that the 30-day engaged cohort is empty, the warmup will fail and there is no point starting it. We will tell you that during the intake call. The honest path forward in that case is list rebuilding — re-engagement campaign on a separate IP first, double opt-in for new sign-ups, then a fresh warmup once you have a 30-day engaged segment that is actually engaged.

Four failure modes show up across warmup engagements. Each has a different recovery path and a different cost in days lost. The playbook below is what we run internally — same checklist for every client because the patterns are stable enough that improvisation makes things worse, not better. Knowing the recovery path before you need it is the difference between a three-day setback and a complete restart.

1 — Early Yahoo soft-block (days 8-14)

The first is the early Yahoo soft-block, usually surfacing between day 8 and day 14. Yahoo is the most aggressive of the major receivers about new IP scrutiny — they evaluate engagement signals harder and faster than Gmail or Microsoft, which means warmup IPs hit Yahoo throttling earlier than other receivers. The signal is 421 4.7.1 deferrals on a meaningful percentage of Yahoo-bound messages, with retry logic eventually delivering most of them but with delivery latency stretched to several hours. The recovery is to cut Yahoo-specific volume by 60 percent for 72 hours while leaving Gmail and Microsoft volume untouched. After 72 hours, resume Yahoo at the volume level from three days before the throttle hit, not the volume that triggered it. Resume the planned ramp from there. Total cost: three days lost, full ramp completion delayed by three days. We see this in roughly one out of every four warmups.

2 — Gmail Promotions tab drift (days 18-25)

Second is the Gmail Promotions tab drift around days 18 to 25. The signal here is subtle — messages still deliver, but the open rate on Gmail traffic drops by 15 to 25 percent compared to the trajectory of the first two weeks. What is happening is that Gmail's classifier has decided your IP is acceptable for delivery but not for the primary inbox, and it has started routing your traffic to Promotions instead. There is no SMTP signal for this. The only way to detect it is monitoring open rate by receiver, which most senders do not. The recovery is two-pronged: reduce Gmail volume by 30 percent for one week to lower the priority signal weight, and immediately review the message corpus for promotional indicators (excessive imagery, link density, sender-name personalisation gaps). Most Gmail Promotions drift recovers within seven to ten days. We see this in maybe one out of three warmups, more often in B2C marketing senders than B2B transactional.

3 — Microsoft SmartScreen full block (days 20-35)

Third is the Microsoft full block, the rarest of the three but the most disruptive. Microsoft's SmartScreen filter is binary in a way that Gmail's is not — when it decides an IP is suspect, the result is 550 5.7.0 with delivery rejection across hotmail.com, outlook.com, live.com, and the entire Microsoft consumer mail estate. There is no soft warning. The block usually appears between days 20 and 35 of the warmup, almost always triggered by complaint rate exceeding 0.3 percent in a 24-hour window. Recovery from a Microsoft block takes between five and ten days and requires three things: complete pause on Microsoft volume, complaint pattern analysis to identify the trigger campaign or list segment, and submission through Microsoft SNDS to acknowledge the issue and request reconsideration. The pause is non-negotiable. Senders who try to push through a Microsoft block extend the block.

4 — When recovery is not the answer

Fourth is total reputation collapse, where the IP is failing across two or more major receivers simultaneously. This is rare in our engagements because the warmup discipline catches the precursors, but it does happen — usually because a list quality problem the client did not disclose surfaces in week three or four. When it happens, the only honest answer is to stop the warmup and start over. We have advised clients to abandon a warmup IP after day 19 once. Once. The IP was permanently flagged across Microsoft and Yahoo, recovery would have taken six to eight weeks at best with uncertain outcome, and the client decided correctly to provision a fresh IP and start the 42-day warmup again. Total cost: 19 days plus a fresh ramp. The lesson is that recognising when not to recover is as much a skill as the recovery itself.