The Reputation Recovery Playbook

A step-by-step recovery plan built from 71,581 real degradation events across 3.7 years of the Warmup Inbox network. Pairs with the recovery predictor tool.

The one rule that governs recovery

Duration beats depth. How long an inbox stays degraded predicts recovery far better than how far it fell.
Day 15 is the line. Recover before it and you are looking at 7 to 34 days. Cross it and recovery jumps to 57 to 74 days.
The odds change too. Short events recover about 75 to 80 percent of the time. Past 15 days, it is roughly a coin flip.

In this article

  1. The finding that should change how you respond to a drop
  2. The recovery timeline, by severity
  3. Step 1: Detect early (watch Gmail first)
  4. Step 2: Diagnose the cause before you touch volume
  5. Step 3: The first 72 hours
  6. Step 4: Days 4 to 14, the closing window
  7. Step 5: If you have crossed day 15
  8. Predict your recovery before you commit

The finding that should change how you respond to a drop

When placement drops, most senders reach for the wrong dial. They panic about how bad the drop is, when they should be racing against how long it lasts. We know this because we measured it. Walking the day-by-day history of about 117,000 inboxes over 3.7 years, we detected 71,581 degradation events and classified each by depth and duration. The result is unambiguous: duration is the driver of recovery, not depth.

That single fact reorganizes the entire response. A deep, scary-looking 40-point drop that you catch in three days recovers in about a week. A mild 10-point drop you leave alone for two weeks can take more than two months. The Playbook below is built entirely around getting you to act inside the window where action still works. For the full data story behind the threshold, see The 15-day cliff.

The recovery timeline, by severity

This is the map. Find the row that matches how long the inbox has been degraded, then the column for how far it fell. The cell shows median days to recover and, in parentheses, the share of events that recovered at all.

Time below threshold10 to 20pp drop20 to 40pp drop40+ pp drop
3 days or less9 days (81%)8 days (76%)7 days (80%)
4 to 7 days16 days (73%)14 days (72%)13 days (76%)
8 to 14 days34 days (62%)27 days (64%)21 days (70%)
15+ days74 days (52%)74 days (51%)57 days (51%)

Source: Warmup Inbox network, 3.7 years, 71,581 detected events. 33% of events are censored (recovery not observed) and reported honestly.

Two things jump out. Read across any row and the numbers barely change: depth hardly moves recovery time. Read down any column and they explode: duration changes everything. The jump from the 8-to-14-day row to the 15+ row is the cliff, recovery time roughly triples and the odds fall to a coin flip. Everything in this Playbook is designed to keep you in the top two rows.

Step 1: Detect early (watch Gmail first)

You cannot beat a 15-day clock you did not know was running. Early detection is the whole game, and the network data tells you exactly where to look: Gmail. In 84.7 percent of detectable degradation events, Gmail’s spam rate spiked first, about a day ahead of Outlook and three days ahead of long-tail providers.

  • Make Gmail placement your primary alert. It is the earliest warning light you have. If Gmail is drifting, treat it as the leading edge of a problem, not noise.
  • Set the alarm on primary-tab placement, not folder placement. Standard IMAP reads Gmail tabs as inbox. Watch true primary placement so a slide into Promotions does not hide as healthy.
  • Alert on rate of change, not just level. A move from 1 percent to 4 percent spam over a few days matters more than any single reading. Gmail escalates fast once it starts.

Gmail is unusual in that it is both the earliest and the harshest filter: when it flips, the typical event already shows about 30 percent of mail going to spam that same day, with no soft-warning stage. That is covered in depth in Gmail is the canary AND the harshest filter. For recovery, the implication is that your Gmail alert is effectively your countdown timer.

Step 2: Diagnose the cause before you touch volume

Recovery is not one lever, it is finding which input changed. Before you cut volume or pause a domain, run a fast differential. The point is to separate a sending-side change you control from network-wide noise you do not.

  1. Check whether it is only you. If placement dropped across many unrelated inboxes at once, it may be a network-level or receiver-level event rather than your reputation. Our history shows most big swings are not caused by policy events, so do not assume the worst.
  2. Reconstruct what changed in the last 7 to 10 days. New list source, a jump in volume, new copy or links, authentication changes, a new sending IP or domain. The trigger almost always sits in this window.
  3. Confirm authentication is intact. SPF, DKIM, and DMARC alignment. A silent break here mimics a reputation drop.
  4. Separate content complaints from infrastructure. Rising spam complaints point to list quality and relevance; sudden folder shifts with stable complaints point to sending patterns.

Step 3: The first 72 hours

This is the highest-value window in the entire Playbook. Events resolved inside three days recover in about 7 to 9 days regardless of how deep the drop was. Depth genuinely does not matter here, so do not let a scary number push you into an overreaction that prolongs the event.

  • Ease volume, do not slam the brakes. A hard stop can look like abandonment and removes the positive engagement signal you need to rebuild. Reduce to your most engaged, most reliable recipients.
  • Lean on known-good recipients. Concentrate sending on addresses that reliably open and reply. Warmup-style positive engagement is what pulls placement back.
  • Fix the identified trigger, one change at a time. If you changed three things, reverting all at once tells you nothing. Undo the most likely culprit first.
  • Hold the line on authentication and list hygiene. Remove hard bounces and unengaged addresses that could be feeding complaints.

Step 4: Days 4 to 14, the closing window

If you did not fully recover in the first 72 hours, you are still in a good position, but the window is closing. Events in the 4-to-14-day band still recover the majority of the time (62 to 76 percent), and median recovery runs 13 to 34 days. The job now is steady discipline, not new experiments.

  • Keep volume conservative and consistent. Predictable, engagement-heavy sending rebuilds reputation. Erratic volume resets your progress.
  • Do not introduce new variables. No new list sources, no new domains, no fresh copy tests until placement stabilizes. You are trying to isolate recovery, not add noise.
  • Track the trend daily against your Gmail baseline. You want to see spam rate falling back toward baseline and primary-tab placement climbing. A flat line for several days is a signal to escalate.
  • Treat day 12 as your decision point. If you are not clearly trending back by then, prepare for the possibility of crossing the cliff and plan accordingly rather than hoping.

Step 5: If you have crossed day 15

Past 15 days the math changes and honesty serves you better than optimism. Recovery time jumps to 57 to 74 days and the probability of recovering at all falls to about 51 percent. This is no longer a quick-fix situation, it is a rebuild, and it is worth deciding deliberately whether to rebuild the existing inbox or stand up a fresh one.

  • Commit to a long, patient re-warm. Treat the inbox as if it were new. Slow, engagement-first ramping over weeks, not days.
  • Weigh a clean restart. With recovery near a coin flip and two months of runway either way, a fresh inbox on a new subdomain is sometimes the faster path to reliable placement. Remember that domain age does not predict placement, so a new domain is not a handicap.
  • Remove the root cause permanently. A rebuild that repeats the original trigger just buys you a second cliff. Fix list acquisition and volume discipline first.
  • Use the predictor to set expectations. Model the likely recovery curve before you invest, so stakeholders know whether they are waiting days or months.

Predict your recovery before you commit

Every step above depends on knowing where you are on the curve. That is exactly what the recovery predictor does: you enter how deep your drop is and how long it has lasted, and it returns the modeled recovery time and probability based on the same 71,581-event dataset behind this Playbook. It tells you, in plain terms, whether you are still inside the window or already over the cliff.

Use the tool

Enter your drop depth and duration. Get the modeled recovery time and probability from 3.7 years of real events, and a recommended action for your position on the curve.

Start with the recovery predictor, then work the steps above from wherever the tool places you. And if you want the full network picture this Playbook sits inside, read the Deliverability Benchmark Report 2026.

Related reading

Frequently asked questions

How fast do I really need to react to a placement drop?

Fast. Events caught and resolved within three days recover in about a week regardless of severity. Once an inbox stays degraded for 15 or more days, recovery time jumps to 57 to 74 days and the odds of recovering fall to roughly 50 percent. The first two weeks are where recovery is won or lost.

Should I stop sending entirely when placement drops?

Usually not. A hard stop removes the positive engagement signal you need to rebuild reputation and can read as abandonment. Ease volume down to your most engaged recipients instead, fix the trigger, and keep consistent, engagement-heavy sending.

Why should I watch Gmail specifically?

Because Gmail is the earliest warning signal. In 84.7 percent of detectable degradation events, Gmail spiked first, about a day ahead of Outlook. It is also the harshest, so a Gmail alert is effectively your recovery countdown starting.

Is it better to fix a badly degraded inbox or start fresh?

If you are still inside the first two weeks, fix it: recovery is fast and likely. Past the 15-day cliff, with recovery near a coin flip and about two months of runway either way, a clean restart on a new subdomain is often the faster path. Domain age does not predict placement, so a new domain is not a disadvantage.

What does the recovery predictor actually tell me?

You enter how deep your placement drop is and how long it has lasted. It returns the modeled recovery time and probability drawn from 71,581 real degradation events, plus a recommended action based on where you sit on the curve.

The Reputation Recovery Playbook, from the Warmup Inbox network. Recovery figures reflect 71,581 detected degradation events over 3.7 years. Individual results vary; modeled outcomes are guidance, not guarantees.