Most cold email programs do not have a deliverability problem until they suddenly have a catastrophic deliverability problem. The drop from healthy to broken is rarely gradual; it is a cliff that arrives on a Tuesday morning when reply rates fall by 60% and nobody knows why. The cause was almost always visible in monitoring data three to five days earlier; the team just was not watching the right signals.
This post is the operational playbook for catching deliverability problems before they crater your campaign. It covers the 8 specific signals worth tracking daily, the alerting thresholds that warrant action, and the recovery protocols when a signal breaks.
The mailbox providers do not announce reputation changes. Gmail will not send you an email saying "we have decided to route your sends to the spam folder." Microsoft 365 will not flag the moment your sender score drops below the threshold for inbox placement.
Deliverability degradation is observed indirectly through behavioral signals: bounce rates rising, open rates falling, replies declining without copy changes. By the time you see the symptoms in your reply data, the underlying reputation damage is days old and partially baked in.
The teams that maintain healthy deliverability do not react to symptoms. They monitor the upstream signals that predict symptoms and intervene before reputation damage compounds. The 8 signals below are those upstream indicators.
Bounce rate measures hard bounces (recipient address invalid) plus soft bounces (temporary failure) as a percentage of total sends. Healthy cold email programs run at 1% to 2.5% bounce rate. Anything above 3% is a problem and above 5% is a crisis.
Why it matters: bounce rate is one of the most direct reputation signals to mailbox providers. Sustained bounce rates above 4% will tank inbox placement within days, regardless of anything else you do.
Common causes: - List quality issues (unverified emails, scraped data, outdated records) - DNS authentication breaking (SPF/DKIM/DMARC misconfigured) - Sending domain newly registered without proper warmup - Catch-all destinations bouncing with delay
The fix: pause sending immediately if bounce rate crosses 3.5%. Run the list through a verification service (NeverBounce, ZeroBounce, MillionVerifier). Re-audit DNS authentication. Resume sending only when verified bounce rate is back under 2%.
Spam complaints are the single most damaging signal a sender can generate. Recipients hitting the spam button trains mailbox providers that your mail is unwanted. The threshold is much lower than people expect: 0.05% complaint rate is the upper edge of what is sustainable. Microsoft applies penalties starting at 0.08%; Gmail starts applying penalties slightly higher but the gap closes if complaints persist.
What this means in practice: sending 5,000 emails per day, you can absorb roughly 2 complaints per day before reputation damage starts. Three or more complaints in a single day from the same domain warrants a sending pause.
Common causes: - List was opt-in for one product, repurposed for cold outreach (terrible idea) - Subject line is misleading or feels spammy - Aggressive follow-up cadence that feels like harassment - Hitting the same prospect from multiple sending domains
The fix: identify the specific campaign or sequence where complaints are concentrated. Pause that sequence. Investigate copy and frequency. Do not resume that exact sequence; revise.
Inbox placement is the percentage of delivered emails landing in the primary inbox versus promotions, updates, or spam folders. This is measured through seed list testing or specialized providers like GlockApps, Litmus, or built-in inbox placement testing inside outreach platforms.
Target: above 92% for cold outreach. Below 88% is the threshold for alerting; below 80% is a crisis.
Why it matters: this is the closest thing to a direct measurement of your reputation. Open rates and reply rates are downstream of inbox placement; if placement drops, opens and replies drop with a 1 to 3 day lag.
The fix: identify the specific sending domain or mailbox where placement degraded. Reduce volume on that mailbox by 60%. Re-warm using a quality pool. Investigate any changes in copy, subject line patterns, or list source in the prior 7 days.
Engagement velocity measures how quickly recipients open and reply to your emails after receipt. A healthy mailbox shows median time-to-open under 3 hours and median time-to-reply under 12 hours. As inbox placement degrades, these numbers grow because the email is sitting in promotions or spam folders, getting found later or not at all.
This signal often moves before raw open rate does, making it a leading indicator. We have seen mailboxes where time-to-open doubled from 2 hours to 4 hours over a week, with open rate dropping only modestly during that window. Two weeks later, open rate had collapsed by 40%. The engagement velocity move was the warning.
The fix: treat engagement velocity changes as an early signal of placement degradation. Investigate the same things you would investigate for inbox placement changes; the diagnosis is usually the same.
Blacklists are the public databases of IP addresses and domains flagged as senders of unwanted mail. Listing on a major blacklist (Spamhaus SBL/CSS, Barracuda BRBL, SURBL) immediately suppresses delivery to a meaningful portion of recipient providers.
Check daily against at least these lists: - Spamhaus SBL, CSS, DBL - Barracuda BRBL - SURBL - SORBS DUHL, RHSBL
Most outreach platforms do this automatically. If you are running on your own infrastructure, set up a monitoring service that polls these lists.
The fix: if listed, identify the cause (almost always a recent campaign with high complaint or bounce volume). Submit a delisting request after fixing the underlying issue. Do not submit delisting requests without fixing the cause; you will be re-listed within 48 hours and the second listing is harder to clear.
Reply rate is downstream of all of the above signals. If reply rate drops without any copy or targeting change, deliverability is the most likely cause. Tracking reply rate trajectory per sending domain catches deliverability degradation that has not yet shown up in inbox placement or engagement velocity dashboards.
The discipline: track reply rate weekly per sending domain, not just per campaign. If domain A's reply rate is 3.2% week 1, 2.9% week 2, 2.1% week 3, 1.4% week 4, you have a deliverability problem on domain A even if your overall campaign reply rate looks stable (because other domains are propping it up).
The fix: pull domain A out of rotation. Investigate the upstream signals (bounce, complaint, placement, engagement velocity) for that specific domain. Address the root cause before reactivating.
A single mailbox going bad inside a fleet of 30 will hide in aggregate metrics. The fleet-wide open rate looks stable while one mailbox is silently underperforming and dragging down the overall reputation pool.
Run the distribution check weekly: rank all sending mailboxes by open rate, identify the bottom quartile, and compare each to the fleet median. Any mailbox more than 12 percentage points below the median needs investigation.
The fix: pause and re-warm the underperforming mailbox. Investigate whether it has unique characteristics (newer domain, lower-quality warmup, different list segment) that could explain the underperformance.
The foundation. SPF, DKIM, and DMARC must pass on every send for the receiving provider to treat your mail as authenticated. Intermittent failures (DNS server issues, expired DKIM keys, DMARC policy mismatches) cause provider distrust that compounds over time.
Daily check: pull a sample of recent sends and verify all three authentication records are passing. Tools like MXToolbox or built-in dashboards in mature outreach platforms automate this.
The fix: if any authentication record is failing, halt sending from the affected domain immediately. Verify DNS configuration. Re-deploy and re-test before resuming. Authentication failures are the single fastest path to deliverability collapse and the easiest to prevent with discipline.
A daily standing review of these 8 signals takes 8 to 12 minutes per day if your tooling surfaces them in one dashboard. The cost of skipping it is measured in weeks of degraded campaigns.
The minimum operational discipline:
Morning check (5 minutes): Glance at the 8 signals across the full sending fleet. Note any signals in alert state. If any are red, the team is on the problem before the day starts.
Weekly deep-dive (30 minutes): Review trajectory data across all 8 signals for the prior 7 days. Look for slow-moving trends that have not yet crossed alert thresholds but are degrading.
Quarterly fleet audit: Retire any sending mailbox that has spent more than 4 weeks below median performance. Replace with fresh, warmed mailboxes. The cost of running a degraded mailbox in your fleet is not just the underperformance of that mailbox; it is the drag on your aggregate reputation.
Our cold email deliverability guide covers the architectural side of why these signals matter at the provider level. Our warmup playbook covers how to bring new mailboxes online without breaking the fleet.
For teams sending under 500 cold emails per day across fewer than 5 sending mailboxes, manual monitoring with tools like MXToolbox plus GlockApps plus a spreadsheet is workable. Cost: roughly $200 per month of tooling plus 30 to 60 minutes per week of operational time.
Above 500 sends per day or 5+ mailboxes, manual monitoring stops working. Either you build internal dashboards (engineering investment) or you adopt an outreach platform with deliverability monitoring built in. We have made the case for the latter throughout this blog because the operational cost of building this from scratch exceeds the platform cost for almost every team.
OnyxSend tracks all 8 signals automatically per domain and per mailbox, with alerting thresholds tunable per fleet. When a signal crosses threshold, the platform throttles or pauses the affected mailbox automatically and surfaces the alert in the operations dashboard. The recovery protocols (re-warm, list verify, sequence pause) are integrated so the response is one click rather than a multi-tool scramble.
Deliverability is the foundation that all other outbound performance stands on. A team running disciplined daily monitoring catches problems while they are 5% issues; a team without monitoring discovers the same problems as 60% issues weeks later. The cost difference is enormous and the operational discipline is modest.
If you want to see how OnyxSend handles deliverability monitoring as part of the integrated outreach workflow, start a 14-day trial. Domain-level monitoring is live within 30 minutes of import.
<!-- onyxsend:related-reading -->
- Cold Email Warmup Playbook: How to Warm 30+ Mailboxes Without Killing Deliverability - Top 7 Instantly Alternatives for Cold Email in 2026 - B2B Cold Email Subject Lines: 35 Templates That Drive Opens in 2026 - Cold Outreach at Scale: 5 Systems Every B2B Team Needs in 2026 - B2B Outreach Platform Buyer's Guide for 2026 - The B2B Automated Prospecting Stack: From Cold List to Booked Meeting - We Analyzed 50,000 Cold Emails. Here's What Actually Works in 2026. - OnyxSend cold outreach services - OnyxSend case studies - OnyxSend API
<!-- /onyxsend:related-reading -->