Deliverability Ops in 2026: The SOP Your Agency Runs Every Monday (So Campaigns Don’t Die Quietly)

If you run outbound for clients, deliverability doesn’t “break.” It slowly bleeds out. Replies dip. Bounces creep up. Gmail starts soft-blocking. Then the client asks why pipeline died “out of nowhere.” It wasn’t out of nowhere. You just didn’t run the weekly checks.

This is the operator-grade weekly cadence agencies run every Monday. It’s a copy-paste deliverability SOP with thresholds and actions. No vibes. No dashboards-as-a-religion. Just “check this, if it’s red do that.”

TL;DR

• Run one weekly deliverability SOP per client, per sending pool (domain + mailbox group).
• Guardrails that matter: authentication alignment, bounce bands, complaint bands, blocklists, and inbox placement sampling.
• When reputation dips: triage, isolate, cool down, rotate, re-warm. In that order.
• Your goal is not “perfect deliverability.” Your goal is stable inboxing at a volume you can repeat.

---

What “Deliverability Ops” means in 2026 (and why agencies need a deliverability SOP)

Deliverability Ops is the repeatable operating cadence that keeps outbound email viable:

• You monitor sender identity health (domain, SPF/DKIM/DMARC alignment).
• You monitor recipient signals (complaints, engagement, bounces).
• You run controlled volume and segmentation rules.
• You stop fast when reputation dips, so you don’t torch the whole pool.

This is different from:

• A deliverability dashboard: dashboards show symptoms. They don’t run the recovery playbook.
• A sending architecture plan: architecture is the setup. Ops is the weekly maintenance.

One reality check: mailbox providers now enforce stricter standards for bulk and promotional traffic. Gmail’s sender guidelines tie mitigation eligibility to spam rate thresholds and require one-click unsubscribe for promotional traffic. They explicitly reference keeping user-reported spam rates below 0.3% and tracking it in Postmaster Tools. (Google sender guidelines FAQ)

Microsoft published requirements for high-volume senders (5,000+ emails/day) that include SPF, DKIM, DMARC, and unsubscribe expectations, with enforcement starting May 2025. (Microsoft announcement)

And one-click unsubscribe is not a “best practice.” It is a defined standard. RFC 8058 spells out the required headers, including List-Unsubscribe and List-Unsubscribe-Post: List-Unsubscribe=One-Click, and it requires DKIM coverage for those headers. (RFC 8058)

So yeah. Your agency needs a Monday SOP. Not a hope-and-pray.

---

The Monday deliverability SOP (weekly cadence)

Run this every Monday before you launch new lists or ramp volume. If you send daily, still run the full SOP weekly.

Time required: 30-60 minutes per client, per sending pool.

Inputs (minimum):

• ESP logs (bounces by code, delivered, opens if you still track them, replies, unsubscribes).
• Google Postmaster Tools (if you have enough Gmail volume to show data).
• Microsoft SNDS / Outlook-facing metrics if available (optional, but useful).
• Blocklist checks (multi-RBL scan).
• Seed/inbox placement sampling (small, consistent test).

Outputs (what you write down every week):

• Status: Green / Yellow / Red per pool
• Which lever you pulled (throttle, segment cut, domain rotate)
• What you will do next send day

Weekly status rubric (copy-paste)

GREEN

• Authentication clean (SPF/DKIM/DMARC aligned, no recent DNS changes)
• Hard bounce rate stable and low
• Complaints near-zero
• Inbox placement stable in your seed sample
• No blocklist hits that matter

YELLOW

• One metric trending wrong, but still recoverable with throttles and segmentation
• Examples: bounce rate creeping, Gmail spam rate flirting with threshold, seed shows more Promotions/Spam than last week

RED

• Complaint spike, hard bounce spike, or multiple red flags at once
• Seed inboxing collapses
• Major blocklist hit
• Action: stop rules kick in (later section)

---

Domain and mailbox health checklist (deliverability SOP core)

This is the stuff that kills campaigns quietly when nobody checks it.

1) Authentication and alignment (SPF, DKIM, DMARC)

Why it matters: mailbox providers don’t trust “I swear this is my domain.” They trust alignment.

Weekly checks

• SPF passes for the exact RFC5322.From domain (not just a parent domain)
• DKIM passes and signs the From domain (alignment)
• DMARC exists and produces reports
• No recent DNS edits from “someone on the client’s IT team”
• One-click unsubscribe headers present on outbound sequences where applicable

Google explicitly calls out From header alignment issues and ties mitigation eligibility to meeting requirements including authentication and spam-rate requirements. (Google sender guidelines FAQ)

Minimum guardrail

• DMARC policy can start at p=none, but it must exist.
• Don’t ship changes on Friday. You will spend Monday cleaning up.

Action if failed

• SPF fail: fix include chain, flatten if needed, remove dead includes.
• DKIM fail: regenerate key, verify selector, ensure ESP signs.
• DMARC missing: publish at least p=none with reporting.
• Alignment fail: fix From domain vs sending domain mismatch.

2) Domain age, history, and usage boundaries

Weekly checks

• No “domain repurposing” (stopped marketing, restarted at scale)
• No cross-client contamination (one domain should not carry two client brands)
• No sudden sending pattern change (volume spikes, new geos, new language)

Action if violated

• Isolate: move the offending segment to a separate pool.
• Cool down: throttle volume 50-80% immediately.
• Rotate: if reputation is damaged, stop using the domain for new acquisition.

3) Mailbox-level sanity

Weekly checks

• Each mailbox sends within its expected daily band
• No mailbox sending 10x more than peers
• No mailbox with abnormal bounce rate compared to the pool
• Reply handling works (no broken Reply-To)
• Unsub processing works fast (see unsubscribe section below)

Action if violated

• Disable the mailbox for 48 hours.
• Re-warm at lower volume after you fix root cause.

---

Bounce and complaint guardrails (with thresholds and next actions)

You need hard stop bands. Otherwise you will talk yourself into “just one more day” while your pool burns down.

Bounce rate bands (copy-paste policy)

Track hard bounces separately from soft/temporary deferrals.

Hard bounce rate (HB)

• GREEN: HB < 1.0%
• YELLOW: HB 1.0% to 2.0%
• RED: HB > 2.0% (or any single-day spike > 3.0%)

What to do

• YELLOW
  • Cut risky segments first: free inboxes, scraped contacts, old leads.
  • Enforce stricter verification: recent validation + role filtering.
  • Reduce volume 25-50% for 2 send days.
• RED
  • Stop new acquisition sends on that pool.
  • Isolate the list source that caused the spike.
  • Rebuild list: re-verify, re-enrich, re-segment.

Operator note: high hard bounces are usually your list, not your copy. Fix inputs.

Complaint rate bands (this is where you die)

If you send to Gmail at volume, the user-reported spam rate matters. Google explicitly references 0.3% and updates daily in Postmaster Tools, with mitigation eligibility tied to keeping spam rates below 0.3% for 7 consecutive days. (Google sender guidelines FAQ)

Complaint / spam rate

• GREEN: < 0.1%
• YELLOW: 0.1% to 0.2%
• RED: ≥ 0.2% (and “drop everything” at 0.3% in any major provider view)

What to do

• YELLOW
  • Cut broad personas. Keep only highest-fit segment.
  • Shorten sequences. Fewer touches, fewer chances to get flagged.
  • Add stronger upfront context line (why them, why now).
• RED
  • Trigger incident workflow: triage, isolate, cool down, rotate, re-warm.
  • Stop sending to the provider where complaints spike (often Gmail-first throttling).
  • Swap to safer segments only after cooldown.

Truth: complaints are a targeting problem wearing a deliverability hat.

---

Blocklist checks (weekly) and what actually matters

Blocklists are not dead. They are also not the only story. The SOP goal is to catch preventable issues fast.

Weekly blocklist SOP (copy-paste)

Every Monday, for every sending domain and sending IP (if applicable):

• Run a multi-RBL scan for sending domains
• Run a multi-RBL scan for sending IPs (if dedicated or known)
• Record any new listings since last Monday
• Tag each listing as “critical” or “noise”

Critical listings

• You see delivery failures tied to it.
• You see a listing on widely used lists.
• Your seed test shows inboxing collapse across providers.

Noise listings

• Random low-impact lists with no delivery impact.
• Old stale entries already delisted.

What to do if listed

1. Stop ramping. Freeze volume.
2. Identify cause:
   • Bad list source?
   • Too-fast ramp?
   • Missing unsubscribe?
3. If it’s a known actionable list, follow that list’s delisting process.
4. Rotate domains if needed. Don’t “argue with the internet.”

---

Seed and inbox placement sampling (the fast reality check)

Open rates are a liar. Inbox placement tests still matter. You just need to run them like an operator, not like a vendor demo.

Weekly seed SOP (simple and consistent)

Build a tiny seed list:

• 2-3 Gmail
• 2-3 Outlook/Hotmail
• 1 Yahoo/AOL (if relevant)
• 1 corporate inbox if you can (Google Workspace / M365)

Every Monday

• Send 1 plain-text message from each pool
• Same subject structure each week (don’t A/B test your monitoring)
• Check placement after 30-60 minutes:
  • Primary vs Promotions vs Spam
  • Outlook Inbox vs Junk
• Log results by provider

Thresholds

• GREEN: 80%+ inbox (Primary or Inbox) across your seed set
• YELLOW: 60-79%
• RED: < 60% or any provider goes mostly Spam/Junk

Actions

• YELLOW: throttle 25-50%, tighten segments, remove risky domains (see segmentation rules)
• RED: stop rules. Do not “send through it.”

---

Volume ramp rules (so you stop nuking reputation)

Volume is not a flex. It’s a risk multiplier.

Weekly ramp policy (copy-paste)

Per mailbox

• Only increase volume if last week stayed GREEN.
• Increase by 10-20% per week max on stable pools.
• If you changed copy, list source, or segmentation, treat it like a new campaign. Ramp again.

Per domain

• If the domain is new or recently cooled down:
  • Week 1: low volume, highest fit only
  • Week 2: expand slightly
  • Week 3+: scale only after seed placement holds

Hard rule

• Never ramp volume and broaden targeting in the same week.
• Pick one. You don’t get two.

For deeper sending pool structure (domains, mailboxes, throttles), Chronic already covered architecture in detail. This SOP assumes you have a sane structure and focuses on cadence. If you need the structure first, read: The 2026 outbound sending architecture.

---

Segmentation rules by persona and recipient domain (where agencies win)

Most “deliverability problems” are actually “you emailed the wrong people” problems.

Persona segmentation rules (copy-paste)

Tier your personas

• Tier A (money personas): direct buyers, high pain, high urgency
• Tier B (influencers): relevant but not direct owners
• Tier C (noise): broad, low-intent, “maybe someday”

Weekly segmentation SOP

• If any metric goes YELLOW, pause Tier C first.
• If any metric goes RED, pause Tier B and Tier C. Keep only Tier A after cooldown.
• If complaint rate spikes, tighten to only Tier A with strongest fit signals.

Want to stop guessing who’s Tier A? Stop doing spreadsheet astrology. Use scoring that blends fit + intent. Chronic’s approach is documented here: AI lead scoring.

Domain segmentation rules (Gmail, Outlook, Yahoo, and corporate)

Why: each provider reacts differently. Your SOP must isolate by domain group.

Weekly domain mix SOP

• Track performance by recipient domain: gmail.com, outlook.com/hotmail.com/live.com, yahoo.com, and corporate
• If one provider dips, throttle that provider first
• Keep corporate domains cleaner. They forward spam complaints to humans with jobs.

If Gmail dips

• Cut volume to Gmail by 50% for 48-72 hours.
• Send only to the best-fit segment.
• Ensure one-click unsubscribe is correct and functional.

If Outlook dips

• Check authentication first. Microsoft explicitly calls out SPF, DKIM, and DMARC for high-volume senders and moves toward enforcement. (Microsoft announcement)
• Cut volume and reduce follow-ups. Outlook can punish repetitive sequences fast.

If Yahoo dips

• Verify unsubscribe and complaint control. Yahoo runs Sender Hub guidance and enforcement patterns change over time. Keep your SOP strict even if volumes are smaller. (Yahoo Sender Hub FAQs)

---

Stop rules when reputation dips (non-negotiable)

Agencies lose clients because they keep sending during a reputation event. They “need to hit activity.” Cool. Enjoy the dead domain.

Stop rules (copy-paste)

Stop sending on a pool for new acquisition when any is true:

1. Complaint/spam rate hits 0.2%+ on any major provider view, or trends toward 0.3%.
2. Hard bounces exceed 2% over a send day, or spike >3% in one day.
3. Seed placement turns RED (mostly Spam/Junk on a major provider).
4. New critical blocklist listing appears plus delivery degradation.
5. Authentication breaks (SPF/DKIM/DMARC fail or alignment breaks).

What “stop” means

• Pause outbound sequences on that pool.
• Do not “just send to warm leads.” Use a different pool if you must.
• Investigate before resuming.

---

Copy-paste deliverability SOP template (weekly checklist)

Paste this into Notion, ClickUp, Asana, whatever runs your agency.

Weekly deliverability SOP (Monday)

Client:
Sending pool: (Domains + mailbox group)
Week of:

A) Identity health

• SPF passes and aligns with From domain
• DKIM passes and aligns with From domain
• DMARC present (reporting active)
• No DNS changes in last 7 days
• One-click unsubscribe headers present where required (RFC 8058)
• Unsubscribe processing time < 24 hours (goal: near-real time)

References:

• Google Email sender guidelines FAQ
• RFC 8058 one-click unsubscribe
• Microsoft bulk sender requirements announcement

B) Bounce and complaint guardrails (last 7 days)

• Hard bounce rate: ____% (Green <1%, Yellow 1-2%, Red >2%)
• Soft bounce/deferral rate: ____% (trend only)
• Unsubscribe rate: ____% (trend only)
• Provider complaint/spam rate (if available): ____%
• Any spike days? list dates: __________

Actions taken:

• None (Green)
• Throttle -25%
• Throttle -50%
• Pause Tier C personas
• Pause Gmail/Outlook segment
• Stop rules triggered (incident)

C) Blocklist checks

• Domain blocklist scan: Clean / Listed (which list?)
• IP blocklist scan: Clean / Listed (which list?)
• Any delivery errors tied to listing? Yes/No

D) Seed/inbox placement sampling

• Gmail: Inbox / Promotions / Spam (%)
• Outlook: Inbox / Junk (%)
• Yahoo: Inbox / Spam (%)
• Corporate: Inbox / Quarantine (%)

Status: Green / Yellow / Red
Notes: __________

E) Volume and ramp plan (next 7 days)

• Current daily per mailbox: ______
• Planned change: +10% / +20% / flat / -25% / -50% / pause
• Segments approved to send:
  • Tier A only
  • Tier A + Tier B
  • All tiers

F) Sign-off

Owner: ______
Approved: ______
Next review date: ______

---

Simple incident workflow: triage, isolate, cool down, rotate, re-warm

When you trigger stop rules, you don’t “optimize.” You run an incident.

Step 1: Triage (30 minutes, same day)

Goal: identify what changed.

• What changed in the last 7 days?
  • New list source?
  • New persona?
  • New copy angle?
  • New sending volume?
  • DNS/auth changes?
• Which provider is failing? Gmail only? Outlook only? All?
• Is it list quality (bounces) or audience rejection (complaints)?

Output: a single sentence root cause hypothesis. Not five.

Step 2: Isolate (same day)

Goal: stop the bleeding without killing the whole account.

• Pause the worst segment first:
  • risky domains
  • broad personas
  • new unproven list sources
• Split pools by provider if needed.
• Disable the worst-performing mailboxes.

If you run outbound end-to-end, isolation should be fast because your pipeline is structured. Chronic’s sales pipeline setup keeps segments and sources traceable. That matters during incidents.

Step 3: Cool down (48-72 hours)

Goal: reduce negative signals.

• Cut volume 50-80% or pause entirely on the affected provider.
• Send only to your cleanest segment if you must keep activity.
• Fix unsubscribe and list issues immediately.

Google’s guidelines explicitly connect mitigation eligibility to maintaining low spam rates over time. Cooldown is how you stop compounding the problem. (Google sender guidelines FAQ)

Step 4: Rotate (only when necessary)

Goal: protect the client’s long-term sending ability.

Rotate when:

• complaints stay elevated,
• inbox placement stays red after cooldown,
• or the domain reputation looks cooked.

Rotation is not “buy 20 new domains and blast.” Rotation is controlled replacement with warm-up.

Step 5: Re-warm (7-21 days)

Goal: re-establish stable inboxing.

• Start with Tier A only.
• Keep copy conservative. No aggressive promo language.
• Increase volume 10-20% per week if metrics stay green.
• Reintroduce segments one at a time.

If you need a fit-based lead intake to keep re-warm clean, stop pulling random lists. Use an ICP definition and enrichment waterfall. Chronic’s ICP builder plus lead enrichment is built for that exact “clean segment first” requirement.

---

What to do next when a metric fails (decision table)

---

How Chronic fits into this (one line, no fluff)

Deliverability Ops fails when list quality and segmentation are guesses. Chronic runs outbound with built-in structure: lead enrichment, AI email writer, and AI lead scoring so your Monday deliverability SOP has clean segments to cut and clean segments to keep.

Also, if you’re still duct-taping five tools and calling it a stack, price that pain honestly. Start here: GTM tool consolidation calculator.

For more context on why deliverability ops is now a core motion, not a side quest: Apollo’s deliverability suite is the signal.

---

FAQ

FAQ

What is a deliverability SOP?

A deliverability SOP is a repeatable checklist and incident process that keeps sender reputation stable. It covers authentication, bounces, complaints, blocklists, inbox placement sampling, volume ramps, segmentation rules, and stop rules. You run it on a schedule, usually weekly.

What thresholds should I use for spam complaints in 2026?

Use internal guardrails tighter than the provider limits. Treat 0.1% as a warning and 0.2% as a stop-and-fix event. Gmail references 0.3% user-reported spam rate as a key threshold in its sender guidelines and ties mitigation eligibility to staying below it. (Google sender guidelines FAQ)

Do I really need one-click unsubscribe for outbound sequences?

If your message is promotional or marketing-like, yes. One-click unsubscribe is defined by RFC 8058 via List-Unsubscribe and List-Unsubscribe-Post headers and requires DKIM coverage for those headers. (RFC 8058) Even when not strictly required, it reduces complaints, which is the metric that actually ruins you.

What do I do if only Gmail deliverability drops but Outlook looks fine?

Throttle Gmail first. Cut Gmail volume 50% for 48-72 hours. Send only to your cleanest Tier A segment. Run seed placement again. Don’t punish the whole pool if one provider is screaming.

How fast should we ramp volume after a cooldown?

Slow. Increase 10-20% per week, only if metrics stay green. If you changed list source, copy angle, or segmentation, treat it like a new campaign and ramp again.

Why do agencies lose deliverability even when authentication is perfect?

Because authentication is table stakes. The failure is usually audience mismatch: bad targeting, weak context, and too much volume. Complaints rise, providers react, and your “perfectly authenticated” emails land in spam anyway.

---

Run the Monday SOP, then hit send

Put this in your agency’s operating system:

1. Assign an owner per client pool.
2. Run the checklist every Monday.
3. Enforce stop rules without debate.
4. Log actions and outcomes for 4 weeks.
5. Promote what stays green. Kill what goes yellow twice.

Deliverability doesn’t need prayers. It needs process.