Loading tutorials…
Loading tutorials…
Open rate dropped from 28% to 14%. Bounces jumped. Customer support is forwarding 'we never got the email' tickets. The instinct is 'subject lines' or 'content' — usually it's deliverability infrastructure. Here's how specialists diagnose SendGrid deliverability without guessing.
Who this is forSendGrid accounts whose deliverability has dropped noticeably (open rate -5+ points, complaint rate >0.1%, bounce rate >2%). If you're not sure whether the problem is auth, IP reputation, content, or list hygiene, this diagnostic sequence isolates the real cause.
What you'll need
Step 1
Apple Mail Privacy Protection inflates iOS open rates 30-50%. Apparent drops in 2026 may be MPP unwinding, not real engagement loss.
SendGrid → Stats → Email Activity → look at recent campaigns. Compare Opens to Clicks.
Healthy ratio: 8-15% of opens become clicks. If opens dropped but clicks DIDN'T, it's MPP unwinding (Apple Mail is gradually unloading pixel pre-fetching).
If opens AND clicks both dropped, this is real engagement loss. Continue diagnosis.
Go-forward: use CLICK rate as your primary engagement KPI. Clicks are unaffected by MPP and reflect real interest.
Build a baseline: average click rate over last 12 weeks (excluding the past 2 weeks). Compare current click rate to that baseline — that's your real deliverability signal.
Step 2
SendGrid → Settings → Sender Authentication. DKIM, SPF, DMARC must all show valid. Then check real Gmail headers for alignment.
SendGrid → Settings → Sender Authentication → Domain Authentication. Click your domain.
Verify status: all CNAMEs should show green. If any are red, DNS was changed (likely by IT or web devs unaware of email auth).
Send a test email to a Gmail address you control. Show original → Authentication-Results header.
Look for: `spf=pass`, `dkim=pass`, `dmarc=pass`. Also verify `header.i=` (DKIM signing domain) matches your authenticated domain (em.yourbrand.com), not sendgrid.net.
mxtoolbox.com checks: SPF Record Lookup (one record, no PermError), DMARC Lookup (record present, valid syntax).
If any auth is broken: fix it (Tutorial 2). Open rate typically recovers 60-90% within 14 days of auth restoration.
Step 3
Pull 90 days of sends. Find the date/campaign where deliverability inflected. What was different?
SendGrid → Stats → Overview → 90-day view. Look at delivery rate, open rate, bounce rate, complaint rate.
Find the inflection: which week did metrics shift? Was there a single campaign that started the slide?
What was different that week? Common causes:
- Volume increase (sent 10x normal — see Tutorial 6 warmup)
- List expansion (added a new segment without warming up to it)
- Content change (new template, image-heavy design, spammy subject)
- Sender change (switched From address, used a new sending subdomain)
- DNS change (someone touched DNS)
- Provider event (Yahoo/Gmail policy update — search Twitter for `gmail email policy [month]`)
Once identified, the fix is to revert or remediate that change. Test a follow-up campaign with the OLD pattern to confirm.
Step 4
Postmaster Tools shows ground-truth reputation Gmail SendGrid stats cannot. If Domain Reputation is "Low" or "Bad," that's your cause.
postmaster.google.com → add and verify your sending domain (if not already).
Check: Domain Reputation (target High, recoverable from Medium). "Low" or "Bad" means Gmail is actively filtering you.
Check: IP Reputation (matters more on dedicated IP). On shared IP, SendGrid pool reputation is what shows.
Check: Spam Rate. Target <0.1%. Above 0.3% triggers heavy filtering. Above 0.5% is account-threatening territory.
Microsoft SNDS (sendersupport.olc.protection.outlook.com/snds/): same analysis for Outlook/Hotmail.
If Gmail reputation is Low: the cause is upstream (auth, list quality, complaint rate). Fix the upstream issue; reputation recovers over 30-60 days.
Step 5
SendGrid → Suppressions. Spike in bounces or complaints in the past 30 days points to list quality issues.
SendGrid → Suppressions → Bounces. Sort by date. Spike in last 30 days?
Bounce rate spike: usually means you imported a stale or purchased list, OR you scaled to inactive contacts without warming up.
SendGrid → Suppressions → Spam Reports. Complaint rate spike? Usually means content + audience mismatch (you sent promotional content to a segment that expects transactional, or you re-engaged a 180-day inactive list).
Healthy thresholds: Bounces <2% per campaign. Complaints <0.05% (under 1 in 2000). Above these, the campaign is doing reputation damage.
Action: PAUSE any send that's historically pushed >2% bounce or >0.1% complaint. Diagnose the segment before resuming.
Step 6
Spammy subjects, image-heavy emails, broken HTML, missing footer all trigger filters. Pull recent campaigns and audit.
Pull last 30 campaigns. Look at subject lines.
Trigger phrases to avoid: "FREE!!!" "Buy now," "Make money," "Risk-free," ALL CAPS, excessive exclamation, emojis-as-decoration.
Email body: text-to-image ratio. Image-only emails (no real text) score high for spam. Aim 60% text minimum.
Links: 3-5 per email is healthy. 20+ links looks like a spam blast and triggers content filters.
Footer: physical address present (CAN-SPAM)? Unsubscribe link functional? Footer missing or fake address triggers Trust + Safety review.
Run a content scan: mail-tester.com gives a SpamAssassin-style score. Free tool, ~10 minutes per template. Aim 8/10+.
Step 7
Pick the most-likely cause from Steps 2-6. Fix ONE thing. Send to engaged segment. Measure for 14 days before next change.
Pick the most-likely root cause based on diagnostic findings.
Fix that ONE thing. Don't fix multiple causes at once — you won't know which one worked.
Send a test campaign to your most-engaged 25% (Engaged 30d segment in Marketing Campaigns).
Wait 14 days. Compare delivery + open + click rates to baseline AND to the campaigns that started the slide.
If metrics recovered to baseline on the engaged segment, expand: send to Engaged 60d, then 90d, then full active list over the next 4-6 weeks.
If metrics didn't recover, the root cause wasn't what you tested. Move to the next most-likely cause from your diagnosis.
Document what worked. Next deliverability issue, the diagnosis will be faster.
Common mistakes
Blaming content when authentication is broken
What goes wrong: Spending 4 weeks A/B testing subject lines while the real cause is broken DKIM (someone touched DNS). Metrics stay low. 4 weeks of test-and-iterate wasted; the actual problem was a 10-minute DNS fix.
How to avoid: Always diagnose in this order: Apple MPP → authentication → list quality → Postmaster Tools → content. Auth is upstream of content; check auth first.
Not setting up Gmail Postmaster Tools + Microsoft SNDS
What goes wrong: SendGrid Stats shows symptoms (open rate dropped) but not cause. Postmaster Tools shows the cause (Domain Reputation: Low). Without it, you're guessing for weeks.
How to avoid: Set up postmaster.google.com + Microsoft SNDS today. 30 minutes one-time. Check weekly during normal operations, daily during incidents.
Continuing to send full volume during diagnosis
What goes wrong: Each campaign during a deliverability incident compounds the damage. 2 weeks of diagnosis at full volume = another 14 campaigns shipped to a degrading reputation. Recovery extends from 30 days to 60-90 days.
How to avoid: When deliverability drops 5+ points, PAUSE bulk sends. Continue ONLY engaged-segment sends (Engaged 30d) at half normal volume. Resume full volume after the cause is identified + fixed.
Sending to 180-day inactives "to clean them up"
What goes wrong: Inactives don't open. They bounce, complain, or ignore. Bulk sending to them spikes complaint + bounce rate. Reputation drops another 5-10 points across the whole list. Worsens the problem you're diagnosing.
How to avoid: Suppress 180-day inactives during diagnosis + recovery. Use a re-engagement automation on a SEPARATE IP/subdomain if you want to recover the list — never on the impaired sender.
Changing many variables at once
What goes wrong: Subject lines, send time, content, segment all changed in one campaign. Metrics either recover or don't — but you can't tell which change caused it. Can't repeat the fix. Can't measure attribution.
How to avoid: Change ONE variable per 14-day test window. Boring discipline beats panicked experimentation. The recovery is faster overall because each cycle gives a clear answer.
Ignoring the SendGrid Stats Deliverability tab
What goes wrong: SendGrid surfaces blocked sends, deferred sends, and reasons in Stats → Deliverability. Most teams skip it. Useful patterns hide there: 'X% of sends to gmail.com were deferred with reason rate-limited' tells you you're hitting Gmail's per-domain rate ceiling.
How to avoid: Check SendGrid → Stats → Deliverability tab weekly. Filter by mailbox provider (Gmail, Outlook, Yahoo). Investigate any provider showing >5% block/defer rate.
Recap
Done — what's next
How to set up SendGrid domain authentication — DKIM CNAMEs, SPF, DMARC, and link branding
Read the next tutorial
Hand it off
Diagnosing SendGrid deliverability the right way takes experience. A specialist who has recovered 50+ accounts runs the right diagnostic in 60-90 minutes and knows which causes to test in which order. Typical engagement is $300-600 for diagnosis + $200-500/mo ongoing monitoring at $14-16/hr.
See specialist rates
14-60 days. Authentication fixes (DKIM/SPF/DMARC): 14-30 days. List hygiene improvements: 30-45 days. Reputation rebuilds from severe damage (purchased list, sustained complaint rate): 60-90 days. Don't expect immediate recovery; sender reputation is a trailing indicator.
Most likely: (1) someone touched DNS (web dev migration, new CMS), (2) list expanded to a new less-engaged segment, (3) ISP policy update (Yahoo/Gmail bulk sender rules in 2024 hit many accounts), (4) you crossed a volume threshold without warming up. Check the inflection-point step — the date of the drop usually points to the cause.
Yes — SendGrid's shared IP pool reputation is genuinely good. Bigger risk is YOUR domain reputation (separate from IP reputation). Your domain + DKIM signing is what mailbox providers track most heavily in 2026. Focus on domain reputation; IP matters less than people think on shared.
Last resort. New subdomain = new reputation = ~30-60 days warmup. Most cases recover with auth + list-hygiene fixes on the existing subdomain. Switch only if reputation is permanently damaged (3+ months of <10% open rates with sustained complaint rate >0.5%).
Check status.sendgrid.com for outages. Then check Postmaster Tools — if your Domain Reputation is fine but open rate dropped uniformly, it might be SendGrid infrastructure. If Domain Reputation is also low, the issue is yours (auth, list, content). 95% of the time it's yours, not SendGrid's.
Only if metrics breach SendGrid's Trust + Safety thresholds: complaint rate >0.5% sustained, bounce rate >5% sustained, spam-trap hits. Below those, SendGrid sends warning emails but doesn't suspend. Heed the warnings — multiple ignored warnings escalate to suspension review.
SendGrid
Gmail and Yahoo's 2024 bulk sender rules made domain authentication non-negotiable above 5K sends/day — and strongly recommended below it. SendGrid hides the link-branding step that most accounts skip, leaving every click flagged 'via sendgrid.net' in Gmail's clip warning. Here's the full auth stack.
SendGrid
Suppressions are the dull, unglamorous part of email that decides whether your account survives the year. Gmail and Yahoo's 2024 bulk sender rules made one-click unsubscribe mandatory at scale, and SendGrid's suppression groups are how you respect that without losing transactional sends. Here's the full setup.
SendGrid
A dedicated IP is the most-recommended SendGrid feature that's wrong for 80% of accounts. Below 100K sends/month, you'll hurt deliverability — the IP can't build reputation on low volume. Above 250K/month sustained, it pays off. Here's the decision framework and full warmup playbook.
Mailchimp
Open rate dropped from 28% to 16% in 30 days. The instinct is 'subject lines.' Usually it's not. Deliverability, list hygiene, and audience drift account for 70% of open-rate drops. Here's how specialists isolate the cause.
SendGrid
DIY SendGrid is the right call until it isn't. The signal isn't 'sending more emails' — it's that the cost-of-mistakes finally outweighs the cost-of-hiring. Here's the honest framework for when that line is crossed.