Module 08 · Diagnosis
Read your numbers. Name the broken row.
If you built the funnel in Module 01, you now have a model that says what each campaign's numbers should be. Diagnosis is just: compare actual to model, find the first row that deviates, fix one variable, re-run.
The one idea
Symptom → broken stage → first fix
Every cold email metric corresponds to exactly one stage of the pipeline. Bounce rate reads your list quality. Inbox placement reads your infrastructure. Reply rate reads your copy. Positive-reply rate reads your offer / segment fit. Knowing the mapping turns "this campaign isn't working" into "row 3 is broken, fix copy."
The diagnostic table
The single most useful table in this guide
| Symptom | Most likely broken stage | First fix |
|---|---|---|
| Bounce rate > 3% | List / verification | Module 02: tighten verifier cascade, re-check email patterns |
| Sends landing in spam at low volume | Infrastructure | Module 03: confirm SPF / DKIM / DMARC, check MXToolbox |
| Inbox at low volume, spam at higher | Warmup / per-inbox cap | Module 04: extend warmup, lower daily cap |
| Decent placement, reply rate < 1% | Targeting or copy | Module 02 trigger, then Module 05 line 1 |
| Replies fine, lots of "remove me" | Spintax / fingerprint | Module 06: spin subject + line 1 |
| Replies fine, positives low | Offer or segment fit | Wrong segment for the offer; revisit Module 02 |
| Positives fine, meetings low | Follow-up speed | Speed-to-lead on positives; refine the close in Module 05 |
Work top-down. A reply-rate problem is meaningless until bounce and placement above it are green. A broken lower row masks everything above it. Always fix from the bottom of the funnel upward.
The metric to ignore
Open rate is lying to you
2026 update Apple Mail Privacy Protection, iCloud Private Relay, Microsoft's image-prefetching, and now Gmail's Gemini-prefetched previews all auto-open emails before the recipient ever sees them. Open rates in 2026 are inflated by 30–60% across most lists. Treat open rate as directional only, never as a decision metric. Decide on reply rate.
The iteration loop
One variable at a time, or you learn nothing
- Run the model from Module 01: expected numbers per stage.
- Run the campaign at a statistically meaningful size — at least 500 sends before drawing conclusions.
- Compare actual vs expected. Find the first row that deviates.
- Change one variable tied to that stage. One.
- Re-run. Compare again. Repeat.
A clean diagnostic
Campaign expected 95% delivery, got 78%. → Bounce / infra row is broken first. Don't touch copy. Re-verify list, check SPF/DKIM/DMARC, fix that. Once delivery is back to 95%+, look at reply rate.
Do this now
Diagnose one real campaign
- Open your funnel sheet from Module 01.
- Plug in actual numbers from your most recent campaign (bounce, delivered, reply, positive, meeting).
- Calculate the actual conversion rate at each row.
- Find the first row where actual is more than 30% off from your expected.
- Use the diagnostic table above. Name the broken stage out loud.
- Pick ONE variable in that stage to change. Write it down before changing anything else.
- Run the next campaign with only that variable changed.
Don't do this
The diagnostic mistakes that keep you stuck
Change three things at once. You lose the ability to attribute the improvement and learn nothing. Discipline is the whole skill.
Kill a campaign before it has statistically meaningful volume. Under 500 sends is noise. Wait.
Optimize copy when the real break is deliverability. If sends are landing in spam, no amount of copy work matters.
Trust open rate. Make decisions on reply rate.
Skip building the funnel model. Without it, you're guessing every time.
You can now
You're ready to run cold email as a system
The whole skill, in one line
Build the funnel model. Build a clean list. Stand up infrastructure. Warm the inboxes. Write tight copy. Spin the fingerprints away. Send with the right controls. Read the numbers. Fix one row at a time. That's the entire job.
One last module — Module 09 covers what specifically changed in cold email between 2023 and 2026 so you know what's new versus what's classic.