The iteration loop

Cold email is a game of iteration. The first version is a draft. The 10th version is a product. The operators who win are the ones who run the feedback loop weekly and compound learnings over months.

The weekly loop

1. Monday, measure. Pull last week's numbers: sends, opens, replies, positive replies, meetings.
2. Monday, read. Read every positive reply and objection. Extract patterns.
3. Monday, decide. What's the one thing you'll change this week? (One variable.)
4. Tuesday, ship. Launch the new variant.
5. Friday, checkpoint. Early signal: is the new variant materially different, or same as control?
6. Next Monday, call it. Winner, loser, inconclusive. Roll forward.

15-30 minutes per week. Compound effect over 12 weeks is dramatic.

What to change each cycle

Rotate through your Tier 1 variables (see what to test):

• Week 1: subject line
• Week 2: first line approach
• Week 3: offer angle
• Week 4: CTA structure
• Week 5: subject again (new challenger)
• ...

Each cycle, keep the winners. Replace the losers. Over 12 weeks your control email is 12 improvements ahead of where you started.

The monthly retro

Once a month, step back:

• What's trending? (reply rate, positive rate, pipeline generated)
• What's breaking? (deliverability drift, cold mailbox aging, list fatigue)
• What's the biggest leverage move for next month?

Sometimes the answer isn't copy. It's list quality, deliverability, or a new segment.

Quarterly reset

Every 90 days:

• Full review of current controls vs start-of-quarter controls
• Kill campaigns that underperform
• Invest more in campaigns that work
• Explore a new ICP or angle if existing are saturating
• Review infrastructure: new sending domains, retired mailboxes, warming cadence

The log

Keep a running doc of every experiment. Format:

Date: [YYYY-MM-DD]
Campaign: [name]
Hypothesis: [what I thought would happen and why]
Variable changed: [one specific thing]
Control version: [exact text or ID]
Challenger version: [exact text or ID]
Sample size per variant: [number]
Duration: [days]
Result:
- Control: [reply rate, positive rate]
- Challenger: [reply rate, positive rate]
- Winner: [control / challenger / inconclusive]
- Effect size: [percentage lift]
What I rolled forward: [did challenger become new control?]
What I learned: [one-sentence takeaway]

After 6 months you have an experimental archive specific to your audience. It's worth more than any playbook, it's the ground truth of what actually works for your segment.

The honest truth about iteration

Most experiments are inconclusive. Most changes don't materially move the needle. You find the 10% that do and scale them. The discipline is in running the process consistently, not in making every experiment a winner.

Operators who iterate for 90 days outperform operators who launch-and-forget by 3-5x on pipeline per email sent. The compounding is real.

What to do with this

• Schedule the weekly iteration loop as a recurring calendar block, without a cadence iteration silently stops after launch
• Aim for 1-2 small tests per week, not one giant overhaul per quarter, the compound of small changes beats heroic rewrites
• Accept that most experiments produce no signal, that's normal, the 10% that move the needle justify the full cadence
• Feed negative-reply objections directly into copy rewrites, prospects' own language is your highest-ROI source
• Commit to 90 days before judging the process, teams that quit iteration in week 3 miss the 3-5x compound that shows up in weeks 8-12