A practical 90-day rollout

It’s taken us 90 days, and sometimes six months, to find message-market fit. The first 90 days on a new account should build a trustworthy learning loop and find enough signal to scale responsibly.

We use this time to figure out what we can scale. Like I said earlier, we test thousands of copy variations to see which CTA, offer, intro, audience, and everything else work together.

Do we test feature-focused explanations or outcome-focused? Lead magnet resource or straight to meeting?

The only reliable way to find that is hard data from hundreds of thousands of sends.

Days 0 to 14: make the data trustworthy

• Stabilize sending infrastructure and record the baseline.
• Model TAM, customer value, channel budget, and the volume the account can support.
• Pick one primary event for this phase: signup, meeting, customer, or another funnel action.
• Connect account, email, click, website, conversion, and CRM.
• Build the offer lab before writing extra copy.
• Decide what evidence is required before a test can win or lose.

Days 15 to 45: expand the accordion

• Test several meaningful offers against a control.
• Run creative batches around the best offer candidates.
• Keep each test to one changed variable.
• Compare segments only after the offer and creative are traceable.
• Log losers with the same care as winners.
• Review downstream events weekly, not just replies.

Days 46 to 75: compress around signal

• Move more volume toward combinations that repeat across cycles.
• Use the 70 / 20 / 10 allocation once a real winner appears.
• Retest the winner in another qualified segment.
• Compare provider, industry, company size, and title one at a time.
• Watch customer quality so cheap signups don't hide weak revenue.

Days 76 to 90: prepare full-market coverage

• Set the 30-to-60-day contact cycle.
• Calculate sendable TAM and annual placements.
• Confirm that infrastructure capacity and account economics support full coverage.
• Build a live exclusion process for opt-outs, customers, bad addresses, and DPA rules.
• Send click audiences to paid media only where consent and platform rules allow it.
• Forecast volume from the observed conversion rate, not a generic benchmark.
• Choose what the next cycle will keep, improve, and test from scratch.