§0 Hypothesis
One frame, four lengths. Find length winner first.
Frame: "We can help extend your runway by claiming hyperscaler programs you don't know about." Direct runway benefit. Programs you didn't know existed. No engineering time required.
Variable: body length only. Subject + frame + CTA held constant.
Persona: CEO / Founder only. (Engineer test deferred to wave 2 to avoid persona × length noise.)
Cohort: funded Series A or B SaaS · 20-50 emp · EU/UK · last 18mo round.
CTA: "Worth comparing notes?" — no calendar link anywhere.
Goal: identify length sweet spot for this audience. Pick top 2, run round 2 confirmatory.
§1 List Spec
A Variant A · Ultra-Short ~20 WORDS
extend runwayHey {{first_name}},
Most Series {{round}} founders qualify for $50-100k+ in hyperscaler programs they've never heard of. Could extend runway a few months.
Worth a look?
MAP, RAMP, Founders Hub, ISV Accelerate. Most founders qualify for two or three of these. Worth the map?
B Variant B · Short ~45 WORDS
extend runwayHey {{first_name}},
Saw {{company_name}}'s Series {{round}}. Most founders your stage qualify for two or three hyperscaler programs they've never heard of. MAP, RAMP, Founders Hub.
$50-100k+ in credits. Real runway extension. Paid by AWS and Google on co-sell, not by you.
Worth comparing notes?
Eligibility rules across these programs contradict each other. We mapped the overlap so you don't have to chase three rep teams. Want the one-pager?
C Variant C · Medium ~85 WORDS
extend runwayHey {{first_name}},
Saw {{company_name}}'s Series {{round}}.
Most Series A/B founders qualify for two or three hyperscaler funding programs they've never heard of. MAP, RAMP, Founders Hub, ISV Accelerate. The AWS rep usually doesn't surface them because they're comped on consumption, not credits.
Adds up to $50-100k+ in credits. Real runway extension at your stage.
Certified AWS and Google partner. Paid by them on co-sell, not by you. No engineering time on your side.
Worth comparing notes on what {{company_name}} qualifies for?
Eligibility rules for these programs contradict each other depending on which page you land on. We've watched founders miss programs they qualified for because the AWS side didn't mention it. Same the other way with Google. We mapped the overlap. Want the one-pager?
D Variant D · Long ~170 WORDS
extend runwayHey {{first_name}},
Saw {{company_name}} closed Series {{round}} in {{quarter}}. Quick context on why I'm reaching out.
Most Series A/B SaaS founders qualify for two or three hyperscaler funding programs they've never heard of. MAP, RAMP, Founders Hub, ISV Accelerate. The AWS account rep usually doesn't surface them because they're comped on consumption, not credits. The Google side is structured the same way.
Combined, they add up to $50-100k+ in credits. MAP alone usually covers a meaningful slice of first-year cloud spend, paid back as credits. The rest layer on depending on cloud commit, product stage, and where you're running workloads.
Real runway extension at your stage. No engineering time on your side. The claim work is on us.
We're a certified AWS and Google partner. Paid by them on the co-sell side, so it doesn't cost you anything.
Worth comparing notes on what {{company_name}} qualifies for?
One more thing on the programs side. Most founders we talk to find out about AWS MAP only after they've finished migrating. By then the funding window has closed and there's nothing to claim. If migration is still ahead of you, the timing right after a fresh round is when the map is most useful. Worth a look?
§2 Volume Math · Why 350 per Variant
| Variable | Value | Source |
|---|---|---|
| Baseline positive-reply rate | ~1.0% | B5 funnel data + Unify 25M study |
| Effect size to detect | 50% lift between variants | Required for ship-vs-kill decision |
| Statistical power target | 80% (directional) | Standard for early-stage A/B |
| Minimum sends per variant | ~350 | Two-proportion z-test at α=0.10 (one-sided) |
| Total sends | 1,400 | 4 variants × 350 |
| Expected positive replies (1% baseline) | ~14 | Across all variants |
| Expected meetings booked | ~4 | 27.8% reply-to-meeting (B5) |
| Expected qualified ($20K+/mo) | ~3 | 73% qualification rate (B5) |
Trade-off honest read
350/variant gives directional read, not statistical certainty. If two variants come in at 1.1% and 0.9% positive-reply rate, that's NOT a real difference. Need 600+/variant for confident kill. Treat round 1 as a pre-screen: keep top 2 lengths, run round 2 at 500/variant for confirm.
§3 Decision Rule
| Day | Action |
|---|---|
| Day 0-5 | Send 1,400 across 4 variants (350 each), ~280/day |
| Day 3-4 | Auto-send threaded bump on no-reply |
| Day 7 | First metrics check: open rate sanity, hard bounces, spam complaints |
| Day 14 | Final read: positive-reply rate by variant + meetings booked by variant |
| Day 14 | Pick top 2 variants by positive-reply rate (ignore opens, ignore total replies) |
| Day 15 | Round 2: top 2 variants × 500 sends each on next list slice |
| Day 30 | Ship winner. Promote to default Wave 2 template |
Kill criteria (any variant)
• <1 positive reply per 350 sends in 14 days → kill that variant
• Open rate <30% → deliverability problem, not length problem · investigate sender pool
• Spam complaint rate >0.1% → kill list segment, not variant
• Negative-reply rate >15% → frame is wrong, kill ALL variants and rethink offer
What we are NOT testing in round 1
• Persona (CEO only — CTO/VPE deferred to wave 2)
• Subject line (extend runway held constant — connector + emoji subjects deferred)
• Hook type (signal-driven held constant — pain-driven and stat-driven deferred)
• Send time (all sends in same morning UTC window)
One axis at a time. Length first.