Campaign Update
Live Campaigns
| Campaign | Leads | Sent | Replies | Status |
|---|---|---|---|---|
| Founder (911) | 7,419 | 644 | 13 | Ramping |
| BrandHijack Takedown (905) | 168 | 148 | 4 | Active |
| Marketing (912) | 5,113 | 2 | — | Launched |
| Product (913) | 3,758 | — | — | Launched |
| Revenue (914) | 1,737 | — | — | Launched |
What Shipped
Wave 2 Segments
Four buyer-role segments sharing one structure (Founder baseline) realigned to each audience's worldview. Same 4-variant split test + passoff, framed per segment.
| Variant | Mechanism |
|---|---|
| V1 | Warm retargeting — capture everyone engaging you + competitors |
| V2 | Competitor ABM — pull competitor followers into warm audience |
| V3 | Competitor funnel teardown — 28-day audit, find the leak |
| V4 | ICP / market map — look-alike accounts loaded to LinkedIn Ads |
Proof by segment:
- Marketing: Toggl $93K→$44K cost-per-closed-won
- Revenue: Toggl — every qualified lead in 2025 closed
- Product: Rosie signup→trial 30%→50%, cost/trial −40%
BrandHijack Takedown
602 competitor screenshots collected. Targeting companies bidding on competitor brand terms in paid search.
1 interested reply from 148 sends — 2.7% reply rate.
TLA Lookalike Pipeline Deep Dive
Discover companies similar to HD's best clients, qualify for ICP fit, generate personalized copy, and stage campaigns — all automated.
7.3% ICP qualification rate • 72 A-tier, 87 B-tier • 46 valid + 9 catch-all + 3 unknown emails
Seed Companies
Why 2,024 Companies Were Rejected
The ICP classifier (gpt-4o-mini) evaluated all 2,183 net-new domains. 159 passed. Here's why the rest didn't:
| Rejection Reason | Count | % |
|---|---|---|
| Too small (<15 employees, but is B2B SaaS) | 1,057 | 52.2% |
| Too small + not B2B SaaS | 674 | 33.3% |
| Not B2B SaaS (right size, wrong type — agencies, B2C, marketplaces) | 178 | 8.8% |
| Other (too large >220, wrong geo, B2B service not SaaS, niche vertical) | 115 | 5.7% |
ICP criteria: B2B SaaS, 15–220 employees, target geos (US, CA, GB, DE, FR, NL, SE, DK, NO). The biggest filter is company size — DiscoLike surfaces many early-stage startups (<15 people) that are too small to run paid media. The "too large" bucket (>220) catches enterprise companies outside HD's sweet spot.
Remaining Work
- 104 domains without contacts — AI Ark + Blitz returned 20% hit rate. ~15 are real standalone B2B SaaS worth enriching via LinkedIn founder search. Rest are regional aliases of large companies (wthubspot.de, zendesk.com.br) or thin shells.
- 2,024 rejected companies — net-new to Supabase, have employee data. Can bulk-insert as icp_qualified=false for future dedup + cross-client prospecting.
- Bison API key — once set, 58 campaigns push in one script run.
PostHog Ideas Campaign Deep Dive
512 companies identified as PostHog alternatives/adjacents. 358 have AI-generated ideas + unique E1/E2 copy. Remaining 154 pending contact pull before copy generation.
How the Ideas Prompt Works
A 6-idea menu (gpt-4.1-mini, temp 0.5) picks the 3 best-fit angles per company. Slot 1 is always the PostHog case study opener. Slots 2–3 rotate between competitor targeting, thought leader ads, pain-first prospecting, and retargeting with proof.
Example 1: Exa.ai (AI search API)
Competitors: Tavily, Firecrawl, Perplexity • Contact: Will, CEO
Ideas generated: (1) PostHog case study credibility opener, (2) Monitor Tavily signals — capitalize when they slip, (3) Thought leader ad on AI search accuracy debates
Example 2: Blumira (SMB cybersecurity)
Competitors: CrowdStrike, Huntress, Splunk • Contact: Jon-Eric, Marketing Director
Ideas generated: (1) PostHog case study credibility opener, (2) Monitor CrowdStrike signals, (3) Founder thought leader ad: why generic security solutions overwhelm lean IT teams
Idea Types Generated
- E: PostHog case study credibility opener (always)
- A: Competitor follower targeting on LinkedIn
- B: Founder thought leader ad (spiky topic)
- C: Case study as promoted post
- D: Pain-first prospecting ad (3 variants)
- F: Retargeting with specific proof point
Pipeline Numbers
Sender: Ajay Mishra
Prereqs remaining: competitor enrichment + contact pull + copy review.
Copy Updates
Non-Founder Copy Refresh (912 / 913 / 914)
Rewriting all non-founder segments to match Founder (911) style: direct-line hard intro offer instead of the softer multi-offer approach. One sharp ask, not three options.
Key changes per segment:
- 912 Marketing: Lead with their paid funnel, lead quality, demand. Drop founder framing.
- 913 Product: Target signup→activation, PLG funnel, trial conversion, competitor users.
- 914 Revenue: Pipeline that closes, in-market deals, competitor deals slipping.
Copy review flagged 6 defect lines across Marketing bucket (2 free-CTAs, 1 price mention, 2 fabricated stats, 1 PMAX-dependent). Founder bucket is clean — the gold standard.
Winning angle on 911: "unqualified leads → lookalike ICP mapping" with Checkly +316% proof. All segments keep this angle, sharpened to direct-line.
Competitor Name-Drop
{COMPETITOR_1} personalization variable in templates.
Coverage: 23,588 / 25,560 (92.3%) — up from 8.8% last week. Clean competitor names (e.g. ChartMogul→ProfitWell). The 7.7% gap gets auto-filled with "your top competitor" at lead level.
Also populated: {COMPETITOR_2} and {COMPETITOR_3} for future use.
Lead Magnet Rotation
Three magnets across all segments:
- Competitor Funnel Teardown — highest-intent offer (audit their competitor's full paid funnel over 28 days, find the leak)
- ABM Account Map — top 50 accounts with buying signals, loaded to LinkedIn Ads
- Thought Leadership Sprint — done-for-you founder LinkedIn content sprint
Founder (911) early signal: 13 replies / 644 sends = 2%. Need more volume on 912–914 before calling winners.
Copy Standards
- 6 subject variants per step — lowercase, 2–4 words, internal-email feel
- Hybrid spintax: word-level swaps + phrase-level for clauses/CTAs
- Target thousands+ combinations per email
- No em dashes. Never use "free" — replaced with {on the house|on us|fully complimentary}
- Spam-checked via EmailGuard before every push