From CSV to Live Campaign
Templates, Rules, Deployment

T1 Template (Personalized - Separate Instantly Campaign)

T1 leads get a personalized subject line (subject_hook) AND a personalized opening line (personalization_line). Both come from the CSV. The subject opens a loop, the first line pays it off. These run in a separate Instantly campaign from T2/T3.

T2/T3 Template (Generic - Separate Instantly Campaign)

T2/T3 leads have no personalization data. They run in separate Instantly campaigns with generic subject lines and a company-name-only opening. Split test subject line variants across these campaigns.

Subject Line Variants (split test on T2/T3)

T1 uses {{subject_hook}} from the CSV. T2/T3 uses one of these generic variants. Run A/B tests across T2/T3 campaigns, 200+ sends per variant minimum. Kill losers after week 1.

Instantly variables: The {{RANDOM | ... }} syntax is Instantly's built-in spintax. Each recipient gets a random combination of sign-off, name, and title. This avoids email fingerprinting and improves deliverability.

Testing protocol: Split traffic evenly across variants. Minimum 200 sends per variant before drawing conclusions. Track open rate (subject line test) and reply rate (body copy test) separately. After week 1, kill any variant below 50% of the leader's reply rate and reallocate volume to the top 2.

Qualification Criteria (must hit at least 2)

Real Example: Adrian Johnston / Elyos AI (ICP 9.3/10)

The highest-scoring lead in the entire batch. Here's how Sherlock's dossier becomes a direct outreach email.

What Sherlock Found

Why T0, Not T1

Adrian hits 4 of 5 qualification criteria. Category definition gap (investor said it explicitly). Funding within 90 days ($13M, 6 weeks ago). Specific strategic gap (international expansion with no global brand). ICP 9.3 with massive media gap. A podcast invite would work. But telling him we already know his problem works better.

What Adrian Receives (Jamie Writes & Sends Manually)

This is NOT an Instantly template. Jamie drafts this email using the dossier page, then sends it directly from his inbox. No automation, no spintax, no variables.

What makes this T0: The email delivers an actual strategic insight Adrian can act on even if he never replies. It doesn't just report what Sherlock found ("you have no podcasts"). It spots a positioning gap through a growth operator lens: his marketing leads with the technology, his customers talk about the outcome, and that mismatch is what will cost him the category race. The CRM/relational database analogy makes the point concrete. The Loom offer feels natural because the email already proved Jamie understands the business.

Loom Overview (3-4 minutes)

Jamie screen-shares their website and does a quick positioning audit. Not scripted. Just pull up their site, point out what you see, and offer value. The founder should learn something even if they never reply.

T0 Follow-Up Sequence (All Manual — No Automation)

Jamie sends each of these follow-ups manually from his inbox. There is no Instantly sequence, no automated timing, no campaign. Jamie checks his sent folder, and when the right number of days have passed without a reply, he sends the next one. Each email is customized using the lead's dossier page.

Other T0 Pain Signal Patterns (from 95 dossiers)

Each pattern requires Jamie to deliver a specific insight the founder hasn't considered. The email angle is not "I noticed X." It is "X is causing Y, and here is what I would do about it." All of these are written manually by Jamie using the lead's individual dossier page — never templated into Instantly.

T0 Airtable Integration

Every T0 lead is pushed to the Jamie Client Database in Airtable with specific field values that mark them for manual-only handling. They are excluded from all Instantly and HeyReach campaigns.

Real Example: Dossier to Sent Email

Here's exactly how a Sherlock dossier becomes a personalized outreach email. This is Alan Zhao, Co-Founder at Warmly, one of 95 deep-dive dossiers from the pipeline.

Step 1: What Sherlock Found

Step 2: What Goes Into the CSV

Step 3: What Alan Receives in His Inbox

What makes T1 different: The subject line opens a loop ("engineer who built a media empire" - what about it?), the first line pays it off and flows directly into the ask. Alan reads something that proves Jamie knows who he is before the pitch arrives. This hook was generated by Claude Haiku from Alan's Sherlock dossier.

When Rule-Based (Not Claude) Provides the Hook

383 T1 leads have hooks generated by a rule-based engine instead of Claude Haiku. The engine scans the existing personalization_line for keywords (funding, pivot, YC, AI, podcast, growth, industry) and generates a matching subject_hook. These are formulaic but functional.

The rule-based engine detected "fund" in the personalization_line and generated a funding-themed subject_hook. Less creative than Claude-generated hooks but still personal enough that Laurent knows you looked him up.

T1 LinkedIn (HeyReach): Runs in Parallel

T1 leads with no email (295 leads) get LinkedIn outreach only. T1 leads WITH email get both channels (LinkedIn connection request + email, staggered by 2-3 days).

Personalized Subject Lines (subject_hook Examples)

The subject_hook is a short, lowercase phrase that opens a curiosity loop. The personalization_line in the email body pays it off. They're always written as a pair. These are actual Claude Haiku-generated hooks from the 69 dossier-matched leads in the final CSVs.

No individual research, no subject_hook, no personalization_line. Uses {{companyName}} as the only personal element. Leads are split into separate Instantly campaigns by segment for tracking reply rates.

T2 vs T1 difference: No subject_hook, no personalization_line. Uses {{companyName}} as the only personal element. Generic subject line variants (A/B tested). T2 LinkedIn-only leads get the same standard HeyReach sequence (no personalization in the follow-up).

Lowest priority. Same template, generic industry phrase. These exist because even low-fit leads occasionally convert when the timing is right.

T3 sends last. Same template as T2. After T1 and T2 campaigns are running and reply rates are confirmed healthy. If domain reputation dips below 95% in Instantly, pause T3 first.

HeyReach runs a 5-step sequence: blank connection request (no note) -> wait 5 days (not accepted = end; LinkedIn auto-withdraws after ~25 days) -> if accepted, 4 follow-up messages with 1-day gaps between each (exits on reply at any step). No profile view step. Blank requests get 10-15% higher acceptance rates. 20 connections/day, 30 messages/day.

Every lead gets the same follow-up cadence. If they reply to any email, the sequence stops automatically. Calendly link ONLY appears in Email 3 (never Email 1).

Hard Rules (Non-Negotiable)

Campaign Structure

T1 and T2/T3 are completely separate Instantly campaigns. LinkedIn runs in parallel via HeyReach with its own pacing.

540/day hits the hard cap exactly (18 accounts x 30/day). If a sending account goes down, the campaign it belongs to slows proportionally. Do not reassign its volume to other accounts. T1 campaigns (820 leads total) get priority - they finish in ~10 days. T3 (907 leads) starts Week 2 after T1 reply rates are confirmed healthy.

LinkedIn Campaign Structure (HeyReach)

4 HeyReach campaigns running in parallel with email. Connections capped at 20/day, messages at 30/day.

Sequence: Send Connection Request > wait 5 days (if not accepted, end - LinkedIn auto-withdraws pending requests after ~25 days) > if accepted, wait 1 day > Send Message > wait 1 day > if no reply, Send Message > repeat 2 more times > End. Each step exits on reply. All 4 campaigns share the daily limits.

Email + LinkedIn Parallel Coordination

The core problem: email can do 250+ new leads/day, LinkedIn can only do 20 connections/day. They can't run at the same speed. Here's how to handle it.

What This Looks Like in Practice (Week 1)

By mid-week, LinkedIn starts overlapping with email - dual-channel leads get both touchpoints with a 2-3 day stagger. The founder sees the email first, then gets the LinkedIn request. If they accepted the connection, they get the follow-up message with the podcast invite.

Sending Account Allocation

18 whitelisted accounts across 6 domains (click, easy, free, scale, team, work) x 3 prefixes (jamie@, jamie.s@, jamie.schneiderman@). Every account sends at the 30/day cap. No account is shared across campaigns.

DL = daily_limit (total sends/day for campaign). DML = daily_max_leads (new leads/day, always DL/2). Days = leads / DML in business days. Account assignment rule: T1 gets priority accounts. T3 gets lowest priority. No account shared across campaigns. If an account goes down, reduce that campaign's volume - never redistribute mid-run.

One-Click Campaign Deployment (API)

When Jordan gives the go-ahead, a single command creates all 10 Instantly campaigns + HeyReach sequences via API. No manual setup.

Don't run all three at once. Start with Approach A (podcast invite) for 2-3 weeks, measure reply rate. Then A/B test B or C on a subset. One variable at a time.

Approach B: Direct Audit

"I looked at your situation and noticed X." Fastest path to call. Feels more salesy but converts faster for founders who are actively stuck.

Approach C: Workshop Invite

"Small group session for founders at your stage." Exclusive feel, limited spots. People who show up are already interested.

A: VC Partners & Associates (~200 leads)

VCs have stuck portfolio companies. Medium-term play. Founders distrust VC recommendations until Jamie has independent credibility.

B: Fractional CMOs, CROs & COOs (~150 leads)

They handle execution inside SaaS companies, Jamie handles strategy. Cross-referral opportunity, zero competition.

C: EOS Implementers & Business Coaches (~150 leads)

When process (EOS) isn't enough, the issue is strategic. Jamie is the strategy layer on top of their structure.