Cold email as a GTM engine

Built from the system Charles Cormier, founder at TopLeads, described on GTM Vault GTM 19 (https://pro.gtmvault.co/episodes/gtm-19). Published free by GTM Vault. Numbers below are the operator's own, from the episode; your results will differ.

When to use this

Use this when a founder or small team wants outbound to produce meetings and closed deals on a small budget, and is tempted to buy expensive tools or paid leads instead. It fits B2B companies that can describe their ICP precisely and can offer something more useful than a sales pitch as the first touch. It also covers what happens after the first call, which is where Charles puts much of his automation.

The system on the record

Charles argues that founders skip cold email for channels where they pay around $250 per lead, when cold email works if you know deliverability, copywriting, product market fit, ICP targeting and how to analyze results (GTM 19, 1:57). The most common failure he sees is landing in spam (GTM 19, 3:09). His system runs as a loop:

1. Sending infrastructure. He buys five .com domains on Namecheap at about $8 each, hosts them on a shared plan at $35 a year, and sends through the SendGrid Essentials 50K plan (GTM 19, 5:25 and 6:12). Instantly handles sending, contacts and warm-ups (GTM 19, 6:12).
2. Data. Contacts come from Apollo, ContactOut and other sources, with scraping through Apify, and enrichment in Clay and Persana. He discloses that he advises Persana and is affiliated with it (GTM 19, 6:12).
3. Targeting. He stacks hard filters, then AI filters, then triggers. In one campaign: ex-Palantir founders, headcount growth above 25%, funded, more than two people on the sales team; then AI filters for who they target, whether they are B2B SaaS, and whether they hired experts recently; then triggers. The output was a list of 97 individuals (GTM 19, 7:22).
4. Two-step offer. He first invites the prospect onto his podcast. If they do not respond, he pitches them directly (GTM 19, 7:22). He treats podcasting as a lead generation play for B2B, not a brand play (GTM 19, 5:25).
5. Copy. He writes with a well-prompted ChatGPT (GTM 19, 9:11).
6. Post-call conversion. A custom ClosedGPT takes everything known about the prospect after the call and writes a closing email: a plan with a timeline, KPIs and ROI if they work together. It runs on Zapier so the email goes out when the call ends (GTM 19, 10:20). The email names their ICP, their type of business and competitors (GTM 19, 11:30). A pain point GPT shapes later follow-ups, then he connects on social channels and runs occasional remarketing (GTM 19, 12:45).
7. Nurture. A daily newsletter that gives away GPTs and lists keeps non-buyers warm until they convert (GTM 19, 13:57).

For scale, he would keep the same stack and add volume under strict limits: at least 100 domains, no more than 10 mailboxes per domain "because that can get you flagged", 20 cold emails per mailbox per day, and at least 100 warm-up emails for every 20 cold emails (GTM 19, 11:30 and 12:45).

The result

1. Stack cost: he says he would only need about 200 bucks a month to rebuild it (GTM 19, 11:30). The line items he gives are $40 for five domains plus $35 a year of hosting, $75 in total (GTM 19, 5:25), SendGrid at $20 a month and Instantly at $80 a month for up to 25K contacts (GTM 19, 6:12).
2. Opens around 50%, though he no longer tracks them (GTM 19, 6:12).
3. Reply rate around 3% to 4% on podcast invite campaigns, around 1% on direct sales campaigns, and around 2% direct conversion on any campaign he starts (GTM 19, 6:12).
4. Sales cycle of 14 to 28 days on $5k deals (GTM 19, 4:23).
5. Newsletter nurture converting 15% to 30% in his best seasons (GTM 19, 13:57).

Run it on your company

1. Ask once, in one message, for: the offer and average deal size; the ICP in plain words; the data sources and enrichment tools available; the monthly budget; current domains, mailboxes and sequencer, if any; the first-touch offer that gives value before a pitch (a podcast seat, a teardown, a free GPT, a list); the sales cycle length; and whether a call recorder or transcript source exists for post-call automation.
2. Build the infrastructure sheet. Size domains and mailboxes to the budget using Charles's limits: no more than 10 mailboxes per domain, 20 cold sends per mailbox per day, 100 warm-up emails per 20 cold. Output: a table with columns Domain, Mailboxes, Daily cold sends, Daily warm-up sends, Monthly cost, plus a total monthly sending capacity line and a cost line by tool.
3. Write the ICP filter spec in three layers. Hard filters (firmographics such as headcount growth, funding, sales team size), AI filters (questions an AI column answers from the website, such as who they sell to and whether they are B2B SaaS), and triggers (recent hires, funding, launches). Output: a table with columns Layer, Filter, Source, Pass rule, and a target list size estimate. Aim for a list small enough to read, as Charles did with 97 names.
4. Design the two-step offer. Step one is the value-first invite; step two is the direct pitch for those who do not respond. Output: two short emails with merge fields, each under 90 words, plus the delay between them.
5. Write the copy prompt. Produce a reusable ChatGPT or Claude prompt that takes one row of the filter output and returns a first-touch email naming the prospect's ICP and situation. Output: a fenced prompt block and three sample outputs for user review.
6. Write the post-call closer. Produce a prompt that takes a call transcript or notes and returns a closing email with a plan, a timeline, three KPIs and an ROI estimate, naming their ICP, business type and competitors. Then specify the trigger: call ends, transcript arrives, draft is created for review or sent. Output: the prompt plus a step list for the automation tool the user has.
7. Write the nurture plan. Define the follow-up count after the closer, the social and remarketing touches, and the newsletter cadence with what it gives away. Output: a table with columns Day, Channel, Touch, Asset.
8. Set the scorecard: replies, positive replies, meetings, closes, cost per meeting, reviewed weekly, separated by podcast-invite and direct-pitch campaigns.

Tools on the record

1. Namecheap: domain and sending infrastructure.
2. SendGrid: email delivery.
3. Instantly: cold email sending at $80/month, in a stack he puts at about $200/month.
4. Apollo: contact data for podcast-led outbound.
5. ContactOut: contact data alongside Apollo.
6. Apify: scraping to build prospect lists.
7. Clay: enrichment and list automation.
8. Persana: lead intelligence feeding outbound.

Any tool doing the same job works; the method is the point. Full stack by layer: https://pro.gtmvault.co/stack/operator/charles-cormier

Where it breaks

1. Deliverability. Charles names landing in spam as the most common mistake (GTM 19, 3:09). Skipping warm-up or overloading domains past his limits defeats the low-cost stack.
2. A loose ICP. The system depends on stacked filters producing a short, qualified list (GTM 19, 7:22). A broad list turns it back into a volume play.
3. No value-first offer. His higher reply rate comes from podcast invites, not direct pitches (GTM 19, 6:12). Without a credible first-touch offer, expect direct-pitch numbers.
4. No post-call system. The closer and nurture are where calls become revenue (GTM 19, 10:20). A team that books meetings but follows up by hand loses that step.
5. Channel risk. LinkedIn suspended his account, which is why he moved follow-up weight to his own newsletter (GTM 19, 13:57). Keep the nurture list on a channel you own.