Most cold email advice comes from marketers who have never set up a sending domain or read an email header. If you're a technical founder or CTO, you already know that deliverability is an engineering problem, not a copywriting problem. The good news: when you treat cold email like infrastructure, you can get 3-5% reply rates and 10-20% positive response rates from decision makers who actually need what you're building.
Cold email is still the cheapest way to get B2B meetings. A well-run campaign costs $300-800 per month in tools and domains, and can book 5-15 qualified meetings. Compare that to $3,000-5,000 per month for LinkedIn ads or trade show booths. The catch is that you need to set up the technical foundation correctly, or you'll land in spam and waste the budget.
In this post, I'll walk through the exact setup we've used on IRPR builds and client projects: domain configuration, list building, copy that gets replies from technical buyers, and a 30-day launch timeline. No fluff, just the numbers and tools that work in 2025.
- Open rate: 35-55% after warmup. Below 30% usually means deliverability issues or bad list hygiene.
- Reply rate: 2-5% for cold outreach to technical roles. Above 5% means your offer or personalization is exceptional.
- Positive response rate: 10-20% of replies should be interested or asking for more info. The rest are objections or "not now".
- Meeting booked rate: 0.5-1.5% of total emails sent. For 1,000 emails, expect 5-15 meetings.
- Cost per meeting: $20-80 including tools, domains, and list data. Far cheaper than paid ads for B2B SaaS.
Technical Setup: From Domain to First Send
- 1
Buy separate sending domains
Never send cold email from your primary domain. If it gets flagged, your company's regular email and website reputation suffer. Buy 2-3 similar domains (e.g., getyourcompany.com, tryyourcompany.com) from Namecheap or Google Domains. Cost: $10-15 per domain per year.
- Use .com or .io only. Avoid .xyz, .info, or other cheap TLDs that spam filters distrust.
- Set up 2-3 mailboxes per domain using Google Workspace or Microsoft 365. Each mailbox can send 30-50 cold emails per day after warmup.
- Total sending capacity: 3 domains x 3 mailboxes x 40 emails = 360 emails per day. That's 7,200 emails per month.
- 2
Configure SPF, DKIM, and DMARC
This is non-negotiable. Email providers check these records to verify you're not spoofing. Without them, your emails go to spam or get rejected outright.
- SPF: Add a TXT record listing the IPs allowed to send for your domain. For Google Workspace, it's "v=spf1 include:_spf.google.com ~all".
- DKIM: Enable in your email provider and add the generated TXT record. This signs each email cryptographically.
- DMARC: Set policy to "p=none" initially, then move to "p=quarantine" after 2 weeks of monitoring. Use a tool like DMARC Analyzer to see who's sending on your behalf.
- Optional but recommended: BIMI record with a verified logo. It increases open rates by 10-15% because your logo shows in the inbox.
- 3
Warm up each mailbox for 2-3 weeks
A brand new domain sending 50 emails on day one is a red flag. Use a warmup tool like Instantly or Smartlead to gradually increase volume. These tools send automated emails to real inboxes and reply on your behalf to build reputation.
- Start with 5 emails per day per mailbox, increase by 5 every 2 days until you hit 40-50.
- Warmup should include a mix of Gmail, Outlook, and corporate domains. Avoid all-Gmail warmup lists.
- After 3 weeks, check your domain reputation with Google Postmaster Tools. Aim for "High" reputation before sending real campaigns.
- 4
Build a targeted list with verified emails
List quality is the #1 factor in reply rates. Don't buy a list of 50,000 contacts from a data broker. Instead, use tools like Apollo.io, Hunter.io, or LinkedIn Sales Navigator to find decision makers at companies that match your ICP.
- Define ICP: industry, company size, tech stack, funding stage, job titles (e.g., CTO, VP Engineering, Head of DevOps).
- Use Apollo.io's filters to export 500-1,000 contacts per campaign. Cost: $49-99 per month.
- Verify every email with NeverBounce or ZeroBounce. Expect 10-15% to be invalid. Remove them before sending.
- Enrich with firmographic data: company revenue, recent funding, tech stack from BuiltWith. This enables personalization.
- 5
Write copy that sounds like a human engineer
Forget "I hope this email finds you well" and "I came across your website." Technical buyers see 50 cold emails a day. Your first line must show you did research and have a specific reason to reach out.
- Subject line: 3-5 words, no emojis, no all caps. Examples: "Quick question about [Company]'s API", "Saw your post on Hacker News", "Your [specific tool] implementation".
- Opening line: Reference a specific fact. "Noticed you're hiring for a platform engineer" or "Saw you raised $5M in March" works 10x better than generic praise.
- Body: 2-3 sentences max. State the problem you solve, one proof point, and a low-friction CTA like "Worth a 10-min call?" or "Can I send a 2-min Loom?"
- Signature: Just your name, title, and company. No logos, no legal disclaimers, no "Sent from my iPhone".
- 6
Set up a sending tool and follow-up sequence
Use a tool like Instantly, Smartlead, or Lemlist to automate sending and follow-ups. These tools rotate mailboxes, track opens and replies, and let you A/B test subject lines.
- Recommended sequence: Email 1 (day 0), Follow-up 1 (day 3), Follow-up 2 (day 7), Breakup email (day 14). That's 4 touches total.
- Each follow-up should be shorter than the previous and add new value: a case study, a relevant stat, or a different angle.
- Set daily sending limit per mailbox to 40-50. Spread sends across 2-3 hours to look natural.
- Track reply rate, not open rate. Open tracking is unreliable due to Apple Mail Privacy Protection. Focus on replies and meetings booked.
Advanced Strategies to Push Reply Rates Above 5%
Use AI personalization at scale without sounding robotic
Tools like OpenAI GPT-4o or Claude 3.5 can generate personalized first lines based on a contact's LinkedIn profile, company news, or tech stack. But raw AI output is generic. The trick is to feed it specific data points and constrain the output.
- Example prompt: "Write a 1-sentence opening line for a cold email to [Name], CTO at [Company]. They use [Tech Stack], recently [trigger event]. Mention [specific fact] and ask about [pain point]. Keep it under 15 words, no fluff."
- Cost: $0.01-0.02 per email with gpt-4o-mini. For 1,000 emails, that's $10-20. Worth it for 2x reply rates.
- Always review AI-generated lines before sending. One wrong fact kills trust instantly.
Leverage trigger events for 3x higher reply rates
Emails sent within 24-48 hours of a trigger event (funding, hiring, product launch, tech stack change) get 3x more replies than cold outreach to a static list. Set up alerts using Google Alerts, LinkedIn Sales Navigator, or tools like Triggr.
- Funding round: "Congrats on the Series A. Most teams at this stage struggle with [problem]. We built [solution] for that."
- New job posting: "Saw you're hiring a DevOps engineer. Many teams we work with use [tool] to reduce infra costs by 30%."
- Website change: "Noticed you moved from WordPress to Next.js. We've helped 20+ teams make that migration without downtime."
Test offers, not just subject lines
Most people A/B test subject lines and call it a day. The real lever is the offer. Test three different value propositions with the same list and see which gets the most positive replies.
- Offer A: 30-min free technical audit (e.g., "I'll review your AWS bill and find $500/month in waste").
- Offer B: Access to a benchmark report (e.g., "We surveyed 100 CTOs on API latency. Want the raw data?").
- Offer C: A relevant case study (e.g., "How [similar company] cut infra costs by 40% in 6 weeks").
- Run each offer to 300 contacts, measure positive reply rate, then scale the winner.
Mistakes That Kill Deliverability and Reply Rates
Sending from your primary domain
If your cold email domain gets blacklisted, your company's transactional emails (password resets, invoices) stop working. Always use separate domains for cold outreach. We've seen founders lose access to their main email for weeks because of this.
Skipping warmup or rushing it
Sending 100 emails from a 2-day-old domain is a guaranteed spam flag. Warmup is not optional. Budget 3 weeks minimum before your first real campaign. If you need leads faster, use LinkedIn outreach in parallel.
Buying a list from a data broker
Those 50,000 contacts for $99 lists are scraped, outdated, and full of spam traps. Hitting a spam trap even once can destroy your domain reputation. Build lists from Apollo.io, LinkedIn, and manual research. Quality over quantity.
Using images, attachments, or HTML-heavy templates
Plain text emails get 20-30% higher reply rates than HTML emails with images. Images trigger spam filters and look like marketing. Keep it to 3-5 short paragraphs, no formatting beyond line breaks.
Ignoring unsubscribe requests and replies
If someone replies "remove me" or "stop emailing", do it immediately. Not doing so violates CAN-SPAM and gets you reported. Set up a rule in your sending tool to auto-unsubscribe on keywords like "unsubscribe", "remove", "stop".
Cold Email Is an Engineering Problem, Treat It Like One
Cold email gets a bad reputation because most people do it wrong: they buy lists, skip warmup, send HTML templates, and blast 10,000 emails from one domain. When you treat it like an engineering project with proper infrastructure, testing, and iteration, it becomes a reliable lead generation channel that costs a fraction of paid ads.
The numbers speak for themselves: a well-run campaign with 3 domains, 9 mailboxes, and 7,200 emails per month can book 30-50 meetings at a cost of $20-50 per meeting. That's a 10x better ROI than most B2B marketing channels.
If you're building this in-house, start with the checklist above and don't skip the technical setup. If you need help with the custom infrastructure (domain management, email sending at scale, CRM integration, AI personalization pipelines), IRPR has built these systems for multiple clients. We can set up a sending engine that handles 10,000+ emails per month with 99% deliverability. Book a discovery call to get a technical plan for your outbound stack.
The IRPR engineering team ships production software for 50+ countries. Idea → Roadmap → Product → Release. 200+ products live.
About IRPR