How to Build a Cold Email System

I've said this before and I'll say it again — we've been struggling.. with cold email deliverability. Cold email infrastructure is frustrating - even when following best practices, deliverability remains inconsistent. I researched everything to solve this problem once & for all. Let me break down what actually works: 1) Infrastructure & Setup: -> Domains & inboxes - Never send cold from your primary domain - Use 3-5 sibling domains, 3-5 inboxes each - Keep branding believable; avoid spammy TLDs (.tk, .ml) - Set up Google Workspace or M365 for legitimacy -> Authentication - SPF covers every sender, DKIM at 2048-bit minimum - DMARC from p=none → quarantine once stable (never jump to reject) - Alignment across From/Return-Path is non-negotiable - Test with mail-tester.com weekly -> Compliance - Clear opt-out, real physical address, legitimate interest docs (EU) - Honor opt-outs within 24 hrs max 2) Sending Strategy: -> Warm-up - New domains need 8-12 weeks minimum - Simulate real engagement (opens/replies/forwards) - Use warmup tools like mailwarm, lemwarm or Instantly.ai -> Volume & Pacing - Start 10-20/day per inbox, add +20-50 weekly if metrics stay green - Randomize send windows; 60-120s gaps b/w sends - Respect recipient time zones (9am-5pm local) -> Timing - B2B sweet spots: Tue-Thu late morning & early afternoon - Avoid Mondays (inbox overload) & Fridays (weekend mode) 3) Content & Copy: -> Subject lines - 6-10 words, human and specific - Personalized context beats cleverness every time - Avoid fake urgency, ALL CAPS, excessive punctuation!!! - Test: "Quick question about [specific company pain point]" -> Body - Short, skimmable, 1 idea + 1 ask maximum - Personalize in layers: hyper-custom for top 10%, segment-level for rest - Use natural language, avoid marketing speak - Images and links kill deliverability - use sparingly -> CTA - Make next step tiny (15-min scan, 1-question reply, "worth a chat?") - Single CTA only - multiple options confuse and reduce response 4) List & Data: -> Sourcing - Prioritize intent and fit over volume always - Dedupe domains (max 1-2 people per company per campaign) - Use Apollo, ZoomInfo or Clay for verified contacts -> Hygiene - Verify syntax + domain + mailbox before sending - Remove hard bounces instantly (never retry) - Prune unengaged cohorts quarterly - Never recycle unsubscribed contacts -> Segmentation - Hot/Warm/Cold bands by recency + engagement - Throttle "Cold" segments heavily 5) Monitoring & KPIs: - Delivery rate ≥98%; investigate anything <95% - Bounce rate <2% (≤1% is excellent) - Spam complaints <0.1% absolute ceiling - Track domain/IP reputation, blacklist status weekly - Use seed accounts & inbox tests ps. Have a response/POA for objections like “not the right person” / “not decision maker” / “No longer at company” / “have in-house team already” / “please contact john from abc” You can also use Valley on LinkedIn - book 2 demos/week for every seat.

In 27 months, I grew my SaaS startup from $1M-$10M ARR in with only 1 salesperson doing 100% OUTBOUND SALES. Here’s my 7-step playbook for building a lean, high-velocity cold email machine: Step 1) Make sure you have product-market fit - My thesis for 2023 and beyond is without word-of-mouth, nothing will work - But if you have strong word-of-mouth (from PMF), many things can help speed you up, including high-volume, low conversion outbound - which we ran Step 2) Identify your target audience - We used BuiltWith to ID domains who used complimentary tech - We then ran the site traffic with SimilarWeb - For us the winning combination was Shopify + Klaviyo above 10k/mo uniques Step 3) Get creative about getting email addresses - The *lazy* thing to do is pull directly from the data vendors - We find the contacts on Linkedin, then use Skrapp and Hunter to find emails - We would include up to 5 potentially relevant people and copy them all - We sell to the high end of SMB (so we always include CEO as well) Step 4) Create short email sequences - We have used a very similar 3-part sequences for the last 3.5 years - Email 1: Present the main idea. For us, “We can get you email addresses of your anonymous web traffic, even if they don’t fill out a form” - Email 2: Social proof from brands they truly admire (hard in the beginning) - Email 3: A super-simple use case email explaining the actual workflow Step 5) Write clear and simple copy. - Each email should be no longer than a few sentences - Make each sentence its own paragraph - Have no more than one idea per email - Keep it conversational - CTA is a Calendly link in both the body and signature of the email Step 6) Manually send the emails, rather than using automation tools. - We have a team of VA’s in the Philippines manually sending from Gmail - Any response gets routed back to a BDR/AE in the USA - The team in the Philippines also does the lead gen, very manually - Cold email gurus will say this is stupid and sucks, but … scoreboard! Step 7) A/B test the sh*t out of it. - Change something, pick a winner - Try to hone in on things to optimize - You’re building a machine that improves itself It may not work for everybody, but it worked for us. Happy to answer questions in the comments! 👇 TL;DR Step 1) Have Product-Market Fit Step 2) Identify Target Audience Step 3) Get Email Addresses Step 4) Create Short Sequences Step 5) Write Clear/Simple Copy Step 6) Send Manually Step 7) A/B Test

Cold outreach can feel like a guessing game, but after sending 1000s of cold emails, I’ve learned 1 thing: Cold email success is in the details. Look at these data↓ - 16.9% of emails never reach the recipient's inbox - Deliverability on Google: 95.54% (57.8% inbox, 37.74% promotions tab) (source: Emailtooltester) Here’s my strategy to improve deliverability and boost replies in 2025: 1. Build a “Spam-Proof” Foundation Your emails are only as good as the infrastructure supporting them. Here’s what works: Authenticate Everything ↳ Use SPF, DKIM, and DMARC to gain inbox trust. Smart Infrastructure ↳ Rely on Smartlead'𝘀 SmartServers for a dedicated setup that avoids shared-IP pitfalls. Start Small, Then Scale ↳ Gradually warm up accounts to 50–100 emails/day. Domain Strategy ↳ Rotate campaigns across multiple domains or subdomains to stay clean. 2. Make Them Want to Open Your Email Subject lines and first impressions matter. Here’s how to make them count: Short Subject Lines ↳ Keep them under 5 words—clear beats clever every time. Personalization ↳ Mention their recent work, role, or company name in the opening line. Text-Only Wins ↳ Fancy designs get flagged; stick to plain-text emails. Curiosity-Driven CTAs ↳ Use phrases like, “Would this solve X for you?” 3. Solve Real Problems, Not Generic Ones Your offer needs to hit home—here’s how to make it stick: - Focus on one pain point that resonates with your audience. - Use data-backed proof: “Cut costs by 20%” or “Save 5 hours weekly.” - Keep your pitch straightforward: No jargon, no fluff. - Add social proof—testimonials and case studies work wonders. 4. Test and Refine Before Sending The best campaigns come from constant iteration. Run Spam Tests ↳ Catch issues before they damage your sender score. Test Placement ↳ Ensure your emails land in the inbox across Gmail, Outlook, and others. Iterate Weekly ↳ A/B test subject lines and CTAs to find what works. A Smarter Way to Scale The formula is simple: Robust setup + Irresistible offers + Strategic scaling = Cold email success. Tools like Smartlead SmartServers handle the technical side, leaving you to focus on what matters: building meaningful connections. What’s been your toughest cold email challenge?

I built outbound systems for 30+ teams. The top performers all did ONE thing first. They built infrastructure before writing a single cold email. This usually looks like: Phase 1: Infrastructure (Weeks 1-2) → Buy 10-15 domains, all close variations of your brand → Set 2-3 mailboxes per domain, never more → Add SPF, DKIM, DMARC on every domain → Use custom tracking domains, never shared ones → Split 50/50 between Google Workspace and Microsoft 365 → Plug everything into Instantly, SmartLead, or Lemlist → Turn on mailbox rotation and start warmup for 6-8 weeks Then. Phase 2: Data and Targeting (Weeks 2-4) → Define ICP with painful detail → Build your TAM and TAL from that ICP → Tier accounts into T1, T2, T3 → Run Clay waterfalls with 3+ providers → Use Apollo, Lima Data, Proxycurl, FullEnrich → Verify every email with SMTP before a sequence → Add intent: hiring, funding, tech changes, job posts → Score accounts so reps know where to focus Finally. Phase 3: Copywriting (Weeks 4-6) → Pick pain or gain angle before writing → Build one framework you can tailor at scale → Open with a real detail, not a fake compliment → Use 3-4 step sequences with new angles each time → Plain text only, under 120 words → One clear CTA in every email → Breakup on step 4 and add LinkedIn between steps If you follow this order, your outbound will start feeling like an actual system. A lot of teams stop giving outbound a try after deliverability starts tanking, but when done right outbound can really start being a growth channel. 

Most cold email campaigns FAIL at the foundation And people wonder why their 10/10 subject lines aren't working The fact of the matter is: You can't copywrite your way out of a broken foundation I see this constantly. Founders will spend HOURS ... – testing different personalization techniques – perfecting their email copy – trying new Clay workflows But either: – their ICP is way too broad, or – their offer doesn't solve a real problem, or – their deliverability setup was never configured properly It's like building a skyscraper on quicksand — The hierarchy of cold email success isn't sexy, but it's the difference between 1% reply rates and 5%+: Layer 1: Message-Market Fit Your offer needs to solve a problem prospects actually care about. No amount of personalization saves weak offers. Layer 2: Deliverability Infrastructure Proper domain setup, authentication, and warming prevent your emails from landing in spam. Get this wrong and nothing else matters. Layer 3: List Quality & Segmentation Hyper-targeted lists with verified contacts outperform massive generic databases every time. Layer 4: Personalization Relevant context that proves you researched them specifically. True personalization ties their reality to your solution. Layer 5: Copy & Messaging Short, conversational emails that lead with their problem before your solution. Write like a human helping. Layer 6: Testing & Iteration Continuous A/B testing on offers, CTAs, and value props separates 1% reply rates from 5%+ performers. Data > opinions. — Most people spend 90% time optimizing layers 4-6 When layers 1-3 are completely broken Stop tweaking subject lines when your fundamentals are missing Fix the foundation first, optimize the details second That's where the real results come from

I generated 47 qualified B2B meetings last month without touching a single lead list. My entire cold email system ran on autopilot while I focused on closing deals. Here's the exact automation setup I used: My situation: I was burning 15+ hours weekly on manual lead scraping, email follow-ups, and reply management. My calendar stayed empty while competitors booked meetings. My process: 1. Connected ICPS and Find Email APIs to Instantly AI (saves 80% on lead credits vs. using Instantly's finder alone) 2. Built evergreen lead filters targeting CEOs and founders at 0-100 employee companies in English-speaking markets using technology and keyword filters 3. Set up a custom enrichment waterfall: ICPS finds emails first (free credits), then Find Email as backup, Instantly credits only as last resort 4. Created campaigns with 100 fresh leads flowing in daily automatically from my saved filters (no manual uploads ever) 5. Deployed AI reply agents in "human in the loop" mode through Slack to approve responses before going full autopilot 6. Configured the AI agent to handle objections, answer questions, and auto-send Calendly links when prospects showed buying signals 7. Switched to full autopilot after approving 20-30 AI replies and confirming quality The results: 47 booked meetings in 30 days, 15 hours per week freed up, and my cost per lead dropped by 70%. This system took me 4 hours to build initially but now runs 24/7 without me touching it. The biggest surprise? My AI reply agent converts warmer than I did manually because it responds within 2 minutes every single time. What's holding you back from automating your cold email outreach? Watch the full step-by-step setup process here: https://lnkd.in/dE25_5gk #LeadGeneration #AIAutomation #ColdEmail

I started running outbound campaigns 6 years ago. Then, grew a cold email agency from 0 to $4,500,000+ ARR in the last 2. Here are 7 things I wish I knew when I started: 1/ Spend most of your time on the list Putting your message in front of the right person matters more than anything else. Thus… → Clearly define your ICP → Precisely figure out what problem you’re solving. → Find companies struggling with the problem you can solve. 2/ Provide as much value as you can Focus on giving instead of getting. This means… → Think of the first thing you’d do if you signed your prospect. → Do it already. → Provide it for free. 3/ Write for 1 person Find 1 ideal prospect and write as though they were the only person on your list. Then… → Find 19 others, and repeat the process. → Identify patterns among these 20 people. → Only then, scale/automate your sending to more people. 4/ Fix your deliverability Bad deliverability guarantees your campaigns will fail. Make sure you… → Send few emails (< 50/day) from secondary mailboxes. → Set up your DMARC, SPF & DKIM records. → Warm-up your mailboxes. Don’t track opens. 5/ Go the EXTRA mile Campaigns that CRUSH aren’t just good. They’re great. This requires extra effort such as… → Don’t stop your targeting at job titles & company size. Look at events, triggers, signals, website, org. charts. → Don’t stop at open jobs signals; scan the job descriptions. → Don’t write just 1-2 email variations, write 5+. 6/ Pair cold emails with other channels Cold Email is often the best B2B acquisition channel to start with. But it works better with other channels. For example, you can… → Use cold emails to distribute valuable content pieces/lead magnets. → Make content, and reach out to people who engage. → Run ads to the same people you contact. 7/ Leverage AI & Technology “A fool with a tool is still a fool”. But the right technology can make you 10X more efficient. You could… → Figure out how to automate repetitive processes with AI & Tools. → Look into ColdIQ’s directory of Sales Prospecting Technologies. → Study how these platforms run their own go-to-market. P.S: How are your cold email campaigns performing at the moment?

We’ve grown into a $2M lead gen agency in just 12 months. Here's the exact framework we used. Core principle: Cold emailing is built on a house of cards. The 3 main components are: 1. Deliverability and infrastructure 2. Messaging 3. Response handling They’re all linked— mess up one, and the whole thing collapses. Most customers come to us after hitting a storm And then think, “ah, cold email isn’t the right channel for us.” So, here’s how we solve their problem: 1. Deliverability We buy a bunch of similar domains and set up hundreds of outboxes. (we never touch your main domain) Then, send 6-8 emails/day/outbox. Our best bet: 1. Low volume sending (earlier this meant 50 emails/outbox. Today, it’s < 10) 2. MDMI sending (multi-domain, multi-inbox rotation) 3. ESP to ESP sending (Gmail to Gmail, Outlook to Outlook) 4. Our own internal warm-up pool In short, we diversify risk by setting up multiple outboxes with low-volume sending. Quick tip: Test as many providers as possible. We use about 4-5 for each client. 1. Messaging Experiment - that’s really it. We divide campaigns into: 1. Evergreen: Damn good copy to test value prop. Think of it like a checkboard and cross out what doesn’t work. 2. Incentive: Play on the law of reciprocity. Send a lead magnet or maybe, gift a bottle of wine to build rapport. 3. Trigger and AI: We use our tech to actively listen. A lead is hiring? They’ve got a problem they want to solve. Send them a timely email and please personalize it! We run them all, wide-net campaigns to hyper-focused ones. Then, stick with what works. There is no one-size-fits-all solution. Experiment and find out. 1. Response Handling If you don’t respond in 30 min, your chances to book a meeting drops by 900% That’s insane. Do you know who’s guilty of this the most? Tenured reps (your costliest investment) These folks are busy golfing, thinking, ‘I’ll get to it later." Well...that’s a cost your company just paid. So we built a simple solution: RevReply. It manages speed-to-lead. Analyzes sentiments > overcomes objections > responds > looks at your calendar > finds a good time > sends back a booking link. All in 3-5 min. Phew! It's really that simple. What aspect of cold emailing are you struggling with?