SDR Playbook for Seed-Stage Startups

The biggest mistake in seed-stage outbound is over-engineering sequences. Six-touch, multi-channel epics with LinkedIn DMs, video messages, and custom landing pages sound impressive but they're a distraction when you haven't proven the core message yet. Start with a 3-touch email sequence. Touch 1: personalized cold email, 150-200 words, problem-focused. Touch 2: 3-day follow-up, 50 words, add one data point or social proof. Touch 3: 7-day breakup email, 30 words, ask if timing is wrong. Use Lemlist or Instantly.ai to send sequences. Both handle deliverability well, have solid analytics, and cost under $100/month at seed volumes. Set up a dedicated domain for outbound (yourcompany.co or yourcompany.io) and warm it up for 3-4 weeks before sending at volume. Domain warming means starting with 10-20 emails/day and increasing by 10 emails/day each week. Skipping warmup is the #1 cause of seed-stage outbound failures — your emails go to spam before prospects ever see them. For email copy, follow the AIDA framework but compress it: one line on who you're emailing and why it's relevant to them (Attention), one line on what problem you solve (Interest), one line on proof it works (Desire), one clear CTA (Action). Your CTA should be low-commitment: 'Worth a 15-minute call this week?' converts better than 'Book a demo' at seed stage when there's no brand trust yet.

The founder should run outbound personally for at least the first 30-60 days. Not because you're the best SDR, but because you'll learn things no hired SDR will ever tell you: why prospects aren't responding, what objections keep coming up, what language resonates. This learning is priceless for positioning, messaging, and product roadmap. Set up a simple system: 1 hour per day, 20 cold emails personalized and sent, track every reply in a simple Notion or Airtable tracker. Don't use a CRM until you have more than 5 deals in flight — the overhead will kill your momentum. Use Gmail with the Gorgias or Text Blaze extension for templating to speed up personalization without losing quality. When you get a reply — positive or negative — respond within 2 hours. Reply rate decays fast. A prospect who replies on a Tuesday afternoon and doesn't hear back until Wednesday morning is already less interested. Set your email app to push notifications during business hours. For negative replies ('not interested', 'wrong timing'), respond with gratitude and ask one question: 'Completely understand — if you don't mind me asking, what would make this relevant for you in 6 months?' This gives you intel that sharpens your ICP and sometimes converts to a pipeline deal 3-6 months later.

Hire your first SDR when you've personally sent 500+ cold emails, achieved a 2%+ positive reply rate, and have a proven sequence that books meetings. If you hire before hitting those benchmarks, you're paying someone to discover what you should have discovered yourself — and they'll be less motivated to iterate. The right first SDR profile at seed stage is different from what you'd hire at Series A. You want someone who's hungry, coachable, and self-directed — not someone who needs a structured program and manager hand-holding. Look for: 1-2 years of SDR experience at a startup (not a big company), a track record of outperforming quota, and evidence that they're curious about your space. Ask during interviews: 'Show me a cold email you'd send to our top ICP.' Their answer tells you everything. Pay ranges: $55-70K base + $15-25K OTE at seed stage is reasonable in most US markets. Equity is important — offer 0.1-0.25% vested over 4 years to attract talent who believes in the mission. Set clear 30/60/90 day targets: 30 days (ramp) = 50 activities/day, 60 days = first 3 meetings booked, 90 days = pipeline quota of 3x monthly revenue target. Build a simple scorecard and review it weekly. Don't wait 90 days to give feedback — weekly 1:1s with an honest performance conversation are non-negotiable.

You don't need a $50K/year outbound tech stack at seed. Here's the lean setup that works: Apollo.io ($49/month for prospecting and enrichment), Instantly.ai ($37/month for sequencing and deliverability), Notion for deal tracking ($0), and Loom for video messages when you land in key accounts ($12/month). Total: under $100/month. Once you're booking 10+ meetings/month, add HubSpot CRM (free tier works until you need deal pipelines and reporting). At 20+ meetings/month, add Clay ($149/month) for advanced enrichment and personalization at scale. The mistake is buying tools you don't need yet — $500/month in SaaS spend at seed stage is real money that should be going to experiments and talent. For LinkedIn outreach, use Sales Navigator ($99/month) when you're ready to add a second channel. But don't add LinkedIn until you've proven email works — multi-channel outbound doubles complexity and if your messaging isn't right, you'll just spam people on two channels instead of one. LinkedIn DMs have a higher open rate than email but lower reply rate for initial outreach, so it works better as a follow-up channel once someone has seen your email.

Track these metrics weekly, in a simple spreadsheet: emails sent, open rate (target 40-60%), reply rate (target 2-5% positive), meetings booked, show rate, and pipeline generated. If you're using Instantly or Lemlist, these are built-in. The most important metric at seed stage is positive reply rate — open rate is vanity, meetings booked is the metric that matters. Run a weekly iteration meeting, even if it's just you and one SDR. Review: what's working (don't touch it), what's not working (kill it or fix it), and one experiment to run next week. Good experiments: test a new subject line, test a new first line, test a new CTA, test a different ICP segment. Bad experiments: redesigning the entire sequence, adding three new channels, or changing your offer. One variable at a time. If you're getting opens but no replies, your first line is weak. If you're getting replies but mostly negative, your offer isn't resonating. If you're getting meetings but no pipeline, your qualification is off. Each problem has a different fix — diagnose before you act. Don't change your sequence mid-test. Let each version run for at least 100 sends before judging it. Statistical significance matters even at low volumes.