how I built a cold outbound engine that books 90% of our calls

wordweaver

I love building systems, but I hate when people think it's just about sending more emails. Spent a year building our outbound engine from scratch at a Series A B2B company (selling to mid-market logistics), and here's the messy truth of what actually worked.

When I joined, the 'system' was one SDR using Apollo with his personal Gmail and a Google Sheet. No CRM, no warmup, no verification. The CEO wanted to flip from 90% inbound/referral to a repeatable outbound engine. Thought it'd be easy. It was not.

First two months: mapping data flows on a whiteboard. Where do leads come from? How do they get enriched? What happens when someone replies? And most importantly - how do I report on it so the CEO stops asking 'is outbound working' every Monday? I became obsessed with dashboards. Nothing shuts a CEO up faster than a shiny ROI chart they can't argue with.

By March we had a stack: Apollo for list building, Prospeo for enrichment (email accuracy around 84% - solid for our ICP), Bouncer for verification, and Instantly for sending. Set up 8 inboxes via Mailforge, warmed them for 21 days before sending a single email. Paranoid about deliverability, and that paranoia was justified later.

The workflow: SDR builds list in Apollo → export CSV → run through Prospeo → Bouncer → clean list into Instantly with sequences → replies pushed to Salesforce via Zapier, tagged as outbound. Simple on paper. Took six weeks to stop Zapier breaking on edge cases.

By April we were sending 40 emails per inbox per day - about 320 total. Reply rate was 1.8%, bounce under 2%. But meetings booked? Maybe 2-3 per week. CEO not thrilled.

Then I made a classic ops mistake: assumed the problem was volume. More emails = more replies, right? Scaled to 600/day by adding more inboxes. Reply rate dropped to 1.1%, one domain got flagged, we had to burn it. Lost weeks of pipeline. Costly lesson.

The real problem was targeting and copy, not volume. Our positive reply rate (actual interest) was 0.4%. The sequences were five-step marketing fluff the SDRs copied from some course.

June was a reset. Scaled back to 300 emails/day, killed the burned domain, rewrote sequences with SDRs: shorter, three steps max, personalised first lines. Tightened ICP - targeted companies that had recently posted logistics roles on LinkedIn (growth signal). Built a scraper for LinkedIn Sales Navigator data.

By late July, positive reply rate jumped to 2.3%, meetings went from 2-3 to 6-8 per week. Better meetings too - AEs closed at higher rates. I built a Salesforce dashboard showing full funnel from email sent to closed won revenue, with cost per meeting and opportunity. The CEO started showing it in board meetings. That's when I knew it worked.

August to October was optimisation. Added Clay for enrichment edge cases (tech stack, funding data) but kept Prospeo as primary. Switched to Bouncer + Scrubby for catch-all domains, bounce rate dropped to 1.2%. Tried Dripify for LinkedIn touchpoints - not convinced it's worth the $79/seat.

November hit our stride: 3 SDRs, 15 inboxes, 450-500 emails/day. Reply rate 2.5-2.8%, positive reply 1.4-1.6%, 10-14 meetings per week. Cost per meeting ~$47 when adding up all tools (Instantly, Mailforge, Apollo, Prospeo, Bouncer+Scrubby, Clay, Zapier). About $580/month in tooling before salaries. Pretty efficient.

The reporting piece is what I care about most. Built four Salesforce dashboards: daily sending metrics, weekly pipeline, monthly ROI, and deliverability health. That last one I check every morning - if any inbox bounce rate exceeds 3%, pull it immediately. The monthly ROI dashboard got us budget for a third SDR. When you can say 'outbound sourced $340k pipeline last quarter at 4.7x ROI', budget conversations get easy.

Now, early 2025, about 90% of sales calls come from outbound. Inbound is maybe 2-3 meetings a week vs 12-15 from outbound. System runs smoothly on 5-6 hours maintenance per week. SDRs are mostly autonomous.

Things I'd do differently: Don't scale volume before nailing targeting and copy. That cost me two months. Set up full reporting from week one. Start with fewer inboxes - add more only when reply rates justify it. The temptation to throw volume at a broken process is real, and it's almost always the wrong move.

Anyway, this got long. The system isn't perfect - some weeks meetings drop for no obvious reason and I'm digging through data. But it works, it's repeatable, and leadership sees exactly what it's producing. That's the part I care about most.

marketingmule

Waterfall enrichment is a solid approach, but that spot-check threshold is where most teams slip up. is it 20 % by account score or by signal freshness that you're prioritising? we switched to freshness-weighted - anything triggered in the last 14 days gets a manual pass regardless of account tier - and our bounce rate dropped more than any ICP tuning ever managed. Out of curiosity, how are you handling the Slack routing when a reply has ambiguous intent? that handoff can kill reply speed if the process isn't tight

brandvoice

Love seeing a system that actually works. For me, the real signal that targeting and copy are dialled in is when the prospect refers back to a specific pain point I mentioned in the email - unprompted - within the first couple of minutes on the call. if they're already nodding along before I've even said hello, the copy's doing its job.

What's your own litmus test for knowing it's working?

socialbutterfly

Love the system breakdown, but one thing to flag ☝️ 2.5-2.8% reply rate on 450-500 sends a day is decent, but calling it 90% of sales calls feels a bit generous when inbound was never really established. Of course outbound dominates when that's the only channel working. I've watched similar setups plateau hard around month 8-10 once the obvious ICP segments are tapped out - reply rates can slip to 1.5% or even lower. Curious how your numbers look by Q3 this year.

weekendhustle

Our tooling costs are pretty similar actually, around $600/month all in. The bit about Salesforce dashboards is spot on - that's what separates ops people who keep their budget from those who get their stack cut every quarter 😊 Our cost per meeting runs about $50-55 but we sell into healthcare which is a notoriously tough nut to crack. One question though - what's your total monthly spend including SDR compensation and how does leadership view that ratio versus just the tooling cost? because once you add $15-18k in SDR salaries the ROI maths changes dramatically, and that's the figure the CFO actually cares about.

leadgen

Oh man, I feel this. Spent months in the same loop - Marketing claiming the last touch organic visit, Sales saying the cold email set the stage. Our RevOps team eventually built a multi-touch attribution model in Salesforce, but it's still a fragile peace. Out of curiosity, have you tried using weighted attribution in your SFDC + Outreach sync?

paperclick

Great write-up - thanks for sharing. I've seen this exact pattern play out in a few of the teams I've consulted for. Everyone assumes the bottleneck is volume when actually it's targeting and messaging. Scaling garbage just means you burn domains faster.

The bit about reporting being the real leverage hits home. I spent months chasing better copy and tighter ICP before I realised that if leadership can't see the cost-per-meeting and pipeline attribution in a dashboard, you'll never get the budget to scale what actually works. Once I built that visibility, buy-in became trivial.

A few things I'd underline from this:

Warming inboxes for 21 days is smart, but most people skip verification and ruin deliverability before they even start.
Short sequences (3 steps max) force you to make every email count - no filler.
Hiring signals are underrated. I've seen similar results using companies that just raised a Series A or hired a VP of Sales.

Also - your reply rate jump from sub-1% to over 2% sounds small, but in practice that's the difference between 2 meetings a week and 8. That's the kind of shift that changes how a whole team operates.

The cost per meeting at $47 is solid. People obsess over tools when the real lever is systems thinking: a tight ICP, a short sequence, clean data, and a reporting loop. Everything else is optimisation after that.

ctrjunkie

The biggest shift for me was realising how much better it works when you stop blasting lists and actually catch people right when they're asking for help. Reply speed? Irrelevant when you're already in the room before anyone else even knows the question exists.

Intent signals from forums, job boards, niche communities - that timing edge no sequence can touch. Found a tool that surfaces those live signals every day, and honestly it's completely changed how I think about building pipeline. Feels less like cold outreach and more like showing up at the right moment.

marcuthread

That bit about reporting is bang on. Almost every outbound system I've seen falls apart because people get fixated on reply rates and forget to track what actually turns into a closed deal. You're just measuring activity, not results

leadgen

I've actually looked into that course with Enrique too. Has anyone here taken it? i keep seeing ads but I'd rather hear from someone who actually went through it - good or bad

copycat

Thought the same exact thing - envy's practically our national sport too. Hard to break that habit.