60% Open Rate. One Prompt. No SDR.

Rob — May 14, 2026 · 5 min read

We dogfooded Sandbox on our own GTM for 90 days. I want to share the honest numbers — not a polished case study, but the actual scorecard.

Because the thing operators ask most often isn't "what can Sandbox do?" It's "does it actually work?" Here's the answer.

What We Set Out to Do

The goal was simple: replace every manual growth task with a Sandbox-run agent workflow. That included:

The constraint: no SDR, no marketing coordinator, no ops person. Just one operator (me) reviewing and approving, then Sandbox doing the work.

The Actual Numbers

Qualified contacts sourced & enriched 851
Cold outreach sequences launched 2 active campaigns
Average open rate 60%
Time to go from "I want to reach pharmacy owners in FL" to live campaign ~4 hours
Blog posts published 30+
Human hours per week on GTM ~4
Equivalent headcount cost replaced $180K–$240K/yr

The 60% open rate is the number people latch onto. I want to be specific about what drove it — because it wasn't a clever subject line trick.

Why the Open Rates Are High

The emails are plain text. No graphics, no headers, no branded footers that signal "newsletter." They read like a message from a human who researched you specifically, because the context was pulled from real signals about each company.

Sandbox pulled company descriptions, LinkedIn signals, and industry context for each contact. The first lines reference something specific to their situation. Not "I noticed you're in marketing" — but a line that reflects actual knowledge about what they're working on.

Industry benchmark for cold outreach: 15–25% open rate. We're at 60% because the targeting is tight and the copy isn't generic.

We didn't write any of those first lines manually. One prompt described the ICP and the angle. Sandbox drafted the personalization for each contact in the sequence.

What It Actually Replaced

Before Sandbox, running this kind of outreach motion required:

Total: $145K–$205K in salary, before tools and management overhead.

What we spent: a fraction of that. The savings compound every month.

What Didn't Work

I want to be honest about the failures too:

The first outreach sequence copy was too product-focused. We described what Sandbox does before establishing what problem it solves. Low response. Took two rewrites to get it problem-led.

LinkedIn content took 3–4 rounds of voice calibration before posts sounded like me instead of "AI marketing content." The early ones were too polished — operators could tell. We had to dial in the rougher, more direct tone.

I still have to personally handle every reply. That's intentional — warm conversations belong to humans. But it's the constraint on scale. When the campaigns get more replies, that becomes a real time investment.

The Honest Summary

Sandbox doesn't replace judgment. Every prompt I gave, every sequence I approved, every piece of content I reviewed — that was still my decision-making in the output.

What it replaced was the execution volume. The research. The drafting. The scheduling. The follow-up. All the work that was always supposed to happen but got pushed out of the week because I was running the business.

For operators at our stage — early validation, running lean — this is the gap between "I know what I should be doing" and "it's actually happening."

Want to see the setup running on your actual use case?

Book a 20-minute demo: cal.com/rob-sandbox

You describe your ICP. I run a live prompt. You see exactly what Sandbox produces — list, sequences, and all — in real time.

Sandbox is the operations engine for operator-led businesses. Prompt in, working business out.