We Built Sandbox to Run Our Own GTM. Here's What 90 Days of Dogfooding Actually Looked Like.

Six months ago, I had a problem I bet a lot of you have had.

I knew exactly what I needed to do for growth. I knew who I needed to talk to, what to say, how often to follow up. I'd done it before with a full team. The playbook was clear in my head.

The problem: I was building the product, managing the money, doing the support. The growth playbook stayed in my head because there was no one to run it.

I didn't want to hire a full-time growth hire for a product that hadn't hit product-market fit. I didn't want to pay an agency $8K/month to send templated emails. I wanted the execution — not the headcount.

So we built Sandbox. And then we used Sandbox to launch Sandbox.

Here's what that actually looked like.

The Setup

We defined our ICP: small business operators, 5–50 employees, founder-led, running GTM themselves (or trying to). Think consultants, field service businesses, marketing agencies, e-commerce operators. People with a clear growth problem and no one to hand it to.

We wrote the outreach copy once, in plain language. Not templates — actual language we'd use in a real conversation.

Then we told Sandbox what we needed: "Find operators in [city] matching these criteria. Run the sequence. Stop when they reply."

What Happened: The Numbers

The time number surprised me most. Not the output volume — I expected the volume. The surprise was that 4 hours felt like a lot. It had been so much more before.

What Worked

The outreach worked better than agency templates because it was grounded in real company signals — things I actually knew about the people we were reaching. Sandbox pulled the context; I approved the framing. The result felt personal because the intent was personal.

The LinkedIn content worked because we used it differently than most founders do. Instead of "thought leadership," we published from what was actually happening in the business. Client conversations that revealed something. Problems we were working through. Honest progress updates. That's the content that makes operators want to reach out — not polished takes.

The follow-up consistency worked because it was systematic, not motivational. It didn't depend on me remembering. Every stalled conversation got a drafted next step. I reviewed and approved. The ones that needed to go, went.

What Didn't Work (Honestly)

The first outreach sequence copy was too product-focused. We talked about what Sandbox can do when we should have led with the problem the operator was experiencing. It took two iterations to get to problem-led copy that felt right.

Early LinkedIn posts took 3–4 rounds of revision before the voice sounded like mine. The first drafts were too "markety" — too polished, not specific enough. The breakthrough was giving Sandbox more raw material: actual Slack messages, real email threads, specific decisions we'd made that week. The more context I gave, the better the output.

And I still handle every reply personally. That's intentional — warm conversations belong to humans. But it's the constraint on scale. Everything upstream can be systematized. The conversation that matters still needs me in it.

Where We Are Now

We're in early validation. First cohort of operators onboarding now. The thesis is holding: operators who are already competent at GTM get the most out of this, because they know exactly what they want and can tell when the output is wrong.

We're not pitching this as "AI does your growth for you." We're pitching it as: your judgment, running at 10x speed.

If you're building something and running out of runway to do GTM alongside the actual product — that's exactly who this is for.

Sandbox is the operations engine for operator-led businesses. Built for founders who know exactly what needs to happen and need it to actually happen.