Hello {{First name|Predictable Revenue community}},

Some of my favourite content has always been founders taking me along on their journey while it's still happening.

I'm building something new, so here's mine. An update on Predictable, what I've been working on, and why.

At the start of this year I was working 60-hour weeks, minimum, just to keep up. At least 15 of those hours were click work. Mostly Clay, building the same plays for clients over and over. It wasn't hard work. It was just hard enough that teaching someone else to do it right, the same way every time, was its own problem.

I tried anyway. I had 12 clients, each needing about an hour of build work a week on top of our calls. Sometimes 20 minutes, sometimes three hours if it was meaty. So I brought in Sam, a really good Clay builder. After every call I'd spend 15 minutes writing up the job, then anywhere from 15 minutes to an hour reviewing it. Some pieces I still couldn't hand off. And now I was managing someone.

I was trading an hour of work I enjoyed for two 15-minute chunks of work I didn't, plus the occasional rebuild when I hadn't explained the job clearly enough. I was net negative on my time with a contractor who was great at the job.

So I tried something nerdier. I spent a week with Chrome DevTools open, downloading HAR files to sniff my own web traffic and recreate Clay's API, so AI could do what I'd been asking Sam to do. About 40 hours that week, on top of my normal workload. I got it to do one job. Not all of them. My plays touched almost every corner of Clay's UI, and recreating the whole thing was a much bigger project. But it worked well enough to make me want to give AI another shot. Maybe build my own thing.

Because honestly, I hate clicking the same buttons over and over. My work was templatizable in theory, but every client had enough little nuances that starting from scratch was easier. And there was a whole list of fun, interesting list builds I never ran, because three hours of work for a few meetings was never worth the trade. I wanted that to be cheap. I wanted clients to be able to tell an agent "build this sequence, run that play" and have it go.

I wasn't starting from zero. I'd built vclist, promoted it here, and landed one customer. But I've never raised venture money. I was building for a journey I'd never been on, in a space I didn't care deeply about. Intellectually interesting, not fulfilling. It was also kind of a terrible business model. You need it once every 18 months, and most companies that raise one round don't raise the next.

What I'd really built was an alternative Clay tool for a very niche process I didn't care about. So in January I asked the obvious question. What if it could get me out of all this Clay work?

I'd prototyped in Lovable, hit its limits, grabbed the GitHub repo and started building in Claude Code. The tooling rabbit hole is its own post. By March I had something I was willing to take a risk on with one client.

Then Clay did something that helped me greatly. They doubled their prices.

To be fair, some things on their pricing page got more accessible. But the package I'd recommend for the way I set up clients went from $350 to $660 a month. My standard stack was Clay, plus Zapmail at $100, Instantly at $100, and $120 for ten domains. Around $550 a month, a bit more the first month. For the infrastructure behind your whole GTM motion, not terrible, but I could still feel people hesitate at the number. Overnight it became almost $1,100 the first month and around $960 after that, for the same value you got a month earlier.

Most clients didn't complain about a couple hundred bucks. It irked me anyway. That was a real feel-bad moment from a company I'd had a strong affinity for. I work with a mix of venture-backed and bootstrapped startups, and I'm bootstrapped myself. I didn't love recommending a tool that effectively excluded anyone who hadn't raised millions. You can still use Clay on the lower plans. You just can't use it the way I do.

That was my kick in the pants. One client moved over, then two, as my trust in the lists grew. I remember the first one. The list worked great, and I sat there thinking, well, shoot, that worked one time. How do I keep this going for months, if not years? Making sure a client never runs out of plays, campaigns, or data drove most of what I built early on.

Time in Clay was never accretive. Even if something took me twice as long in Claude Code, I could come back and reuse it.

Today that goal is 100% done. One straggling client is moving over in the next month. I spend effectively zero hours a week in Clay and carry a higher client load than I did in January.

That alone would be a win. But a couple months ago I went back to the market to see if people would pay to use the tool themselves, with a bit of my help onboarding. It took about two months to get to 10 paying customers.

Then I took my foot off the gas.

Call it experience or call it baggage. At Carb.io we went from zero to a million in three months and the wheels fell off. Everything stopped working and we couldn't respond to customer demand fast enough.

So I've held at 10 or 12 customers and I meet with them every week. For a while now I've been asking one question: what do I need to build for you to feel like you're getting good value? I'm nearly through that run. The next question is, what do I need to build to get you to recommend me to three friends?

Version one does what it promises. You can build a list and send a sequence. There's a CLI and an MCP server, so you can be like me and run everything from Claude Code without clicking a button. The major holes are plugged. Now I'm adding niche things people ask for, porting over plays I used to run in Clay, and figuring out where to take it next.

About half of what I build comes from those customer calls. The other half comes from my own domain experience. I'm building this for founders, not salespeople, not GTM teams, not folks coming from Clay. That means making something complex easy. It also means I'm comfortable building things nobody's asked for yet, because I know the space well enough to have good taste for what's coming. Some of the time.

My guess for what's next is social listening. A few clients want to act on LinkedIn posts, and I've felt this one myself. I don't mind engaging with people on LinkedIn. I do mind wading through an algorithm to find the right ones. Yes, you can set up searches, but then you still have to read through all the noise every day. I want only the relevant conversations with potential buyers surfaced to me. We'll see what next week's check-in calls say.

That's the September update on Predictable. I'm taking on 10 more beta customers. It's a paid beta, but it includes AI credits and a bunch of my time, so I think it's a pretty good deal. If any of this sounds interesting, hit reply.

What's the work you're still doing by hand every week that you already know how to do in your sleep?

Collin

The Terrifying Art of Finding Customers is the playbook for founders who never planned to do sales but have to figure it out anyway. It pulls together 13+ years of building software startups, building outbound systems for hundreds of B2B startups, and more failed experiments than I'd care to admit, into a practical guide for finding your first 10, 100, and 1,000 customers.

It will help you build a repeatable way to find customers without losing your weekends (or your soul).

Order now and you'll unlock a stack of exclusive bonuses I actually use with my consulting clients: Email me the receipt and I’ll share the resources.

Email me your receipt and I’ll share the resources.