I kept my stack cheap on Cloudflare so I could learn how to sell.
In a September 2026 live stream I laid out what I'd tell a technical founder starting today. Set up the company properly, keep the infrastructure bill small, then spend your energy on selling. I'm a self-described developer gone rogue: building means little if you don't know how to make money from it.
The business case
Most developers I talk to know how to build tools. They do not know how to promote them or put the pain into words in a way that drives demand. That gap is the whole point of the stream.
You get three routes, depending on whether you have money or only time. With a budget, run small hyper-local ads written with the Problem-Agitate-Solution framework and launched with AI. With no budget, earn know-like-trust on forums for about 30 days before you mention your app. For B2B, use cold email. Two traps run through all of it: selling to peers who have no money, and optimizing payment tools before they cost you anything. Pricing and plan fees change, so treat the figures below as September 2026 and recheck the vendor pages before you spend.
Set up the company
If you have never set up a U.S. company, start with Stripe Atlas. On stream I quoted about $500 to incorporate and about $150 a year to maintain, with Stripe handling the paperwork for a Delaware company investors already understand. Incorporating separates you from the business. That is general orientation from the stream, not legal advice. For banking I named Mercury and Brex. Stripe also offers its own business account and cards now; check legality with Stripe for your country before you rely on that path.
Keep the stack cheap
I looked for alternatives to Cloudflare. Vercel and Supabase lost me on pricing. I run nearly everything on Cloudflare. You get global scale from about $5 a month on the cheap end via Workers. My September 2026 usage was 19 million requests for about $2.70. I contrasted that with another cloud bill of about $1,500 that I was cutting. Calling models from a Worker is a few lines through Workers AI. For a deeper map, see Cloudflare SaaS architecture. On checkout, use a Stripe Payment Link plus Stripe's email notification and move on. Your focus is making money, not a seamless checkout on day one.
Why developers sell
Being a developer means nothing if you do not know how to market and sell what you build. My younger self had something valuable and no idea what to do with it. AI can help you build almost anything if you keep iterating. The hard skill is getting the right people to care.
Paid ads with AI
With unlimited money you just run Facebook or YouTube ads. Most developers, me included, do not have that budget. AI now does the ad work. Describe your app, audience and problem to ChatGPT or Claude. Ask for ads using the PAS framework (Problem, Agitate, Solution), have the AI make the images, and connect Meta's ads tooling to build the campaigns in Meta Ads Manager.
Target markets that are not saturated. That is the blue-ocean idea from Blue Ocean Strategy. My real test was ads for my dad's equipment business in Antigua, an island of about 108 square miles and about 80,000 people. About $27 in Facebook ads brought 11 qualified leads, worth about $15,000 of business. Naming the country and keeping the ads hyper-local made people trust they were not spam.
Jump to
- 0:00 Recap: a U.S. company with Stripe Atlas, banking, and my background
- 12:28 Why I run my business on Cloudflare
- 21:53 Stripe checkout questions: use a payment link and move on
- 28:09 Why developers need to learn to sell
- 35:25 Paid ads: the pain, the PAS framework and an AI-built campaign
- 1:04:04 Ad math, Stripe vs RevenueCat, and not optimizing early
- 1:14:23 No budget: Upwork, Reddit and know-like-trust
- 1:34:35 Cold email, my small team and what's next
I walk through the company setup, the Cloudflare bill and the three selling routes on screen in the full video on my channel: Building on Cloudflare, then selling on YouTube.
Ad math and tools
Start small, around $10–20 a day or about $100 for a week. Use the Meta Pixel for retargeting. Set a fixed budget and end date, because the platform will keep spending. Start with three to five ad variations, read the funnel numbers, then make more of what works.
Across four or five experiments for my dad's equipment business I spent about $600 for about $15,000 in returns, not all profit. For every $1 in, you should get at least $3 back, ideally around $6. If you do, keep spending, and keep churn low. After a week of ads, have the AI review the campaigns, diagnose them and propose the next set. In one to two weeks, about $200, you will know whether you have a business.
I use only Stripe, not RevenueCat or Lemon Squeezy, because those sit on top of Stripe and take a cut at volume. If RevenueCat already works for you, do not switch. Do not prematurely optimize. I used to run a web agency that charged about $15,000 to set up plus a $5,000-a-month retainer with a six-month minimum. What took manual setup then takes under an hour with AI now.
No budget routes
If you have time instead of ad spend, find real problems on Reddit and on freelance job boards where people say what they will pay. Do the work, turn the client into a customer, then build the tool around it.
Read the subreddit, note the problems, and use AI to plan genuinely helpful replies. Do this five to ten times a day for about 30 days before mentioning your app. Know, like and trust come first, so do not spam links. Then ask for beta users, get on calls, ask about their problem and what they would pay, and offer early users a big discount later. Document the journey on YouTube. Content wins over the long game.
Solopreneurs are often overwhelmed and working in the business rather than on it. Ask ChatGPT to list your audience's problems, then position for problem-aware buyers first. They are closer to buying.
For B2B, cold email from a bought or built list is the low-cost option. Check cold-email law for your reader's country before you send. Recap of the three routes: ads at about $20 a day for two weeks; forums such as Reddit and Facebook groups; cold email for B2B.
Pick a route today
Do not build a tool to sell to peers who have no money. Sell to people who have money, or who are actively looking for a solution. Keep the stack cheap so the bill never becomes the excuse, then pick one sales route and run it for two weeks.
If you are a developer with a working app and no clear path to first customers, that is what I cover when we start as a solo founder.
About Dwain
I'm Dwain Browne, Founder of LeadScore AI and SnapSuite. I'm a Toronto-based software architect with 25+ years across software, sales, operations and entrepreneurship. SnapSuite helps contractors and trades teams get paid faster, and LeadScore AI shows you which leads deserve your next call. I also run Vibe Code to Profit, a community helping people build software with AI.