The missing layer

Building software got cheap. Running it didn't.

That sentence explains most of what has changed at Cloudburst this year, including the pages on this site. It's worth writing down plainly, because the businesses we talk to are living the second half of it and mostly blaming themselves.

What we kept seeing

Over the last two years a lot of software got built fast. Some by freelancers, some by small shops, some by owners using tools that didn't exist in 2023. Much of it works. Almost none of it came with someone accountable for what happens after launch.

The pattern showed up in nearly every conversation we had this spring. A strategic-planning firm with a live product and a weekly dev meeting that had become the owner's second job. A healthcare operator whose platform was part of what they sell, paying for it through project invoices that never looked the same twice. An app that was three-quarters finished and had been "almost there" for six months — because the one person who understood it had moved on.

Different businesses, same shape. One person held all the context. Requests piled up. Nobody could say what was fragile or what a change would cost. The owner became the project manager by default.

That isn't a you problem. It's a missing layer.

What the layer is

Every piece of software a business depends on needs three things after it exists. Someone has to keep it running: hosting stays up, dependencies get updated, security patches land before they matter. Someone has to answer when it breaks, with a real response time. And someone has to make it better on a cadence the business can plan around, so the software keeps pulling its weight instead of slowly becoming a liability.

Builders are set up for the build. Most of them are honest that they aren't set up for the rest. That gap is where Cloudburst now works. We enter after the build, and we run and grow what's there.

What we changed

The front door is now an Owner's Manual. Six plain-language documents about your software, each reviewed and signed off by one of our engineers: what it's built on, who can do what, how the data is organized, a security review, what needs updating, and how well it's tested. One week, $1,500, and it's yours to hand to any developer you ever work with. It ends with a recommended path and a firm quote for our work. If the software still doesn't make sense to you after reading it, you don't pay and you keep the manual.

After that, Operate: one flat monthly number on a two-week cadence. The tiers differ on who owns the backlog, how much moves at once, and how fast we respond. At Run you bring ticket-sized changes. At Grow you bring us tickets. At Scale you bring us priorities and we handle the rest, including the things you'd never think to ask for. No hourly rates, no overages. Hosting and third-party services stay in your name and are billed by those providers, so the number is ours alone.

We retired hourly work entirely. Hourly pricing puts the client in the position of managing our time, which is exactly the job they were trying to get out of.

How we know it works

We run our own products this way. Localyst and winwinGO live in the same workspace, on the same cadence, with the same engineer review before anything reaches production, as the software we run for clients. When we changed how a release gets approved, our own products were the first to feel it. That's the test we trust.

The other thing we trust is portability. Code, data, and backlog belong to the client at every tier and can leave with them at any time. If the only reason someone stays is that leaving would hurt, we've built the wrong thing.

If this sounds like your software

Start with the Owner's Manual. Or talk it through first. Either way, the first conversation is about what's actually there, not what we'd like to build.