How We Work
What happens from the first call to the day you own it, and the promises you can hold us to.

What happens from the first call to the day you own it, and the promises you can hold us to.
Every build starts with a written scope and a fixed price. You pay 50% to start and 50% when an acceptance checklist you signed passes at launch. We build in a staging environment, run the checklist with you, and launch through automated gates with automatic rollback. A 12-month defect warranty follows. The custom code and data are yours, hosted in accounts in your name, and an optional care plan adds monitoring and a monthly health report.
Before any build starts, you have a proposal that says what we will build, what we will not, and what it costs. It lists the included work as itemised lines, names the exclusions in plain words, and carries a ship date that goes into the contract. We publish our build prices on the custom software page, so the number does not come as a surprise.
Payment is 50% when you sign and 50% when your acceptance checklist passes at launch. If the scope changes after sign-off, the change goes through a written change order at $175 an hour, agreed before any work starts. Nothing is billed that you have not approved in writing.
For bigger or less clear jobs, a paid Blueprint produces the written scope, the acceptance checklist, a fixed quote and a clickable prototype. Credited in full if you sign the build within 30 days.
We build where you can see it, in a staging environment, never on your live business. You get a written status every week with the risks named, not an "almost done."
The checklist you signed before we started is run together, in the view your staff and customers will use. The build is accepted when every item passes.
Launch goes through automated gates and has an automatic rollback. If a release fails its post-launch checks, it is reversed instead of left broken.
For 12 months we fix defects in what we built at no charge. If a checklist item cannot pass within 30 days of launch, we refund the final 50% payment.
We do not tell you something is ready until four things are true. It is built and finished, not scheduled or almost done. You have been able to check it in the same view your staff and customers will use. You have the direct link to the live thing, or the exact click path to it. And it is clear who does what next, including when the answer is "nothing is needed from you." A change that only passed on our side has not met that bar.
The process is the protection. These are the steps every release goes through.
Every release passes automated checks first. Checks reduce risk; they can't promise no bugs.
A release that fails its post-launch checks is rolled back to the last working version.
Changes are built and checked in a staging copy before they touch the system your customers use.
During a build you get a plain written update each week, with named risks and the next step.
Scope changes are priced and agreed in writing before work starts, at $175 an hour.
Anything risky gets a runbook: every task has an owner, success criteria and a rollback step.
When a build replaces a tool your business depends on, we do not switch it off on a Friday and hope. The old tool keeps running while the new one proves itself in parallel. We return work to the new system in stages, with written rollback triggers agreed before we begin. We scope the parallel period to how critical the tool is, and we do not promise a switchover date until that period is scoped. Where a tool is worth keeping, we build on top of it instead of replacing it.
On a care plan, we monitor your critical paths, such as booking, payments and forms, and alerts come to us. We cannot promise to catch every problem before you do. What we do promise is a record: a plain-language monthly health report covering uptime, response time, certificate expiry, incidents and what we fixed, and a written root-cause note after any significant incident that says what broke, how we fixed it and what we changed to prevent it. We keep the security basics an insurer asks about. Secrets are kept in a managed vault, never in code or chat, and email authentication (SPF, DKIM and DMARC) is set up on your sending domain. Backups include a quarterly restore test. These are controls we have in place. They are not an audit or a certification.
For systems on a care plan, our target is 99.5% availability each month, measured by an external monitor. In a 30-day month that allows about 3 hours and 36 minutes of downtime. It is a target, not a guarantee. If a month falls below it, you receive a pro-rated service credit, capped at one month of your care fee, and that credit is the only remedy. The target does not cover outages at providers we depend on, such as your payment processor, your domain registrar or the cloud network itself, or maintenance you ask us to schedule.
Handover is part of the build, not an extra.
The custom code lives in a repository in your name, so another developer can take it over.
Architecture overview, deployment steps and an access list, so nothing depends on our memory.
The system runs in hosting and service accounts in your name. Nothing is locked inside ours.
Hosting, monitoring, backups and a monthly health report from $149/mo plus tax, billed apart.
A care plan is month to month. Cancelling takes one step and costs nothing.
Take your data with you anytime. Platform components we license stay ours, named in the contract.
We use AI coding tools; every change is reviewed by an engineer before it ships. Your contract has a short Tools and AI section that names the tools we use, what of your data may go into them, and confirms that you own what we build for you. If you want to know how a specific piece was built, ask and we will tell you.
Short answers about process, reliability and leaving.
Done means every item on the acceptance checklist you signed passes when we run it with you, in the view your staff and customers will use. Until then we will not call it ready. After launch, the 12-month defect warranty covers anything that does not work as the signed scope describes.
No. We publish a target of 99.5% availability each month for systems on a care plan, measured by an external monitor, with a pro-rated service credit as the only remedy if a month falls short. It is a target, not a guarantee, and it does not cover outages at third-party providers.
You keep the code, the data and the runbook, because they are in accounts in your name. A care plan cancels in one step and you can export all of your data. Platform components we license to you stay ours, and your contract names them.
In stages. The old tool keeps running while the new one proves itself in parallel, and we return work to the new system step by step with written rollback triggers. Where a tool is worth keeping, we build on top of it instead of replacing it.
Tell us the workflow you want fixed. A free 30-minute scoping call, then a written proposal with a fixed price within 24 hours.