Cloud, hosting & cost control

What we do

Cloud, hosting & cost control covers four jobs: deciding where your systems should run, checking what you already pay for, moving them safely when they need to move, and running the systems we build once they are live.

Architecture. AWS is where most of our client builds run. Six of them, from a native iOS and Android app to a legal-advice portal used across 45+ countries, were designed, hosted and maintained by us on AWS. We also work with Microsoft Azure, Microsoft 365, Cloudflare and DigitalOcean, and run our own systems on several of them. We size for the traffic you have, with a written path to the traffic you expect. We have recommended Kubernetes, a system for running many containers at once, in plans for start-ups expecting fast growth; for most small systems a simpler setup is cheaper and easier to look after.

Cost review. A cost review lists what is running, what each part costs a month and what it does. Anything that does nothing gets switched off. On our own estate, that meant retiring an unused Azure content delivery network (CDN) endpoint that was still billing every month. We do not quote typical savings, because every estate is different; the review gives you your own figure.

Migration. Moving a system to a new server or provider follows a written plan: a test copy first, a rollback ready, and a switch-over date agreed with the people who depend on the system. Setups we design avoid features that lock you to one provider unless they earn their place, so a later move stays possible.

Hosting. We host, maintain and support the software we build, on whichever provider fits it. When something needs fixing, you deal with the person who wrote the code. We do not sell hosting as a standalone package for systems we did not build.

What you get

• An architecture with the reasoning written down
• A cost review: what runs, what it costs and what to cut
• A migration plan with a tested rollback
• Hosting, updates and support for what we build
• A setup that can move to another provider

How it runs

1. Inventory. We list every service, server and subscription, and what each costs. Gate: you confirm the list is complete.
2. Recommend. A written set of changes, each with its cost and risk. Gate: you choose which to make.
3. Change. Each change is made with a rollback ready and tested before the old setup is switched off. Gate: you confirm it works before we remove anything.
4. Run and review. For systems we host, we maintain them and review the bill with you. Gate: anything that has stopped earning its cost is raised, not left running.

Who it is not for

• Organisations that need round-the-clock operations with a contractual response time measured in minutes. We are deliberately small; a managed-service provider with an on-call team is the right fit.
• Anyone looking for hosting on its own for a system we did not build.
• Enterprise cloud financial management (FinOps) programmes across dozens of accounts.