What a first phase looks like
One core workflow, its data model, validation, a basic report and an admin view. Enough to replace the most painful part of the old process.
Overview › Software development
We build business software around information: capturing it cleanly, combining it reliably and presenting it to people who act on it.
Small web applications that replace a process held together by email, shared drives and memory. Typical examples: a request and approval flow, a register of assets or contracts, a scheduling board, a case tracker. They are built for a known group of users with clear permissions, and they tend to pay back quickly because the people using them are the people who suffered the old way.
Inputs: a walk-through of the current process, examples of the forms and files involved, and a sponsor who can settle disagreements. Outputs: a working tool, user guide and administration notes.
Most reporting problems start at the point of entry. A free-text field where a list was needed, a date typed three different ways, a required value that is optional in practice. We design forms and imports that check information as it arrives, explain errors in plain language and make the right answer the easy one.
We favour a small number of views, each tied to a decision someone makes regularly. For each figure we record its definition and source, so a manager asking "where does this number come from?" can find out without calling a developer. Where an established reporting product suits you better than custom software, we will say so and help configure it.
Many organisations run a handful of good systems that do not talk to each other. Where those systems provide documented interfaces or exports, we build reliable links with logging, error handling and a clear owner for each piece of information. We do not claim existing integrations with any named platform; each connection is assessed for your case.
Moving from a spreadsheet to a system is mostly about trust. We document the existing rules, import historical data with a reconciliation report, run both side by side for an agreed period and switch only when the figures agree. Anything that cannot be migrated cleanly is listed rather than quietly dropped.
Code lives in a repository your organisation owns. Hosting is set up on accounts in your name with established providers. Documentation covers how to run, update and restore the system. Ongoing support can be agreed with us, but it is never a condition of owning what you paid for.
One core workflow, its data model, validation, a basic report and an admin view. Enough to replace the most painful part of the old process.
Secondary reports, edge-case automation and integrations that are nice rather than necessary.
Calculations, permissions, imports and anything that would be expensive to get wrong.