Data and internal tools
Build the tool your team has been waiting three years for
A tool cut to fit how you work, built in weeks rather than quarters, with the data behind it under control.
Weeks
to build, where a conventional project counts in quarters
Monthly
automatic check that the data pipelines are still running
The problem
Every company has a tool that nobody sells: the one that would do exactly what your business does. Since it does not exist, its place is taken by a shared workbook with forty tabs that one person knows how to run and nobody dares change.
Until recently, having that tool built cost too much to justify. That is no longer true. The cost of building an internal tool has collapsed, to the point where building it often comes in cheaper than paying a subscription for software that does not quite fit.
Then comes the half that everyone forgets: the data that feeds the tool. Importing external reference sets, checking them, loading them into a warehouse and then into the application databases, and confirming every month that the chain still runs. It is less impressive than the tool itself, and it decides whether anyone is still using it in two years.
This concerns you if
- Your core process lives in a workbook nobody dares touch
- You pay a subscription for software you use a fifth of
- One person is indispensable because only they can run a given job
- Your reference data is updated by hand, when somebody remembers
- Your teams have waited months for a change nobody has time to make
The builds
What gets built
Les délais indiqués sont ceux d'une mise en production réelle, pas d'une démonstration. On n'en lance qu'un à la fois.
- 01 4 to 8 weeks
A tool built for how you actually work
The application your teams use every day, built with them in a few weeks, hosted wherever you want it. We start from what people actually do, not from a specification document.
- 02 3 to 6 weeks
Checked imports of reference data
Jobs that import large external or internal reference sets, check them, version them and load them into your warehouse and then into your application databases.
- 03 4 to 8 weeks
A data warehouse at the right scale
A warehouse sized for a company your size, not for a multinational: enough to join your sources together and produce figures you can trust, without an eighteen month programme.
- 04 ongoing
Monitoring and handover
Monthly supervision of the jobs, alerts when something looks wrong, operating documentation, and a gradual handover to the teams that want to run it themselves.
Tools Google CloudSupabaseBigQueryReactPython
Data privacy
On this lever specifically
This is the area where confidentiality is easiest to hold, because the tool and its data stay with you or on infrastructure you choose. Where a model is involved, it sees only what it needs, and the architecture separates what may leave your walls from what must not.
The full framework, level by levelData and internal tools
Common questions
What happens to the tool if we stop working together?
It is yours. The code sits in a repository you own, hosting can be on your own account from the first day, and the operating documentation is part of the deliverable. That is a starting condition, not an exit clause.
Will a tool built quickly age badly?
That is the real risk, and it does not come from the speed of building. An internal tool dies because it has no tests, no monitoring and nobody to take it over. Those three things are part of the build, or there is no point starting.
When is buying off the shelf the better answer?
When your need is standard and regulated, such as payroll or accounting, or when a product exists that genuinely covers how you work. Building makes sense where your way of working is your advantage and no vendor has modelled it.
The other levers
-
Marketing and acquisition
A steady flow of inbound enquiries, built once and kept running in a few hours a month.
-
Sales and prospecting
The write-up, the quote and the follow-up are ready before you leave the meeting. You read them and you send.
-
Customer service and support
The real causes of customer frustration, drawn out of your own conversations, and the tools to help your team answer better and faster.
-
Admin and finance
Data moves between your systems without anyone retyping it, and discrepancies reach you before the money leaves.
Next step
Where would you start?
Ten days to map where your teams lose time, put a figure on each lever and name the first system to build. If AI is not the right answer, I will tell you so.
Diagnostic
Étape 110 days
€9,500 excl. VAT · fixed fee, one area of the business
- A map of where the time actually goes in the area audited
- Each lever costed in hours freed, ranked by effort and by gain
- The data note: what leaves your systems, and where it is processed
- A three-month plan, naming the first system to build