You will recognise some of this
- The package you bought does 70% of the job and the other 30% is done in Excel
- Two systems hold the same information and neither is trusted
- A person exists mainly to copy data from one screen into another
- Every report is someone exporting a CSV and rebuilding it by hand
- The developer who understands it has left, or is about to
- You have been quoted for a rebuild and cannot tell whether the number is real
What makes this different from an agency build?
An agency delivers a repository and an invoice. The risk of running the thing — the migrations, the 2am alert, the schema change that breaks a report — transfers to you on delivery day. We keep it. We already run a production SaaS with paying customers, so operating software is not a service line we bolted on; it is what we do daily.
How long does a first version take?
For a focused system, eight to twelve weeks to something real people use — not a demo. We work in short cycles against a live environment, so you see working software every week rather than a status report. Anyone promising a full ERP in six weeks is selling you a prototype and calling it a platform.
What happens after launch?
Monitoring, alerting, backups, database migrations and a support channel with a human on it. We instrument the paths that carry money and alert on those first, because a green dashboard next to a broken checkout is worse than no dashboard.
How it runs
- 1
Discovery
Two weeks
We sit inside the operation rather than reading a requirements document — including the workarounds nobody wrote down, which is where the real requirements live. You get a scope, a shape and a number. If the answer is that you should buy something off the shelf instead, you get that too.
- 2
Build
Eight to twelve weeks to a first real version
Short cycles against a live environment. Working software every week that you can click, not a status report. Scope moves in the open; nothing is discovered in month three that was known in month one.
- 3
Cutover
Planned, rehearsed, boring
Parallel running against real data before anything is switched, and a rollback path that has been exercised rather than documented.
- 4
Operate
Ongoing, monthly
Monitoring, alerting, migrations, upgrades and a support channel with a human on it. This is the part most vendors leave out and the reason the other three hold up.
When not to hire us for this
- A fixed-price build from a specification nobody has tested against reality
- Rescuing a project where the previous team is still mid-flight — one of us has to own it
- A team augmentation contract; we take responsibility for outcomes, not for filling seats
- Anything where a package like Odoo would do the job — we will tell you, and we implement those too
What this rests on
The clearest evidence is not a testimonial. It is that we run our own multi-tenant SaaS in production, with paying customers, on the same pipeline and monitoring your project would get.
See what that involvesQuestions
- Do we own the code?
- Yes. Full source ownership and repository access from day one, regardless of whether we continue to operate it for you.
- Can you work with our existing systems?
- Usually. Most projects involve integrating with something that already exists — an accounting package, a supplier feed, a legacy database. That integration work is normal, and we scope it during discovery rather than discovering it in month three.