Method
How we work
Clear stages, identified deliverables at each one, and a usable product well before the end. At any point you know what has been done and what remains.
- First usable batch
- 6 to 12 weeks
Overview
The whole path at a glance
- Understand (1 to 2 weeks) — A map of the processes concerned, A list of pain points ranked by real cost, An inventory of existing systems and data
- Frame (1 to 2 weeks) — Scope of batch 1 and what is explicitly out of scope, Target technical architecture, Success criteria expressed in figures, Estimate and schedule
- Design (2 to 4 weeks) — A clickable prototype, A validated data model, Business rules written down
- Build (Two-week iterations) — An acceptance environment you can reach, Automated tests on the critical rules, Systematic code review
- Deploy (1 to 2 weeks) — Verified data migration, Training sessions by role, Switchover and rollback procedure, Operations documentation
- Evolve (Continuous) — Monitoring and alerts, Support commitments, A quarterly roadmap
Stage by stage
Every stage, one by one
- 1 to 2 weeks
Understand
We start from your actual operations, not from a theoretical specification. Field interviews, process observation, an inventory of the tools already in place and of what genuinely gets in the way.
Deliverables
- A map of the processes concerned
- A list of pain points ranked by real cost
- An inventory of existing systems and data
- 1 to 2 weeks
Frame
We turn the findings into a scope: what the first batch covers, what it deliberately leaves out, and what it will cost. You stay free to continue or stop there.
Deliverables
- Scope of batch 1 and what is explicitly out of scope
- Target technical architecture
- Success criteria expressed in figures
- Estimate and schedule
- 2 to 4 weeks
Design
Data model, screens, permissions, integrations. The database schema comes before the framework: it is the part that costs the most to correct afterwards.
Deliverables
- A clickable prototype
- A validated data model
- Business rules written down
- Two-week iterations
Build
Delivery in usable increments. You see and use the product throughout, rather than discovering it at the end.
Deliverables
- An acceptance environment you can reach
- Automated tests on the critical rules
- Systematic code review
- 1 to 2 weeks
Deploy
Data migration, role-based training, production start-up under supervision. The switchover is prepared and reversible.
Deliverables
- Verified data migration
- Training sessions by role
- Switchover and rollback procedure
- Operations documentation
- Continuous
Evolve
Corrective maintenance, security updates, and new functionality prioritised with you as your needs change.
Deliverables
- Monitoring and alerts
- Support commitments
- A quarterly roadmap
What guides the method
The rules we do not negotiate
The need before the technology
We spend time with the teams who do the work before proposing anything. A process that needed simplifying should not be automated.
Batches that stand alone
Each batch delivers value independently of the next. You can stop after the first one having lost nothing.
The data model first
A well-formed relational model outlives several changes of framework. It is also the part that costs most to fix afterwards.
Security from the design stage
Role-based permissions, logging, server-side validation. Not an option added at the end, when it is too late to do it properly.
Africa Tech Services
A project to frame?
Framing is costed separately: you stay free to continue or not.