The Roadmap We Build With You
Last updated:
Why we do not write the roadmap first
A two-year technology roadmap written before anything has been built is a work of fiction. It rests on assumptions nobody has tested, and it commits you to a shape before you have learned anything.
So we build one project, watch how it lands, and then plan. The roadmap that comes out of a delivered project is worth more than the one that comes out of a workshop, because it is informed by how your business actually reacted to change.
The three horizons
- Next two quarters — specific. Named projects, scoped, priced, in order. These we will commit to.
- Quarters three and four — directional. Named, sized in bands, sequence provisional.
- Year two — a hypothesis. Themes rather than projects. Written down so the architecture does not close doors, not so anyone plans against it.
Being explicit about the confidence level at each horizon is the whole trick. A roadmap that presents month twenty with the same certainty as month two is misleading everyone, including the person who wrote it.
The sequencing rules
- Data before intelligence. No forecasting or AI until the underlying data is trustworthy.
- Internal before external. Prove it with staff before customers see it.
- Cheap and certain before expensive and speculative. Fund the ambitious phase with the credibility of the boring one.
- Each phase must stand alone. If you stop after any phase, you still have something worth having.
- Never two hard things at once. A new integration and a new data model in the same quarter is how quarters get lost.
The rule we defend hardest is the standalone one. A roadmap where phase two is worthless without phase four is not a plan, it is a hostage arrangement.
What a real roadmap looks like
| Phase | What it does | Why now |
|---|---|---|
| Q1: order capture | Removes daily re-keying | Highest score, fully contained |
| Q2: stock reconciliation | One trustworthy stock number | Needs Q1's clean order data |
| Q3: supplier portal | Suppliers update their own lead times | External — needs internal proof first |
| Q4: demand forecasting | Buying suggestions | Only sensible once Q1–Q3 data exists |
| Year 2: customer portal | Self-serve order status | Directional — depends on Q3 adoption |
Note that the AI-flavoured phase sits fourth, not first. That is not conservatism; it is that a forecast built on the data as it existed in Q1 would have been confidently wrong.
Reviewing it, and being willing to tear it up
We revisit the roadmap quarterly. Roughly a third of the time something changes materially — a phase gets cut because the manual workaround turned out to be fine, or a new one jumps the queue because the business changed.
A roadmap nobody has changed in a year is not a sign of good planning. It is a sign that nobody is reading it.
What it costs and who owns it
The roadmap session is included for clients we have delivered a project for. It is a half-day with the people who run the affected processes, plus a written document afterwards.
It is yours, and it does not commit you to us for any phase in it. Several clients have used it to hire an internal team, which is a perfectly good outcome and occasionally the one we recommend.
Frequently asked questions
Do we have to commit to the whole roadmap?
What if our priorities change completely?
Can you write a roadmap without building anything first?
How far ahead is it worth planning?
Not sure which part of the business to fix first?
A 30-minute call and a process walkthrough is usually enough for us to say where the money is. There is no charge and no follow-up sequence.
Related services
What we build for problems like this one