Most transformation roadmaps are a list of projects arranged by when somebody would like them to finish. They survive until the first dependency bites, usually in week nine, and are then quietly replaced by whatever is actually happening.
What a roadmap is not
It is not a Gantt chart with swimlanes. It is not a set of horizons labelled now, next and later. It is not a slide. Those are all presentations of a roadmap, and an organisation can produce every one of them without having done the work underneath.
A roadmap is the sequenced answer to a gap analysis. If there is no gap analysis behind it, the arrows are guesses drawn confidently.
Start from the gap, not from the wish list
The input to a roadmap is not what the executive team wants. It is two things: which dimensions of the organisation are actually live for this change - out of the 42 in our register - and, for each live one, the distance between the current maturity level and the target level.
That gives every line on the roadmap a reason to exist that can be stated in one sentence: this dimension is at level 2, it has to be at level 3 for the benefit to be claimable, and these are the practices that carry it. A line that cannot be stated that way is on the roadmap because somebody asked for it.
Sequence by dependency, not by appetite
The natural instinct is to start with the dimension that has the most energy behind it. The correct order is the one that respects dependency. Data quality before analytics. Decision rights before process redesign. Capability before the system that assumes it. Supplier contracts before the operating model that depends on them.
Almost every expensive transformation failure is a sequencing failure that was visible on day one. The technology went first because it was fundable and legible, and the four dimensions it depended on went second, which is to say they went never.
Four things on every line
An owner who is not the programme. Named, in the business, with the benefit in their own numbers. If the only owner is the transformation director, the line has no owner after the programme closes.
A date that is a commitment, not a placeholder. Quarters are acceptable for month 18 onwards. For the first two quarters a date is a date.
A cost, including the internal cost. The consultancy fee is the visible part and usually the smaller one. The line should carry the days of internal capacity it consumes, because that is the number that collides with change saturation.
The evidence that will close it. Decided at the start, not at the end. What will be true, measurably, that is not true now. This single discipline kills more decorative workstreams than any governance process.
Set the target level, not the top level
Maturity models tempt everyone toward level 5 in everything, which is both unaffordable and wrong. Most dimensions have a right level for a given organisation at a given moment, and it is often 3. Deciding the target deliberately - and writing down why it is not higher - is what turns a maturity assessment into a plan with a finite cost.
It also gives the roadmap an ending. Roadmaps without target levels do not finish; they get renamed.
Three to thirty-six months, and no further
Beyond three years a transformation roadmap is fiction, because the operating environment that justified it will have changed. Under three months it is a project plan. The useful horizon is three to thirty-six months, with the first two quarters at task level, the next two at workstream level, and the remainder at dimension level with target maturity and nothing else.
Anything more detailed at month 30 will be wrong and will be used against you.
How to tell a real roadmap from a slide
Four questions. Which dimensions did you rule out, and why - a roadmap with no named exclusions has not been through an analysis. What is the target maturity level of each live dimension, and why not one higher. Which two lines are on the critical path, and what happens to everything downstream if either slips. And who, by name and outside the programme, carries each benefit.
A roadmap that answers those survives contact. One that does not will be replaced by events, and the replacement will not be written down.
Our one-week engagement ends in exactly this artefact - a costed, owned, dated plan over three to thirty-six months, built on the dimension shortlist and the maturity gaps beneath it. Readiness sets the sequence; the method sets the content; the engagement page sets out what you get. We run it in Ireland, the United Kingdom, the United States and Australia.