DelloiteConsultingAsia Talk to us

PerspectiveTechnology

What a technology roadmap should leave out

A roadmap earns trust by what it declines to promise. A way to decide what stays off the list.

DelloiteConsultingAsia Technology practiceNovember 20262 minute read

Neatly bundled cables in a rack

A roadmap is a set of commitments

Whatever its authors intend, a roadmap is read as a list of promises. Every item on it will be chased by someone. A roadmap with forty items, built by collecting every request, tells the business nothing about what matters and tells the technology team that everything does.

The useful work is in deciding what does not go on it.

Five things to leave out

  1. Dates you cannot defend. Beyond the next two or three quarters, show the order of the work and what depends on what, not months. A precise date for something two years away is still a guess, and it will be quoted back to you.
  2. Projects with no owner in the business. If nobody outside technology will say what the change is for, and release their own people to work on it, the project is not ready for the roadmap.
  3. Technology names in place of outcomes. Moving to the cloud or adopting AI describes a means. State the change in the business, such as orders confirmed the same day or one customer record across the group, and let the means follow from it.
  4. Keeping the lights on. Upgrades, patching and licence renewals are necessary and they consume capacity, so show the capacity they take. They are not choices, and they do not belong in a list of choices.
  5. Anything the team cannot staff. Add up the people each item needs against the people there are. What does not fit goes on a separate list.

Publish the list of what is not being done

This is the most useful part of a roadmap and the least comfortable. A visible list of what has been left out lets a business unit head see the decision instead of discovering it months later.

It also gives them a fair way to argue. To add something, they name what comes off. The conversation moves from whether a request is worthwhile, which it nearly always is, to whether it is worth more than something already on the list.

Three questions for every item

  • Who asked for it? A named person in the business who will own the result.
  • What will be different when it is done? Described in terms the business would use.
  • What are we not doing in order to do it? If the answer is nothing, the capacity sums are wrong.

Items that cannot answer all three wait. Run the same test every quarter, on the items already on the roadmap as well as the new ones.

Working on something like this?

Talk to the technology team