Roadmapping: The key to getting your digital house in order

Here's what a best-practice roadmapping approach looks like, plus a structure to set your course

Does your to do list look something like this?

  • Clean up existing legacy IT mess
  • Get rid of duplicated systems, data islands, inflexible infrastructure
  • Move to the Cloud
  • Leverage DevOps
  • Support Digital Transformation

What I've seen in many engagements with enterprises is that the key to all of these is a strong roadmapping capability.

What is roadmapping? Roadmapping involves identifying your current state architecture, and also the future state your business is aiming to achieve together with transitions between the two. The most effective roadmaps consider the multiple ways of completing the journey from one to the other.

Holding all of these permutations in your head is, of course, impossible. And you'll quickly outgrow basic business tools such as Excel and PowerPoint.  So how do you gather data to weigh these multiple options in order to navigate with confidence towards a desired business endpoint?

Using an enterprise architecture toolset which maintains a discrete collection of architectures (i.e. current, transition, target) and is capable of simulations and metrics-based analysis against them.  The toolset should let you set goals and select analytics which will allow you to quantitatively assess potential options.  You might want to consider using a range of metrics: e.g. Financial, Experience, Risk, Compliance, Technical and Environmental. It's also worth testing a range of simulation options such as equational, structural, activity-based, discrete-event and Monte-Carlo techniques.

Setting the course

A useful way to provide structure to your planning is to work through the questions below:

  • What elements should be on the roadmap?
  • What properties or lifecycle stages should we use?
  • How fine-grained should the roadmap be?
  • What is the direction of / are there dependencies between the roadmaps?
  • What visualizations will work best? Gantt chart, Milestone chart, Sunset chart?

I find a useful starting point is to consider the 4 basic types of EA roadmaps:

Tagging with recommendation and heatmaps

This style uses a metric like Retain-Redesign-Refresh-Retire or Tolerate-Invest-Migrate-Eliminate to color, or "heatmap," your business capability maps or technology landscapes to show what is changing. This is simply about making a recommendation of what should happen, not how or when.

Lifecycle properties and Gantt charts

These roadmaps explore in-flight and proposed change initiatives as the how and use actual date/time properties for when changes are going to happen. They require a deeper understanding of PPM (Project Portfolio Management) approaches and the transformational side of the business. However, they are limited when dealing with interdependencies.

The "ripple-effect" or "dynamic" roadmap

Dynamic roadmaps exploit the hidden dependencies or n-degrees-of-separation structure inherent in every business. You can watch the cascading effect that a given change might have on every other area of the business. Say a project is proposing to upgrade a service which requires an application that is hosted in a data center being decommissioned by another project. What will happen?

The "multiple architectures" roadmap

Expert roadmappers will find themselves asking two critical questions: is our target state possible in and of itself? And is it actually possible to get to our desired target state from where we are today? We call this feasibility analysis and reachability analysis. You can model multiple states as separate architectures, comparing planned states with actual states. The planned states can be project architectures, transition architectures, and ultimately target architectures. Comparing these to a baseline or current states will help determine their viability.

The art of the possible

This style of future state modeling and optimization will allow you to identify improvements, such as which legacy systems can be safely retired (and the most prudent way to do this), a plan for moving data safely to the cloud, or which areas need to be upgraded to improve customer experience.

And most importantly, it will also give some real metrics-based certainty to business planning and digital transformation initiatives.

Copyright © 2015 IDG Communications, Inc.

How to choose a low-code development platform