~ 11 min read

Agile Roadmaps - Why and How Roadmaps can be better

Explore the need for change in agile roadmapping and discover how hypothesis-led roadmaps can transform the agile process.
Image credit: midjourney.com

We feel that roadmapping today can be better. While over the years we as an industry have gotten very much used to Jira and agile, at times it seems we’ve become enslaved by our tools. Particularly the way roadmaps are created and visualized often seems to create tendencies that work against the iterative approach. This post is about discussing better tools, why we need them, and what they might look like.

Challenges with Roadmapping TodaySection titled Challenges with Roadmapping Today

The Cost of MeetingsSection titled The Cost of Meetings

When aligning priorities and estimating items, it is easy to get lost in this process. The cost of these planning meetings is a blind spot for many companies. Instead of going for the quick-win and getting insights, things tend to be planned 10x over, before the first move is made. What we need is more bias for action.

Planning ParalysisSection titled Planning Paralysis

It is hard to say when a plan is finished. For roadmaps, epics are created. Individual tickets are created and groomed. The estimated size of the epic becomes clear and gets approved somewhere in the area around C-level. To get approval for an idea, all the details need to be mapped out, and you might find as many reasons to do a thing, as there are to not do that thing.

Planning SoftwareSection titled Planning Software

The industry has significantly evolved over the years. Collectively, we have learned that making a long plan for a thing doesn’t really work in software. That’s why we’ve turned from Waterfall to Scrum or Agile. Working in iterations, we’re supposed to aim at some result at the end of each sprint. From this result we can learn, reflect and adapt, in order to stay on course. Thinking up a huge plan ahead of time and following it 1:1 does generally not work in software development.

The Concept of EpicsSection titled The Concept of Epics

Epics are one big chunk of work on the roadmap. They can represent a feature, an effort or an initiative. They are time-bound and generally have a concrete goal. Epics are generally prefilled with all necessary tickets, and estimated before the epic is started. This is to get the approval for the idea given the complexity, effort, dependencies and projected result, while also fixing the start and finish. Because epic configurations are approved by leadership, changing the epic after it is started is problematic, even by iterative findings that came up in the sprints. This is how epics tend to stand opposite to the iterative approach.

Tool BlindnessSection titled Tool Blindness

Today, almost every software company has adopted agile methodologies. There is hardly any place that doesn’t use a tool like Jira, ClickUp or Linear to facilitate project planning. While we have gotten pretty used to a planning status quo, and putting things in their right places in our tool of choice, hardly anything ever goes according to plan. We use modern tools and fill out all the colorful boxes, but we get disappointed every time as the deadline has to be moved, as that epic has to be extended that little bit.

Visual BiasSection titled Visual Bias

The way that roadmaps are visualized creates a certain bias for something that we know not to be true in software: Long-term plans don’t work out as expected. Yet looking at the roadmap tools, we set start and end-dates for everything. You can have a team of product managers raging in excitement, because each of them marked an epic as done on the roadmap. Meanwhile none of them have talked to the customer or looked at the numbers. Tools today create a bias to execute things perfectly just as planned, not to learn, reflect and iterate.

A new ideaSection titled A new idea

One CEO we both knew always used to say:

Verify your hypothesis as early as possible.

It seemed to be a core learning of his after many hard years in the startup world. We decided to take this quote as the foundation to a new way of planning. What C-Level cares about at its core is not that colorful bubbles stay exactly where they were originally created in Jira. What C-Level cares about is to validate hypotheses. To learn critical information about the customer and the future of the business.

While it is true that probably behind every epic are one or more core ideas, epics themselves are not hypotheses. Epics are planned items of work on a high level, but including all individual tasks as well, bundled up with a start and estimated end. Following a huge plan in software development is something that historically must end in disappointment, and which lacks the iterative components. In contrast, trying to verify a hypothesis as early as possible seems to be close to the heart of the iterative approach.

Hypothesis-led RoadmapsSection titled Hypothesis-led Roadmaps

If we can reduce roadmaps and their epics to their essence, which are the hypotheses, we can start thinking of a new way of roadmapping, and new tools.

When roadmaps are just a set huge plan for people to execute, not a lot of additional visual information is needed. As hypotheses are thinner than epics, as they are more condensed, we need additional information to help us out here. This is what we need from a truly iterative roadmap:

  • Central KPIs with their start and end date visually in the roadmap
  • Hypotheses as the center piece
  • Big releases featured (ideally they affect the KPIs)
  • Iterations as visual features. Based on a release you may receive customer data or feedback and make a decision. This needs to be on the roadmap to create a new bias. An iterative roadmap needs to be full of circles, instead of long bars!
  • Sprints near the bottom to again highlight the visual nature. A sprint is a planned iteration and thereby connected to the releases.

Sketch of the proposed agile roadmap layout.

Imagine a roadmap that evolves and adapts, mirroring the dynamic nature of agile workflows. That’s the essence of a truly iterative roadmap. That’s why we dub this concept as “agile roadmaps”. It is more in line with what we as engineers know to be true in software engineering, while it also caters to the most important need of C-Level: Validating hypotheses.

At the same time it addresses all mentioned problems from earlier:

  • The Cost of Meetings
    • Extensive planning and approval of long-winded plans is avoided as initiatives are condensed to their hypothesis. Deadlines can exist, but generally hypotheses should be confirmed as soon as possible. Weekly status meetings on selfmade deadlines can be skipped, and we can instead focus on the metrics and the next iteration.
  • Planning Paralysis
    • Reducing the amount of planning that goes into approval of a hypothesis, reduces the amount of planning needed. The only thing that’s needed is an idea for the first sprint, and from there we can go iteratively.
  • Planning Software
    • Moving software planning closer to the iterations and learnings based on user data puts software development in the place where it can work. Delivering iteratively is the way to go in software, while trying to hit the target a bit closer every time.
  • The Concept of Epics
    • Focusing on hypotheses changes roadmapping fundamentally. Instead of pre-planning everything, we expose the core ideas of our plans to the brain hive of the entire company. We focus closer on what truly matters to leadership: To confirm or to disconfirm an open hypothesis.
  • Tool Blindness
    • Celebrating “marking something off the list”, without talking to the customer or looking at the KPIs, is hardly possible with such a visualization. If no iteration was done or the KPI didn’t move, the real impact of the work clearly shows.
  • Visual Bias
    • Agile roadmaps create a bias towards sprint-based releases, and reminders to create true iterations chasing a measured goal. Today’s roadmaps reward product managers for finishing an epic. Roadmaps of the future will reward product managers for iterating.

The ComponentsSection titled The Components

  • Hypotheses: Can be either open / confirmed / disconfirmed / dropped or inactive
  • Iterations: While originally every sprint was supposed to be an iteration, in reality this often does not lign up. Iterations in agile roadmaps require three things:
    • Customer data / feedback
    • A release this feedback relates to
    • An insight, decision or learning

A real brain teaser will be the fact that we have to stop thinking exclusively in deadlines. Yes, deadlines can and do exist, but deadlining everything that moves has honestly proven inefficient. Putting a deadline on something that gets then moved again, again and again is not only stressful, dull and repetitive, it also drives up the hidden cost of meetings and frustration.

With a hypothesis-led roadmap, once a hypothesis is open, it should be confirmed as soon as possible. Iterations will be created iteratively, as we learn more about the problem and take action. Hypotheses that are open clearly have this indication that they need to be verified asap, but they do not have a deadline on the roadmap and simply stretch to the current length of the display. It is the job of the project manager to verify such a hypothesis with the insights that are gained from iterations.

More than a SketchSection titled More than a Sketch

We decided only having it as a sketch was not enough, and we created an initial prototype for this idea. The plan is to integrate this with all major Agile planning systems. A new visualization in our mind can make the real difference in the daily lifes of many people.

Experience the future of agile roadmapping firsthand. Explore our demo now: Agile Roadmaps Prototype

If you would like to see this prototype further build out, please let us know either by starring the repository or by reaching out in some other way. We will base our future actions on this project purely on the feedback we receive for this blogpost.

This prototype shows hypotheses and iterations as the center pieces. Tickets can still be connected to epics if needed, and in the demo you can toggle them on in the settings on the top right.

FAQsSection titled FAQs

What is the Line of Control?Section titled What is the Line of Control?

Condensing the epics into their hypothesis and ligning it up next to KPIs, releases and iterations, also gives a new finer level of delegating control. Before with epics, everything is one unit, from the individual tasks to timing to dependencies and scope. Epics get approved as one packed bundle somewhere near or by C-Level.

With a hypothesis-led roadmap, leadership gets more options where they want to draw the line. Do they want to only set the KPIs? KPIs and hypothesis? KPIs, hypothesis and sprint goals? Kpis, hypothesis, sprint goals and iterations? The answer to this question will vary from leader to leader. More delegation could in the end lead to more empowerment, more drive by the employees, and more time freed up from all of that.

Why Reinvent Roadmaps?Section titled Why Reinvent Roadmaps?

Isn’t it the case that iterations already exists today in the form of sprints? Epics and hypotheses are also not too far apart. Releases can be displayed on a Jira roadmap, etc. All this is true, but still we believe a new way to create roadmaps is needed.

With the right mindset and with aligned priorities, you can work truly agile today, even in tools like Jira.

The issue is that current roadmap tools create a certain tendency away from what we know to be best for software engineering: The iterative approach. That’s why we need a new way to look at roadmaps and a new tooling. Whether this is our current approach or somebody else creates a smart plugin within these tools remains to be seen. But in our minds, something needs to change. Software project planning needs to be better.

Where do I start?Section titled Where do I start?

Given that a few of the ideas in this post convinced you to try and start moving closer to a hypothesis-led agile roadmap, what could be done? These are the steps you could take if you’re not doing so already:

  • Clearly write down the hypotheses behind your epics.
  • Talk to your users and use this feedback as input for your next sprint planning.
  • Make sure you have a live dashboard somewhere with your KPIs.
  • Avoid ‘big bang’ releases at the end of epics. Instead, focus on rolling out partial solutions iteratively to end-users.
  • At the end of an epic, take the time to analyze the numbers and the real impact.

Rounding UpSection titled Rounding Up

We hope to have hit a nerve here with our take on better roadmaps. We truly believe in our daily lives that roadmaps can and must be better. A dusty roadmap that was planned a long time ago in the end can only lead to dusty meetings. Let’s change this!

Any Comments? Join the Discussion on Twitter: