We often receive RFPs (Request for Proposals) that demand a firm committed five year product roadmap. Similarly we are often criticized for not having such a “golden” roadmap when other competing products have. Having now worked in an Agile world for some time now, these requests seem stranger and stranger.
The quibble here isn’t with having a plan, it’s with the inflexibility these requests imply. That a company needs to set its course for five years and then any change in that plan is somehow a failure to deliver. That as knowledge and circumstances change that you need to stick to the plan and not adapt to the new situation.
Products are now introduced in “Internet” time. This means they are updated far more frequently (sometimes several times a day). All companies are looking to be “disruptive” and to “redefine” their market. Under these fast moving and fast changing conditions does it make sense to have a fixed long term roadmap?
On the other hand a product needs direction. A product needs long term thinking. You need to decide when to do something quick and dirty versus laying more groundwork and infrastructure to support future features. Stakeholders need to have an idea where a product is developing and what might be coming down the road.
There are quite a few types of roadmaps, there are technology roadmaps, feature roadmaps, release roadmaps, stop list roadmaps, marketing, strategy and many others. This articles generally applies to any of these.
Waste Not, Want Not
One of the key tenants of Agile Development is to reduce and if possible eliminate waste. Waste is any extra work that is being performed by team mates that doesn’t directly add value during the agile sprints. One main source of waste is doing too much and too detailed estimating. If you want a team to commit to estimates then they have to spend a lot of time working through those estimates so that they have the necessary precision. However when you do this, this work is often just waste, since then the work isn’t done due to changing priorities or another team does the work and insists on repeating the process or the project is postponed and when its resumed things have changed.
Roadmaps tend to generate a lot of wasted work. Once a company wants a roadmap that everyone is committed to, then far too much time will have been spent working on the estimates. The trick here is to be willing to accept inaccurate estimates. Many studies have shown that inaccurate WAG (Wild-Ass Guesses) type estimates aren’t really any less accurate than carefully constructed ones. All you need to know for building a roadmap is the order of magnitude of an item, not the details.
Detailed estimates are only done when the stories are going to be performed by the agile team. This work usually happens as a part of backlog grooming in the sprint before the work is actually going to be done. This then ensures that the stories are properly broken down and that the work can fit into one sprint.
Accept the Roadmap as a Guideline
The best way to think of a roadmap is as a guideline for current thinking. It is a mechanism to elicit feedback which can then be used to produce a better roadmap. Publishing a roadmap as a “fait accompli” doesn’t serve nearly as well as using a roadmap as a starting point for a conversation.
Often getting customer feedback on direction without providing any context or ideas is quite difficult because customers don’t spend their time thinking about how you need to develop your product. With a roadmap they can see how your product will fit in (or not) with their future business directions. Then they can provide useful feedback on what will be useful, what will be irrelevant and what will actually be harmful.
Keep in mind that conversations are two way things and the only way to be successful is to incorporate the feedback received and to show that the time spent by the customer talking to you is worthwhile. Corporations that can incorporate and synthesize the feedback from hundreds of customers in an effective manner tend to be the companies that really shine.
Accept that Agile Works
A lot of times the push for a fixed roadmap is a result of the organization outside of R&D not being comfortable with the Agile idea of working on the most important story all the time. They liked the old days where a giant requirements document was produced and upper level management reviewed this and then felt comfortable that they could let R&D go off for a year or two to work on this without paying any more attention.
Generally it’s proven out that Agile is much more efficient and produces better products that meet customers’ needs much better than the old waterfall execute the requirements approach. But if upper management wants to know what R&D is doing they have to pay attention since things are fluid and always changing. This can be hard to accept, but now its being found that Agile can be applied to other parts of the organization and rather than older parts of the company dragging down the Agile parts, now all departments are going to Agile and its working very well for modern companies. In fact many people now believe that if a company doesn’t make this transition then it will become less and less competitive. The sad part is that Agile produces far better artifacts showing the progress of a project, you just need to learn how to use the tracking software to see them (another wastage is producing specialized reports just for upper management consumption).
In the end, the goal is to be as customer connected as possible. Always working on the item in the product backlog with the highest value for the customer. This is now a proven principle for success. Dictating to customers what is good for them will just alienate your customers and send them elsewhere.
Creating a general roadmap that is used to get customer feedback and buy-in is just one tool of many to being a better customer connected company. And again the key secret ingredient is always adapting to change and not becoming fixed in your direction.
Of course when talking to customers and often other stakeholders, they will start out with how they need everything yesterday. But you have to steer the conversation to choosing priorities and not being brow-beaten into accepting that any estimates need to be shorter.
Roadmaps are great tools for having a conversation with stakeholders on the direction of a product. You just have to be careful not to fall into the trap that the roadmap is somehow a commitment that can never change. If done properly it can serve quite a few goals that are fully compatible with an Agile methodology.
If you are presenting a roadmap at a conference or WebEx, always prefix the presentation with that this is our current thinking and that we are always looking for feedback and ways to make the roadmap better.