New Product Development with Lean Startup and Scrum
The Agile Zone is brought to you in partnership with Hewlett Packard Enterprise. Discover how HP Agile enterprise solutions can help you achieve high predictability and quality in your development processes by knowing the status of your projects at any point in time.
Discovering Lean Startup was inspiring for me: I felt I had found an approach that could complement Scrum nicely. Since then I have been experimenting with combining the two approaches in my own new product development work as well as helping my clients to do so. This post shares my experiences and insights. It maps out a high-level process for creating new products within existing businesses focusing on product management practices and tools.
From Vision to Launch
Lean Startup encourages us to first investigate if there is a need worthwhile serving before we worry about the details of the new product. The former is called “Problem Validation”, the latter “Solution Validation”. While traditional approaches also suggest carrying out market research and analysis before we engage in product planning and definition, lean approaches increase the speed at which we operate. This allows us to fail and learn faster, to adapt our product strategy and tactics quickly, and to hopefully launch the right product with the right features sooner.
The following diagram illustrates the process I apply to create new products:
The new product development process depicted above starts with a vision. It then uses a series of cycles or iterations to transform the vision into a product ready for launch. The focus of the early iterations is to understand what problem the product solves, who the customers and users are, and how the product benefits the company creating it.
I find it important that the product owner leads this effort and carries out the necessary work together with a cross-functional team. While the team composition depends on the product, having a marketer, sales representative, service representative, a UX designer, and a developer on board is usually helpful. Product owner and team capture their assumptions about the market by using a tool like the Vision Board, Business Model Canvas, or Lean Canvas.
Employing techniques such as direct observation, problem interviews, and competitor analysis helps them validate their assumptions and update the board or canvas, as the following pictures illustrates:
Additionally, the developers may create spikes (throwaway prototypes) to explore the feasibility of the product. Target users and customers participate in the process, for instance, by acting as interview partners.
Once product owner and team have shown that there is a problem that’s worthwhile addressing, the focus changes to validating the solution and developing the actual product: The team needs to learn what the desired user experience is, what functionality the product should provide, and how it should be built.
The new focus requires adapting the team composition: The product owner, the UX designer and the developer stay, but the marketer, sales and service representatives leave the team. They continue to participate in the development effort as stakeholders. New developers, testers, and other roles required to create a great product join the team.
Product owner and team capture their assumptions about the product including the user interaction, the user interface design, the functionality, and the nonfunctional aspects using a tool like the Product Canvas and techniques such as scenarios, user stories, and design sketches. The assumptions are tested by collecting and analysing feedback on prototypes, mock-ups, and product increments/MVPs. Useful techniques to gather the feedback include product demos, users tests, and releases (to selected users).
As you may have noticed, the picture above suggests that “Stakeholders incl. users” participate in the process, and provide feedback or data on work results. While using target users and customers to validate ideas is generally helpful, it is not always appropriate. Imagine addressing a key technical risk in your first solution validation iteration: It probably makes little sense to invite users to the review meeting to understand if the approach chosen is viable. Similarly, if a disruptive product is developed it can be hard to find target users that are not too attached to their current solution.
Scaling and Launch Preparation
Once the Product Canvas and the architecture have started to stabilise, you can start adding more people to the project. I find it useful to break-up the original team so that at least one or two members are part of each new team, as this helps the new members get up to speed. I also suggest you grow your new product development project in a step-wise fashion: Scale from one to two teams, then from two to four, and so forth. This allows you to understand the impact on the people and the process including product ownership .
Be aware of the danger of premature scaling: Adding too quickly too many people. This tends to lead to a bloated, over-engineered product in my experience, and it prevents you form being able to experiment effectively. Therefore delay scaling until you have resolved the main risks – until you can focus on completing features and adding new ones.
Finally, as you make progress with your solution validation work, don’t forget to prepare the product launch. Having a marketer present at the sprint review meetings should help, but you may find that a dedicated marketing may be required to prepare and execute the launch well.
The following table summarises my recommendations for transforming an idea into a shippable product:
The process described in this post is based on work by Steve Blank, Ash Maurya, Eric Ries, and Ken Schwaber. The eagle-eyed process historians amongst you may have noticed that the idea of progressive scaling has its roots in the (Rational) Unified Process. Experimental, iterative product development has been around for quite some time of course, I believe at least since the late 19th century when Thomas Edison established the Menlo Park laboratory.
If you have any feedback, comments, or questions, then I’d love to hear from you!