A case study by Matteo Cavucci
How does a product discovery strategy look like for a software development team?
In most cases, it’s still a long list of things due by a specific date. The expectation of the managers, even in digital products, is to see these new features going out as soon as possible, without considering too much the motivation behind. So it’s not uncommon to experience a lack of alignment and accountability between people at different stages to a common business goal. Moreover, despite the best intentions of the original planners, the level of details defined in the product strategy is limiting the possibilities for teams to act based on what they learn.
For an organization stuck in this situation, it’s tremendously hard to succeed in a very demanding environment like the digital realm.
Getting into the challenge
A couple of months ago, I had the chance of supporting a large organization producing one of the most beloved software product of internet history. I can’t say the name for obvious reasons, but I’m sure you spotted their logo even multiple times per day.
The product team that contacted me had an ambitious mission: find a new way to attract and delight the next four million users through their ability to connect with different people.
Every day a large community of developers, ranging from semi-professional practitioners to well-known brands, are working to extend the native software capabilities with new functionalities and superpowers.This vast and diverse pool of creators is continuously demanding the best technology and infrastructures to develop and distribute their products. On the other side, users from all around the world want to discover new extensions, enjoying a tailored experience, and be in control of their setup.
For three days, a varied group of people made by product owners, tech leads, UX people and other specialists gathered together to rethink legacy procedures and reimagine the entire ecosystem.
The principles behind the action
Ideas are never a problem. Everyone working in an organization has multiple thoughts on how to improve the management, the process, or the product itself. But not all the ideas are the same. Some reveal unexpected complications and dependencies. Some others may create opportunities for specific customers only while generating backlash for others.
How can we move through this wealth of insights and transform them into an actionable strategy?
In some cases, it’s just a matter of the managers’ intuition and experience: they decide what is right and what is wrong based on what worked in the past. Sometimes, though, this approach reveals dramatic consequences, as it ends up producing features that are disconnected from the constantly- shifting customer needs.
Making software products means moving through uncertainty, and the best way to operate in this condition is through experimentation and test, instead of upfront decisions.
In this scenario, selecting ideas is an entirely different game. When you want to innovate through experiments, it’s not essential to avoid wrong ideas; instead, what is crucial is to quickly and cheaply test as many promising ideas as possible, and to learn which ones will work.
A simplistic, linear strategy wouldn’t have been so successful in this case. Our plan can’t be a set of specifications of what to do and then handed down to developers. Instead, we believe that customer-facing teams, which have a stable connection with end-users based on constant feedback loops, are the best people to decide how to do something. What these teams need is a clear direction, which needs to be shared and discussed focusing on framing the desired outcome.
As strategists, our primary job is to provide teams with just enough information to balance local autonomy with global product identity. In this case, our focus shifted to the problem we aimed to solve and the impact we wanted to observe from the customer perspective, entirely delegating the “how to do it.”
Let's do it
Let ideas merge
However, before we start organizing ideas, it’s necessary to let them appear, discuss them, and start collecting them in a proper form. From observing the way business ideas flow through different teams to partners and customers, we can spot much room for improvement. For this reason, creating a model of the current implementation allows people to share their thoughts, the issues they’ve been noticing, and the importance of specific insights.
EventStorming really shines when it’s time to review the current process and discuss possible changes. It was our starting point to approach the problem in a systemic, holistic way, and to gather everybody’s understanding of the service from different perspectives. Through the flow of the events (known facts occurring in the service domain, captured with orange stickies along a timeline on a very long wall), we explored in detail how developers submit their code, how users discover and install additional features for their setup, how the communication of a new API is handled, and the entire lifecycle of themes and extensions. Mapping together with the existing process made it easy to spot areas where changes can lead to significant impacts.
Orchestrating experiments at scale
After a brief discussion, the team produced a short statement picturing the future state of their product. It became, and still is the leading vision for the new product we were planning, a guiding direction toward which all the work should contribute. The Vision articulated how the new product would have to create value for customers and how it would have to change their behaviours.
Using a value/confidence matrix, we selected – through a quite challenging conversation – the most promising prospects. We noticed that some opportunities were going in the same direction, sharing the same purpose. Grouping them gave us the chance to investigate each high-level challenge, framing them as major strategic approaches that will have led from the current situation to the desired state, each one referring to a distinct customer segment and with a specific description of how success looks like.
Five, and then four high-level goals emerged from the discussion. The relationship between the leading Vision, the Goals, and the sub-relationship between each Goal and the possible opportunities created the first level of a tree-shaped representation of our aims.
It was the start of our Lean Value Tree.
The Lean Value Tree is a visual tool that makes it easy to get from plans for a value-driven strategy.
It allows leaders and teams to structure and manage options for change and improvements. It offers an effective way to share and analyze desired outcomes, examine risks and opportunities, and continuously redefine what is needed. It’s based on the idea of prioritization by results (and not features) introducing optionality as a crucial element of planning.
Each aspect of the tree contributes to the accomplishment of the leading vision through experiments directed to produce behavioural change.
At this point, the ideas that emerged from the previous exploration were refined into what we call Bets.
Each one is the current best thinking about how the respective Goal can be achieved and not a specific deliverable.
A Bet includes the fundamental hypothesis, the customer profile, and potential risks and possibilities.
Most importantly, each one contains a precise measure of success, based on the change we’d like to observe in the customer’s behaviour. The term “Bet” connotes the non-deterministic essence of the plan. Like in gambling games, we don’t expect that every bet will be successful. What is essential is having at least a single Bet producing impacts at a large scale.
With the Vision, Goals, and Bets articulated in an understandable visual scheme, it was now time for development teams to look at the picture and agree on the best way to proceed. They decided how to explore a Bet with multiple experiments, trying to verify (or falsify) the hypothesis from different angles. Each of them could include the creation of prototypes and (minimum) testable products, A/B or multivariate testing, and also contextual research and data analysis.
A first step into something new
The conversation around the Lean Value Tree had a tremendous impact on the way each team managed discovery activities.
It produced a complete roadmap for adapting the organization work to emerging opportunities, leveraging on collaborative decision-making.
This roadmap includes the customer-centric analysis of where the business is going, but also included budgeting, timeframes, dependencies, and risks. The strategic work is not acting as a constraint for innovation, but instead, it provides a structure for continuous integrating new findings and evaluating new ideas.
From what we achieved together, the organization is developing a path toward an adaptive, learning culture.
It’s not an easy journey. It requires the business to step outside the traditional habits and focus on the possible ways to serve the customer. It makes it necessary to imagine a way of working that is different from what perhaps worked very well in the past.
Learn with Matteo Cavucci
Matteo is the trainer of the Product Discovery Workshop with the Lean Value Tree.