Continuous Discovery and the Opportunity Solution Tree

Premise — why one-off studies aren't enough

Conducting a user study once a quarter with a lump-sum budget has its limits. The product's situation keeps changing between studies, yet the material for decision-making is only updated right after a study. As a result, roadmap meetings tend to be debated on the basis of memories from several months ago rather than "recent user voices." In her book "Continuous Discovery Habits," Teresa Torres proposes an operation that prevents this staleness of information by accumulating small weekly touchpoints. However, continuous discovery is not a complete replacement for large-scale studies. For questions such as estimating market size or statistically significant comparisons between segments, a substantial quantitative study is still needed.

Procedure

1. Secure a weekly touchpoint as a fixed slot

Someone on the team (in many cases the PM and the designer) fixes time in the calendar to have at least one user contact every week. About 30 minutes per session is fine, conducted as a shortened version of the kind of deep-dive questions covered in 07-user-interviews. What matters is guaranteeing the frequency of "we'll do it again this week, without fail," not the scale of any single session. For continuous recruiting, permanently installing a simple sign-up path for research participation within the product keeps down the cost of reaching out each time.

2. Record statements as opportunities

Write down the realizations gathered from interviews, support inquiries, and behavioral data not in the form of solutions but in the form of "the needs, pains, and desires the customer holds." Rather than "we should add a search feature," record it as an opportunity: "unable to quickly find the one needed item among a large amount of past data." If you decide the solution first, you lose the room to compare and consider other solutions later.

3. Assemble the opportunity solution tree

The tree is made up of four layers.

Layer 1: Desired outcome
  e.g., Increase the weekly active rate of existing users

Layer 2: Opportunities
  e.g., Unable to quickly find needed information from a large amount of data
  e.g., Unable to grasp the progress of multiple teams on a single screen

Layer 3: Candidate solutions
  e.g., Add a search feature / Saved filters / Summary view

Layer 4: Validation experiments
  e.g., Usability testing with a prototype /
        Comparison of usage rates via A/B test

By placing the outcome at the apex, you can take a bird's-eye view of whether multiple opportunities are contributing to the same goal, and conversely, whether you are spending time on opportunities that don't contribute to the outcome. Laying out multiple candidate solutions for a single opportunity prevents jumping at only the first idea you come up with.

4. Update the tree weekly

Each week, decide whether statements obtained from new interviews should be linked to an existing opportunity or added as a new branch. If you keep using a tree you built half a year ago without updating it, it becomes the basis for a prioritization that has drifted from reality.

Applied example — connecting to the roadmap meeting

The tree's Layer 2 (opportunities) and Layer 3 (candidate solutions) directly become subject matter for the roadmap meeting. When you proceed to the prioritization stage, combine them with a scoring method like RICE, covered in 04-roadmap-prioritization, to decide which opportunity to tackle first. Because the tree remains as a record of "what was considered," it becomes easier to explain afterward why that initiative was chosen.

Verification

Confirm whether each opportunity in the tree has corroboration from statements or behavioral data within the last one to two months. An opportunity whose corroboration is old, or which rests on only a single statement, is marked as a priority target to confirm in the next weekly interview.

Common pitfalls

A pattern teams that start continuous discovery tend to fall into is that conducting weekly interviews becomes an end in itself, and the statements obtained accumulate without being reflected in the tree or the roadmap. Another trap is proceeding to an experiment with only a single candidate solution produced at Layer 3. The value of the opportunity solution tree lies in being able to lay out multiple solutions and compare them. On the other hand, if your team is small and lacks the capacity to secure weekly contact, it is more realistic to start from every other week or monthly and gradually raise the frequency rather than forcing yourself to continue every week. Rather than holding up a frequency you can't sustain and having it become a hollow formality within a few weeks, keeping to a frequency you can sustain preserves the freshness of information in the long run.

Read next

The question design and procedure for weekly interviews is covered in Designing and Conducting User Interviews. For the procedure to translate the candidate solutions that come out of the tree into an actual roadmap, see How to Build a Roadmap and the Practice of Prioritization.