Directed vs. Undirected Problem Discovery
Product Discovery is often described as the process whereby teams decide which problems to solve ("problem discovery"), and which solutions to ship ("solution discovery"), through the gathering of evidence to reduce risk.
A large part of finding the right problems to solve requires speaking to your users. Not just any users, but the right users. And not just about anything, but the right things. And while there’s rarely a good time to speak to the wrong users, as PMs there are times when it’s OK to talk to the right users about… anything.
The hunt for problems
There’s a fun saying for when someone is looking for something:
It’s always the last place you look.
But what if you don’t know exactly what it is you’re looking for? When you find something resembling it, do you stop looking?
Answering those questions is often more art than science, but I've come to realize through years of working on, advising, and speaking with product teams that there are two distinct modes PMs find themselves in when they're running problem discovery. I've named those modes Directed vs. Undirected Problem Discovery, and learning to recognize which of those mode you're in or should be, is going to make running problem discovery much easier and productive. Here's how I define them:
Directed Problem Discovery: Problem discovery with an agenda. Apply this to answer a specific question, assumption, or hypothesis to create immediate utility.
Undirected Problem Discovery: Problem discovery without an agenda. Apply this to gain a deeper understanding of your users and discover new opportunities for your product. (This may or may not have immediate utility.)
In practice
Directed Problem Discovery can be used either for known knowns (to challenge your own knowledge) or known unknowns (you have a hunch that you will find useful information that you are not aware of in the field of inquiry). While Undirected Problem Discovery can only be used to discover unknown unknowns (the insights you didn't know were there).
Another fun way to think about the difference (if you enjoyed calculus) is whether you’re on the hunt for a single, local maximum, or instead trying to map the location of multiple maxima in a larger search space. In Directed Problem Discovery, you’re guiding the conversation in a way similar to that of running a hill climbing exercise to converge on an answer. However in Undirected Problem Discovery, you’re mapping a larger search space and need to apply periodic jumps (or “perturbations”, to keep with the mathematics theme) to the conversation to discover other interesting maxima. Undirected problem discovery is a balance between uncovering more areas of interest, and shifting into Directed Problem Discovery to measure the maximum. While running Undirected Problem Discovery, you might find yourself saying something along the lines of “OK, let’s switch gears!” often.
The two exist on a continuum of whether your intention is to identify areas for future investigation, or immediate utility.
Both skills are invaluable to have in your toolkit. And knowing when to apply one or the other depends on your understanding of the user, their context, and your current needs. When you trip on gold (an unexpected insight) during Undirected Problem Discovery, an experienced PM knows it’s important to quickly transition to Directed Problem Discovery. Conversely, when you just can’t get to an answer for the specific question you’re trying to answer with Directed Problem Discovery – why not fall back on Undirected Problem Discovery, rather than just bailing on the interview? (we’ve all been there.)
Getting explicit
Formalizing concepts we've recognized ourselves or others apply helps us to reason about, collaborate on, and improve them.
Do you have a name for these modes of problem discovery? Do you think about it differently? We love to chat with product teams about their discovery practices – grab 20m with our founders!
