How do you do customer discovery, and what is the hardest thing for you about it?

Talk to customers, talk to customers and talk to customers!! Any time you have a chance to talk to a customer or a potential customer take it!

I find that mixing qualitative discussion and validating with quantitative research that we are not solving a problem for a loud customer or the most recent account review helps ensure you are working on a problem with real reach and that solving it will have real impact for customers.

For me the best signal of a customer with a real problem is one who is already trying to solve a problem with a workaround. A common example of this I have run into during my career is in building simple user dashboards for customers in a product. This shows up in many ways during customer conversations. You may get questions about an API to pull data, rate limits on an API, specific questions about how to get usage data, etc. Those are good hints that customers are trying to get data for themselves to present or analyze so there is an opportunity to provide it for them in a dashboard OR if a data interface is not your core competency provide them a richer data feed so they can get at the data they need.

When it comes to measuring reach and impact of a problem I like to mix a quantitative and qualitative approach.

For reach once I have validated there is a problem I'll try to extrapolate how many other customers might have this issue. Using the dashboard example above we might look at how many users are also hitting those specific endpoints a customer asking about the APIs for. If there is not enough attempted usage here (or enough conversations) then it may not be the most important problem to solve right now.

For impact I like to work with customers to answer this by comparing and contrasting problems to understand the urgency of the problem. This might work by asking them to stack rank against other requests they have made or as simple as "are you trying to solve this now and if not what are you trying to solve?"

Using these approaches helps me keep the recency of the most recent conversation from influencing what I prioritize next.

Customer discovery should never be treated as a 1-time planning activity. It should be an ongoing process to define a running backlog of opportunities.

Customer discovery should be done across all available & valid channels - customer interviews, competitive intelligence, product usage analytics, support interactions, customer facing team interactions to name a few. The hardest thing is to identify and  keep an eye on all valid channels to build a holistic view. This can be very time intense.  Across everything I have to do a PM leader, I make a deliberate effort to block some time for customer discovery every week but honoring that time myself is a challenge many times - given all the other priorities that come my way.  Given customer discovery is time intense,  availability of time and resources to conduct the due diligence by other cross functional teams such as UX researcher, data analyst can many times be a challenge. Embracing AI co pilot products in this space is what I see as an unlock to making PMs impactful in customer discovery space. 

Validating the problems before scaled investment is table stakes - and how it is done depends on whether it is a B2B or B2C product. For B2B, reviewing the roadmap with customers periodically is an effective way to make sure you are building that has both customer and business value.You can test PMF of B2B products by unlocking a few paying customers first before expensive GTM motions. From a  B2C perspective, low-investment way to A/B experiment and validate the impact with data before scaling is an effective way to drive the right product investment.