How to Master Continuous Discovery for Product Success | Teresa Torres | Glasp Talk #31
This is the thirty-first session of Glasp Talk!
Glasp Talk delves deep into intimate interviews with luminaries from various fields, unraveling their genuine emotions, experiences, and the stories behind them.
Today's guest is Teresa Torres, a renowned product discovery coach and author of Continuous Discovery Habits. Teresa is widely recognized for her work in helping product teams develop sustainable discovery practices that are both customer-centric and valuable to the business. She shares her expertise in opportunity mapping, hypothesis-driven decision-making, and cross-functional collaboration, all of which have revolutionized the way companies build products.
In this interview, Teresa discusses her influential book, which has sold over 100,000 copies worldwide, and elaborates on the core principles of continuous discovery, emphasizing the importance of feedback loops and understanding both customer and business value. Teresa introduces the concept of the Opportunity Solution Tree and provides real-world examples, including insights from her coaching experiences with product teams.
Throughout the conversation, Teresa touches on how product teams can prioritize opportunities strategically, the overlap between opportunity mapping and frameworks like Jobs to Be Done, and how visionary companies can apply these principles even when customer needs are not yet well-defined. She also shares her perspective on the evolving role of AI in product management and how teams can balance visionary leadership with practical, customer-centric product development.
Read the summary
Glasp
Transcripts
Glasp: Hi everyone, welcome back to another episode of Glasp Talk. Today, we are very excited to have Teresa Torres with us. Teresa is a renowned product discovery coach and author, best known for her influential book Continuous Discovery Habits. She has guided countless product teams in adopting customer-centric approaches and developing sustainable discovery practices. With her expertise in opportunity mapping, hypothesis-driven decision-making, and cross-functional collaboration, Teresa has revolutionized the way companies build products that deliver both customer value and business value. Beyond coaching, Teresa shares her insights through her blog, Product Talk, at producttalk.org, where she provides actionable advice on product discovery and management. Today, we’d like to explore her journey, her approach to continuous discovery, and how teams can create products that truly resonate with their customers. Thank you for joining today!
Teresa: Thanks for having me!
Glasp: Thank you. First of all, it's really impressive that your book has sold over 100,000 copies worldwide! That’s impressive, and we love the book. Could you tell us about it for the people who may not know about it yet? Also, what are Continuous Discovery Habits in your words?
Teresa: Yeah, my goal in writing the book was to create a really practical guide for product teams. When I say "product teams," I mean the cross-functional team: product managers, designers, and engineers. I wanted to give them a guide for how they can make good decisions about what to build. There are a few key tenets to that. If we want to make good decisions, we need to build in feedback loops. How do we know if we made a good decision? How do we measure it? How do we know if it was the right thing? In the product world, we’re trying to build for our users, for our customers. We want to have a feedback loop with them: Is this working for them? Is it solving a need for them? Are they going to continue to engage, use it, or buy it again? Continuous discovery is the work that we’re doing when we’re deciding what to build. Continuous discovery means any digital product is always evolving; it’s never done. So, if it's never done, we're continuously making decisions about what to build, and we need to be looking at continuous feedback loops. How do we make sure that on a daily basis, we're getting feedback from our customers that what we’re building is the right thing to build for them? The book is meant to be a practical guide to help teams do that.
I look at it as three big components, and then each of those three components has smaller subcomponents. The first one is really understanding how we're going to create value for the business so that we earn the right to serve the customer over time. That means defining clear outcomes to make sure that your business will be around for the long term. The second is making sure that we're understanding customer value, so how do we create value for the business in a way that also creates value for the customer? And then the third piece is discovering the solutions that address that — creating customer value in a way that also creates business value. That's super high-level, but we can get into the specifics. My intent with the book was to create a practical guide to help teams make better decisions about what to build.
Glasp: Thank you. You mentioned opportunity mapping, like starting with a desired outcome — not output but outcome-focused — and breaking it down into opportunities. Then for each opportunity, you can create solutions and experiment for each solution. Could you elaborate on opportunity mapping?
Teresa: Sure. I created this visual called the Opportunity Solution Tree, and it’s meant to represent those three components I talked about. The top of the tree is the outcome. Our outcome represents business value. From there, we look at what I call "opportunities." Opportunities are unmet customer needs, pain points, or desires. So, how can we positively intervene in our customers’ lives in a way that drives our business outcome? We're aligning that business goal with customer value. For any individual opportunity, we’re looking at solutions that can solve that opportunity in a way that drives business value. As you work your way down the tree, we're starting with the business need, then there’s the opportunity space. We're mapping the opportunity space to understand customer needs, which allows us to make a more strategic decision about where we want to play. We’re looking at all the areas where we could create customer value, allowing us to make a more strategic decision about where we want to create customer value. Then, we can iterate through smaller opportunities so that we can continuously deliver value. A well-structured opportunity space using an Opportunity Solution Tree allows us to find, week over week, what we can do that creates value for our customers in a way that also creates value for the business.
Glasp: That’s a really insightful mapping. People can visually understand the concept, and that’s great. Thank you. Do you have an example? I think you've coached many product teams, and you’ve explained this concept hundreds or thousands of times. Do you have a good example that people can easily understand?
Teresa: Yes, one of the examples I use throughout my book relates to streaming entertainment, like a company such as Netflix. Part of the reason I use that example is it’s a product everyone around the world is broadly familiar with. So, you can imagine if you worked at a company like that, one of the outcomes or business needs might be to increase viewing engagement, to get people to watch your service more often. That’s tied to a business goal of driving retention, so in a subscription business, we want people to retain. Maybe our theory of retention is that the more you watch, the more likely you are to retain.
So, the top of our tree might have an outcome like "increase average viewing minutes per week." Then, we want to go out and interview our customers to understand how our product fits into their world. Presumably, they’re buying our product to be entertained, so we’d want to learn when and where they’re watching, who they’re watching with, and so on. As we collect those stories, we’ll learn things like common pain points, such as "I can't find something to watch" or "I want to have a good viewing experience, but I have a slow internet connection, so it constantly buffers."
As we hear customer stories, we’ll hear these opportunities, needs, pain points, and desires that, if addressed, could help us drive our outcome. So, if I interview you and try to figure out how to get you to watch Netflix more, you might tell me a story about wanting to watch Netflix but not being able to find something. You might say, "I can’t tell if I’m going to like this show." The high-level need is finding something to watch, but a sub-need might be that you want to know if you're going to like a show. We can visually map this. An Opportunity Solution Tree is like a decision tree – we’re breaking a hard problem down into its subcomponents, and for each of those, breaking it down even further.
What that allows us to do is that we might spend years helping people find something good to watch, but if we work our way down the tree to something small, like "Who's in the cast of this show?" – that’s something we could solve in a short period and deliver a solution, then move on to the next thing that’s neighboring it. As we do more of that, we start to solve the higher-order problem of "I can't tell if I’m going to like this show."
Glasp: I see! So, in my understanding, through user interviews and so on, we should collect user stories to understand the context and situations in which people use the product. Isn’t this similar to the jobs-to-be-done framework? Or do you have another way of looking at it?
Teresa: Yes, there’s a lot of overlap. In a jobs-to-be-done interview, we tend to focus on the purchase decision. There’s this concept that when you buy a product, you’re hiring it to do a job, and what we’re listening for is what that job is. I teach something called story-based customer interviews, where I collect specific stories about what actually happened. In my Netflix example, I might ask, "Tell me about the last time you watched Netflix" or, if I’m on the mobile team, "Tell me about the last time you watched Netflix on the go." Or, if you're on the search team, "Tell me about the last time you searched for something on Netflix." All those are story-based questions. None of them is a jobs-to-be-done interview.
A jobs-to-be-done interview might be something like, "Tell me about why you decided to sign up for Netflix," because jobs-to-be-done is really focused on the purchase. I’d say jobs-to-be-done is a great story-based interview format, but it works when we’re talking about the highest-level job.
And here’s the thing: I’m familiar with Tony Ulwick’s work on jobs-to-be-done and especially Bob Moesta’s work on it as well. They both, I believe, go more specific than just the purchase. But the way most companies use jobs-to-be-done is at that high level around the purchase. They end up with a job like, "I want to be entertained," or "I don’t want to be bored after dinner at home." That’s the job I’m hiring Netflix for.
With my opportunity mapping, though, we’re getting much more granular. Yes, there’s that high-level job – wanting to be entertained – but there’s also a very specific need if I can’t find something to watch. The reason why I can’t find something to watch might be because I can’t tell if I’m going to like a show, and the reason for that might be because I like character-driven dramas and not plot-driven dramas, and I don’t know which type this show is.
So, my goal with opportunity mapping and getting more specific is to tackle these hard, intractable problems iteratively over time. Whereas with jobs-to-be-done, the way most people use it, it’s more at that high, purchase-decision level. I think a lot of people who invented or popularized jobs-to-be-done would say it can be done at narrower scopes, but I rarely see companies apply it that way.
Glasp: I see. Thanks for clarifying the difference and the granularity. At the same time, when we evaluate opportunities, there are so many metrics or ways to prioritize – like the ICE metrics, where you look at impact and ease, or OKRs, and so many others. Do you have a favorite framework or process for evaluating opportunities?
Teresa: Most of those frameworks, like ICE, are applied at the solution level, and I think that’s the wrong level. Most of the time when people are using ICE or RICE, they’re looking at a list of features or ideas and trying to estimate the impact of those features, the confidence in that impact, and the effort required. I think the challenge there is that it skips the more strategic question: what’s the most important customer need or problem for us to solve?
I rarely see teams apply those frameworks at the opportunity level, and I think we fall into this trap of prioritizing solutions instead of opportunities. The more strategic decisions are made at the opportunity level – deciding which opportunities to go after. This is where understanding the opportunity space and choosing opportunities that nobody else is solving can create differentiation for your product.
So, my first recommendation is to prioritize opportunities, not solutions. Once you’ve chosen a target opportunity, you can then compare and contrast solutions based on how well they address that specific opportunity. Instead of needing a scoring framework, you’re using prototyping and assumption testing to evaluate if a solution addresses that need.
Now, as for how to evaluate opportunities, there are lots of frameworks for assessing them, but I think it’s most important to use a framework that matches your stakeholders' expectations. Choosing a target opportunity is where we often need to defend our decisions internally, and different companies at different times will value different criteria.
I think about it as four broad categories of criteria:
- Opportunity sizing – how many customers are impacted and how often?
- Market factors – is this opportunity a differentiator or a table-stakes requirement? How does it affect our market position?
- Company factors – how well does this opportunity support our company’s strategic initiatives?
- Customer factors – how important is it to our customers?
So, for example, if you’re at Spotify and they’re making a big strategic push for podcasts, it doesn’t matter if you’ve uncovered an opportunity in music that customers care about. The company’s strategic context is saying, "all hands on deck" for podcasts and audiobooks. Or if I’m working on generative AI and feel my company’s behind, maybe we focus on opportunities that are table stakes for AI rather than broad customer impact.
It’s not about one best framework; it’s about tailoring the criteria based on what your company needs at that moment.
Glasp: I see. And does the same concept apply to more visionary companies, like consumer companies where the founder is very forward-thinking? I’m thinking of companies like Snapchat or others where customers might not even realize they need the product.
Teresa: Yes. We’re seeing this with generative AI. The average person might not even fully understand what generative AI is or what it can do, but that doesn’t mean we can’t do discovery. Instead of interviewing customers about solutions, we’re interviewing them about their lives and needs. We’re listening for opportunities where generative AI might be a good fit.
A big mistake teams make is using interviews to explore solutions. It’s just not effective. The purpose of an interview is to understand different contexts and needs. For example, I can ask people how they spend their evenings after work and listen for unmet needs like boredom or gaps in entertainment, regardless of the solution I plan to build. This gives me valuable insight.
We’ll never interview a human who says, "I need generative AI." But people will express needs where AI can provide a solution. So, it’s not about how visionary the founders are; it’s about the match between that vision and what people need in their lives.
Glasp: That makes total sense. But sometimes we hear product teams say they’re following this process, then suddenly the founder or executive suggests doing something entirely different, creating some miscommunication. Any tips to fix this?
Teresa: Yes, I think there's a misconception here. People think an "empowered team" means they get to do whatever they want, but that’s not it. An empowered team is empowered to make the daily decisions they need to serve the business, but they’re still operating within a strategic context.
Let’s take Spotify again. Historically, they’ve been focused on music, but if the executive team decides they’re making a big push for podcasts, the entire organization is going to focus there. And that’s fine. Leadership is responsible for steering the ship, setting direction, and sometimes changing direction based on the market or company needs. Product teams then need to align their work with that new context.
People think of it as a false dichotomy – either team are empowered, or they’re not. But it’s more about the scope of empowerment. When cash flow is ample, teams are often empowered in a broad scope. When cash tightens, that scope might narrow. But every team, regardless, still makes dozens of important decisions every day.
Glasp: Yes, that makes sense. You also mentioned technology trends. As we’re seeing AI and LLMs become more prevalent, do you think they’ve affected how product teams work or customer needs?
Teresa: Yes, we’re already seeing it make a big impact. There’s a product called ChatGPT that’s helping product teams write better PRDs, user stories, and other documentation. It’s tedious to write detailed stories that cover all edge cases, but an LLM can fill in those details and help find potential edge cases.
People use ChatGPT for data analysis, synthesizing themes, and identifying patterns across large data sets. I use it personally for social media analysis, finding out what’s resonating with audiences and driving engagement. I think AI has tremendous potential to change the way we work.
What I don’t like is when people outsource understanding humans to AI. Companies are creating synthetic user interview tools, but technology needs more humanity, not less. Some tools, like one-click opportunity solution trees, also miss the point – you’re supposed to be synthesizing and aligning as a team, not just automating it.
So, I think AI is valuable, but it’s about understanding its role as a tool, not a replacement for human insight.
Glasp: Yes, technology should support but not replace that human touch. Just curious, what AI tools do you use for your daily work?
Teresa: I use both ChatGPT Plus, the paid version, and Claude. They’re useful in different scenarios. ChatGPT is great for quick answers, like Google. For example, if I need to know if my dog can eat something, I can get an immediate answer without clicking through articles.
For work, I use ChatGPT for data analysis, uploading spreadsheets, and asking questions about them. For creating summaries, I prefer Claude because it matches my writing style better. ChatGPT won’t return text verbatim, which is a guardrail from a New York Times lawsuit, I believe, but Claude can pull text from an article and summarize it in my own words.
I use ChatGPT for SEO and web-browsing needs, like analyzing how to rank articles better, and Claude for tasks that require preserving my voice in writing.
Glasp: That’s awesome. Since you’ve coached tens of thousands of people, have you ever thought about starting a company again or joining a big company as an executive?
Teresa: I think I’m past the stage where I’d want a traditional job, though I loved my time at startups. The energy at early-stage companies is amazing. But starting a software company is a different beast, and I think I’m at a stage in life where I’m more focused on sustainable impact.
I work with some companies closely and consider that enough. I miss working with engineers and managing teams sometimes, but I feel like I can have more impact doing what I’m doing now.
Glasp: That makes sense! And, as our audience includes aspiring product managers, founders, and writers, do you have any advice for them?
Teresa: Sure. One thing I learned the hard way was to choose carefully which hierarchies you want to climb. Every decision you make places you on a different hierarchy. Many people chase success without asking if it’s the path they truly want.
For founders, especially in the Silicon Valley area, it feels like being a founder is the ultimate success. But it’s also one of the hardest jobs, with a very low chance of success, and it can lead to burnout if you’re not careful.
So, my advice would be to periodically reflect on whether you’re on the path you really want. I became a CEO at 32 during the 2008 downturn, and it was so hard that it forced me to reevaluate what I wanted to do. Now I’m doing work that feels meaningful to me. I think that’s a pretty good legacy – just spending time understanding human needs and helping product teams connect with their customers.
Glasp: Beautiful advice. Thank you so much, Teresa, for sharing your insights and advice today!
Teresa: Thank you! It was great to be here and share my journey with you.
Glasp: Thank you for joining us. We loved the discussion and all the valuable insights. Thank you so much!
Teresa: Thank you, and thanks for having me!
