Norms
- Communicate if you can’t finish, if you would want to know anything
- Communicate any small disagreement or confusion in low stakes – don’t leave anything in the locker room
- Don’t be mean
- Established meeting times, not ad hoc
- Together time is precious
- Decisions together, execution solo
- You are accountable for your work
Affinity map
AI second pass
The AI second pass caught several patterns that we missed – one being the importance of customer understanding with one interviewee saying that it is a PM failure when a product doesn’t solve their problem and another interviewee saying it was hard to engage customers and ask the right questions. This second pass also picked up on a nuance as I had grouped several interviewee problems into one group under “lack of meaningful work”, but the AI patterns identified these challenges in a more detailed manner by pointing out that two interviewees both described stale meeting where nothing was decided, and that the main issue itself was not that days were packed with meetings. AI flattened several distinct points into something generic by classifying five different points under the pattern header “influence without authority and stakeholder juggling.” This pattern grouped a disagreement of one team with the other teams, a disagreement in product strategy between two executives, an overload of feature suggestions from one stakeholder, and a need for trust into one bucket. While several of these points are similar because they talk about disagreement, this header is too general and could be broken into smaller and more nuanced pattern buckets. The AI pass did assert several things that were not in the transcript – one false assertion being that it claimed the two interviewees who worked at startups were more “go, go, go” with less formal testing, but only one of the interviewees explicitly mentioned or implied that there was less formal testing than traditional companies. Additionally, the AI second pass asserted that one interviewee said that it was hard to keep up with tools, but while this was raised by another interviewee, the first interviewee actually said that there are a lot of PM tools coming out, but that it is important to keep track of them and up to date on how to use them and that it is exciting that so many are coming out.
Long list
- PMs lack authority which makes cross-team buying difficult
- Mentioned by 6 interviewees
- “Being a PM is like being the CEO and the janitor because you have to move the ball over entities you don’t have control over”
- PM roles vary by company and are ambiguous, so it is hard to transfer skills between roles
- Mentioned by 4 interviewees
- “In name it’s the same role, but because the size of the companies were different, the way the teams communicate and operate are different”
- Companies adopt AI tools differently and it is difficult to keep up with all of the new tools coming out
- Mentioned by 3 interviewees
- “Even with the technologies that are coming out now, it’s hard to stay on top of everything or just only learn one specific tool”
- There are too many competing requests and inputs to prioritize
- Mentioned by 3 interviewees
- “It wasn’t manageable to do all that”
- Failing to understand the user’s real problem and struggling to reach users
- Mentioned by 2 interviewees
- “How important talking to the customers and also how challenging it can be to engage them and ask the right questions”
- Execution and routine works crowds out strategic thinking
- Mentioned by 2 interviewees
- “If all you you’re doing is executing, anybody else can execute, given a plan”
- Large amount of manual QA required, especially when validating AI-generated work
- Mentioned by 2 interviewees
- “You really need to make sure that you’re checking all the work that’s been done with these types of AI tools”
- Time taken up by inefficient meetings that produce no decisions
- Mentioned by 2 interviewees
- “Whenever meetings are stale and nothing is decided. It’s like I just wasted time”
TAM for the PM market
We are making the assumption that we are only considering the PM market in the United States. In class, we learned that an estimate for the number of PMs in the US is 300,000 to 500,000. Based on research about the tools that PMs use (https://www.getproductpeople.com/blog/product-management-tools-how-to-build-the-right-stack), we think that they will have different tools for strategy and roadmap, feedback and discovery, delivery tracking, analytics, and collaboration. Choosing a popular tool for each of these domains, we estimated that the average PM would use Productboard, Jira, Figma, Claude Code, and Miro. We then looked out the websites of each of these products to figure out the yearly cost of an enterprise account for each tool. Combining $220 (Productboard), $150 (Jira), $60 (Figma), $240 (Claude Code), and $240 (Miro), we estimated that a product manager spends about $910 on tools every year. Multiplying this number by the number of PMs in the US, we estimate that the yearly US market is from $273,000,000 to $455,000,000.
Reflection
From round one to round two, we all got better at asking follow up questions to dig deeper into how our interviewees solved problems to invite stories instead of just asking a question, listening to the answer, and then asking the next question on the list. We also changed the focus of the interviews, as the round one interviews were more broad and focused on daily PM life and challenges and successes while round two dug more specifically into these challenges by asking about navigating conflict and competing ideas and the tools they used.
One idea from this week’s reading
The article says that a good interview question is open-ended, is not a leading question, asks about the past, and is grounded in a specific instance, with “tell me about the last time you watched Netflix” identified as a good question. We tried our best to ask questions that fit these criteria, and as one example, when asking about challenges after an interviewee discussed receiving constructive feedback, one of our questions was, “tell me about the last time you received critical feedback.”
