Unorthodox PM Tips: Automating User Insights, Unselling Candidates, Decision Logs, and More | Kevin Yien
Product management advice often centers on familiar rituals: roadmap planning, prioritization frameworks, customer interviews, and stakeholder alignment. In this conversation on Lenny’s Podcast, Kevin Yien offers a more unconventional operating system built around clear constraints, direct customer exposure, deliberate reflection, and unusual hiring practices.
Yien leads product for merchant experiences at Stripe. Earlier in his career, he helped build Square’s restaurant business and ecosystem teams and served as head of product and design at Mutiny. His ideas draw on those varied environments, but they share one premise: a product manager’s job is to turn a team’s latent capability into value that customers can actually experience.
Product Management Is a Function, Not a Mandatory Position
Yien argues that every product team needs certain outcomes, but not necessarily someone with the formal title of product manager. A team must still understand its users, decide what matters, establish direction, coordinate execution, and judge whether the finished work creates value. In some organizations, engineers, designers, founders, or salespeople can collectively perform those duties.
The need for a dedicated PM becomes stronger as the distance between builders and customers grows. When employees closely resemble the people using the product, the team may acquire context almost automatically. When the users operate in an unfamiliar industry, face specialized constraints, or interact with a complicated ecosystem, someone must bridge that gap deliberately.
This view reframes the PM role. Its purpose is not to own every decision or manage a ceremonial process. It is to ensure that vital product work does not fall between organizational boundaries.
Entering Product Management Through an Adjacent Discipline
For aspiring PMs, Yien recommends beginning with roles whose responsibilities product managers frequently absorb. Engineering develops technical judgment and an understanding of implementation tradeoffs. Design builds sensitivity to user behavior and product quality. Sales teaches customer discovery, persuasion, and commercial awareness.
Moving from one of these disciplines into product management should not be interpreted as failing to become an exceptional engineer, designer, or salesperson. Yien suggests treating it as evidence that a different combination of strengths may offer greater leverage.
The practical lesson is to build a strong foundation before chasing the title. A person who can already make, design, or sell something brings a useful perspective to cross-functional product work.
Writing Turns Ambiguity Into Coordinated Action
Yien considers writing one of the PM’s central skills because the role repeatedly requires converting incomplete thoughts into shared understanding. Writing exposes weak reasoning: an idea that feels coherent internally may become obviously vague when placed on a page.
Good product writing does more than document a conclusion. It helps the writer understand the problem, makes assumptions visible, and gives colleagues something specific to question. A clear document can connect strategy to execution by explaining the customer need, the intended outcome, and the boundaries within which the team can operate.
To improve, Yien recommends reading widely and paying attention to prose that changes how the reader thinks or acts. Style matters as well as logic. Sentence length, rhythm, and variation can keep important material from becoming exhausting. The goal is not literary ornament; it is communication that people can absorb and use.
Give Engineers and Designers a Perimeter, Not a Script
Yien believes product organizations frequently underinvest in design. Increasing design capacity relative to engineering can improve not only visual polish but also the coherence and usability of what the team ships.
Within that partnership, the PM should define the problem and establish meaningful constraints without prescribing every detail of the solution. Yien describes this as drawing a perimeter. Inside it, designers and engineers should have room to explore.
A useful perimeter might clarify:
who the product is for;
which need deserves attention;
what platforms must be supported;
which tradeoffs are acceptable;
whether speed, reliability, consistency, or another quality takes priority.
Constraints can increase creativity because they remove unproductive uncertainty. Telling a team that responsiveness matters more than immediate data consistency, for example, provides a concrete basis for technical and design choices.
Empowerment does not mean detachment. Yien still expects PMs to inspect the finished experience closely and advocate for customer value. If the details do not hold together, the PM may need to challenge an internal convention or argue for more time.
Make Feedback Synchronous When Speed Matters
Sending a document to several colleagues and requesting asynchronous comments can appear efficient, yet the review may linger for days. People postpone reading, interpret passages differently, or leave comments that create another round of clarification.
Yien proposes an alternative: schedule a working session in which participants read and comment on the document together. Although it adds a meeting, it also reserves focused time for the review. Questions can be answered immediately, ambiguous feedback can be explored, and the author can begin revising while the context remains fresh.
This practice is most valuable when a decision is blocked by feedback. It is not an argument that every document deserves a meeting. Rather, it recognizes that a concentrated hour can sometimes replace a long chain of fragmented exchanges.
Build Product Sense With a Decision Log
“Product sense” can sound like intuition that certain people simply possess. Yien makes it more trainable: make a judgment, record why you made it, observe what happens, and revisit the reasoning.
A decision log should capture the available evidence, the assumptions in play, the expected result, and the chosen course. When the outcome becomes visible, the PM can compare reality with the original prediction. The point is not to punish incorrect calls. It is to detect recurring blind spots and improve future judgment.
The exercise can also use decisions made by other teams or companies. A PM might document what they believe an organization is trying to accomplish, predict the consequences, and return later to evaluate the result.
Yien recommends making the habit small enough to sustain. Ten minutes each week spent examining one decision is more useful than designing an elaborate system that is quickly abandoned.
Apply the Same Learning Loop to Interviews
Hiring decisions offer a particularly clear use case for decision logs. Interviewers can record the questions they asked, the evidence they considered, and the reasons behind their recommendation. Those records become far more valuable once the candidate has worked at the company for six, twelve, or eighteen months.
By comparing actual performance with the original interview scorecard, a team can test whether its signals were meaningful. Perhaps a heavily weighted answer did not predict success, or a concern that seemed minor later proved important. Without retrospective review, companies may repeat weak interviewing habits while assuming their process works.
This approach also encourages intellectual honesty. Interview quality cannot be assessed solely by whether everyone followed the prescribed steps; it must be evaluated against the eventual hiring outcomes.
“Unsell” the Role Before Making the Offer
One of Yien’s most distinctive hiring practices is the unsell email. After interviews but before the offer, the company sends the candidate a candid explanation of the role’s potential disadvantages and the organization’s difficult realities.
The message is not designed to repel people theatrically. It surfaces concerns candidates may already have and provides information that ordinary recruiting conversations often soften. A person considering the role should understand its constraints, frustrations, and risks before committing.
Yien acknowledges that greater candor can increase the number of declined offers in the short term. He sees that as an acceptable filter. Candidates who proceed are doing so with more realistic expectations, which can produce a stronger match, greater satisfaction, and better retention.
The broader principle is that recruiting should optimize for durable fit, not merely offer acceptance.
Automate Access to Customers, Not the Understanding Itself
Yien warns PMs against relying exclusively on research summaries or interpretations passed through sales, support, and other departments. Processed insights are useful, but they cannot fully replace hearing customers describe their circumstances in their own language.
The repetitive logistics surrounding research, however, can be automated. Yien points to services such as UserInterviews.com for recruiting and scheduling people who match a target profile. The PM still conducts and interprets the conversation, while software reduces the administrative friction that makes interviews easy to postpone.
Existing sales calls provide another research channel. Sales teams speak with customers continuously, and conversation-intelligence tools such as Gong can flag relevant words or themes. Those signals can trigger a workflow inviting selected customers to a follow-up interview.
This creates a practical division of labor: automate discovery, alerts, outreach, and scheduling, but preserve direct human exposure to the raw conversation.
AI-Native Users Will Bring Different Assumptions
Yien illustrates changing expectations with a story about a seven-year-old using an AI image generator. The child described the activity simply as “drawing,” without treating the model as categorically different from crayons or other creative tools.
For adults, AI-generated imagery may invite debates about interfaces, authorship, and technical novelty. For children who grow up with these systems, AI may become an ordinary material for making things. Yien suggests that this generation could eventually approach technology and product development with assumptions that current teams do not yet share.
The anecdote is a reminder to look beyond how existing users describe new technology. Tomorrow’s users may not inherit today’s conceptual boundaries.
Separate Personal Failure From Environmental Mismatch
Yien also recounts losing a startup PM job while his wife was nine months pregnant. His reflection is not a simplistic call to treat failure as positive. Instead, he distinguishes between lacking a capability and working in an environment that no longer needs what someone does best.
After a setback, people should examine what they could have changed: their decisions, communication, execution, or relationships. They should also identify external conditions, including company stage and strategic need, that were outside their control.
That distinction supports a more accurate response. Sometimes the answer is skill development. In other cases, the better move is to find a context where the existing skill set is valuable.
Yien’s reading recommendations reinforce this theme. He favors memoirs and autobiographies for their access to other people’s reasoning, highlights The Courage to Be Disliked for its focus on controllable choices, and recommends Ken Liu’s The Paper Menagerie and Other Stories.
Across the conversation, his unconventional tactics form a consistent philosophy: make assumptions explicit, shorten feedback loops, stay close to customers, tell candidates the truth, and review decisions against reality. Product management becomes less mysterious when judgment is treated as a practice that can be documented, tested, and improved.
