How to Win Friends & Influence Decisions: Julie Zhuo at Lenny & Friends Summit 2024
Product prioritization is rarely difficult because a team lacks ideas. It is difficult because several worthwhile ideas compete for the same time, people, and attention. A reliability improvement, an onboarding redesign, and a new generative AI feature may all have legitimate advocates—and each advocate may be working from a different definition of urgency.
In her Lenny & Friends Summit 2024 talk, design leader and founder Julie Zhuo presents a five-step process for navigating these conflicts. Her central argument is that influence is not simply the ability to tell a persuasive story. Better decisions emerge when teams create shared ownership, combine different perspectives, examine evidence, assign responsibility clearly, and revisit what happened afterward.
Why Prioritization Becomes a People Problem
Zhuo opens with a practical choice among three product investments: rewrite a piece of infrastructure, redesign onboarding, or introduce a generative AI capability. The divided audience vote demonstrates the problem immediately. Without more context, none of the options is obviously correct.
Each proposal can be supported by reasonable assumptions. Weak infrastructure might threaten performance and retention. Better onboarding could help new customers reach value sooner. An AI feature might materially improve the product’s promise or protect its position in the market.
The temptation in such situations is to improve the pitch for one preferred option. Zhuo says she once treated influence largely as a storytelling challenge: if she could present an idea clearly enough, other people would support it. Her experience as both a design leader and a founder showed that persuasion alone does not resolve the deeper issue. A productive process must first reveal why intelligent people disagree.
Step One: Draw a Circle Around the Group
The first step is deciding who belongs inside the decision-making circle. A person can draw that boundary tightly around their own position and treat everyone else as someone to defeat. Alternatively, the circle can include the people who hold different views.
This reframing changes the opening question. Instead of asking, “How do I convince them that my project should win?” the team asks, “How do we make the best choice for the outcome we all want?”
Zhuo recommends beginning with an explicit common purpose, such as creating a product customers genuinely value. This may sound elementary, but it changes the emotional structure of the conversation. Colleagues stop acting like rival campaigners and begin working as contributors to the same result.
Shared purpose does not eliminate disagreement. It makes disagreement safer and more useful. Once everyone is inside the circle, changing your mind no longer means losing to another department. It means helping the group reach a better answer.
Step Two: Treat Each Perspective as a Partial Truth
Zhuo uses the familiar parable of people encountering different parts of an elephant. One person touches the trunk, another the leg, and another the side. Their descriptions conflict because their experiences are incomplete, not necessarily because anyone is dishonest or incapable.
Product debates often work the same way. Engineering sees operational fragility. Design notices confusion in the user journey. Sales hears objections from prospective customers. Growth focuses on acquisition and activation, while another team may optimize revenue or retention. Each discipline observes real signals, but no single function sees the entire system.
According to Zhuo, a product leader’s job is to hold these truths together. That requires acknowledging the value of each contribution before attempting to reconcile the differences. People become more willing to consider unfamiliar evidence when they know their own expertise has been understood.
This approach is more rigorous than simply agreeing to disagree. The goal is to assemble a fuller picture from incomplete views and then connect that picture to the objective established in the first step.
Ask What Would Need to Be True
Once competing perspectives are visible, Zhuo recommends examining the assumptions beneath them. One useful question is: What conditions would have to exist for option A to be the right priority?
The reverse question is equally revealing. If someone favors option B, what evidence or circumstance would persuade them that A would better serve the shared goal?
These questions move a debate away from fixed preferences and toward conditional reasoning. Consider the three proposed investments:
An infrastructure rewrite becomes urgent if performance is deteriorating and that deterioration is causing customers to leave.
An onboarding redesign becomes more valuable if a large influx of new users is expected and too few currently reach the product’s core benefit.
A generative AI feature deserves priority if the technology significantly strengthens the value proposition and early users respond positively.
Extreme scenarios can help expose the logic. If nearly every new customer abandoned the product because onboarding was incomprehensible, most reasonable people would support fixing it. If severe latency were driving valuable accounts away, reliability would likely take precedence. The purpose of the exercise is not to exaggerate the forecast, but to identify the variables on which the decision truly depends.
Step Three: Convert Assumptions into Evidence Questions
After identifying the decisive conditions, the team can ask whether those conditions exist. Zhuo calls for turning the disagreement into questions that evidence can address.
For a performance project, the team might examine whether service quality has actually worsened and whether that trend corresponds with customer churn. For onboarding, it could review expected user growth, current retention, and relevant market benchmarks. It might also compare self-service behavior with the outcomes achieved when someone personally guides a customer through the product.
An AI proposal can be explored through prototypes and bounded experiments. Customer interviews may indicate whether the concept solves an important problem, while market research can show how competitors are moving and where differentiation remains possible.
Zhuo uses “data” broadly. Useful evidence may include product analytics, customer feedback, sales observations, research sessions, experiments, surveys, or informed reports from employees close to the problem. The objective is not to manufacture mathematical certainty. It is to replace vague conviction with information relevant to the underlying assumptions.
Evidence also makes collaboration easier. The team is no longer debating whose intuition deserves the most status. It is investigating a set of questions that its members helped define together.
Step Four: Choose the Right Person to Make the Call
Even excellent evidence cannot remove every uncertainty. Product decisions involve forecasts, and forecasts eventually require judgment. Zhuo describes this as a necessary leap of faith.
When the available information still supports more than one plausible path, she recommends moving the conversation up one level. The question becomes not only what should be done, but who is best positioned to decide.
The appropriate decision-maker is the person with the strongest combination of context, relevant skill, and sound judgment for that particular choice. It need not always be the most senior participant, nor should one individual automatically decide every issue. A leader can deliberately nominate the colleague closest to the customer experience, the technology, or the strategic tradeoff.
The selected person should be able to assess the evidence calmly and without excessive attachment to a preferred answer. Zhuo notes that leaders may nominate themselves when appropriate. What matters is that the broader group has had a genuine opportunity to contribute before ownership becomes singular.
This distinction prevents inclusive discussion from turning into ambiguous accountability. Many people can shape a decision, but someone must ultimately make it.
Step Five: Document the Decision and Watch the Replay
A careful process can still produce the wrong outcome. Zhuo therefore recommends recording what was decided, who made the call, which evidence mattered, and why the team expected the choice to work.
The team should then return to the record after an appropriate interval—perhaps a month, a quarter, or a year. The review is not merely a verdict on whether a metric increased. It is an opportunity to ask several deeper questions:
Which assumptions proved correct?
What did the result reveal about customers and the product?
Did the team measure success appropriately?
Was the decision process too slow, too hasty, or missing important evidence?
What does the outcome reveal about the decision-maker’s judgment?
A review may show that the original framing was flawed. What appeared to be an acquisition problem, for example, may actually have been a retention problem. The relevant definition of success may also change as the team learns more.
Keeping a decision history allows patterns to emerge. Over time, individuals can recognize recurring biases, leaders can improve how they allocate authority, and teams can refine the way they collect and interpret evidence.
Look Beyond the Original Menu of Options
Near the end of the talk, Zhuo introduces “mu,” a concept that rejects the premise of a question when neither the offered yes nor no is adequate. Applied to prioritization, it is a reminder that the initial list of choices may be too narrow.
The best answer might be a sequence rather than a winner. A small experiment could precede a major commitment. Further investigation might uncover a fourth option that serves the common goal better than any proposal on the original ballot.
Zhuo connects this broader thinking to Facebook’s early growth work. A small group was asked to address a particular growth challenge, but its testing and iteration eventually contributed to a wider, more data-informed operating culture. The enduring result was not merely one solved problem. It was a repeatable way of learning that spread across teams.
Influence That Outlasts a Single Decision
Zhuo’s process can be summarized as five linked practices: create common ground, combine partial truths, seek relevant evidence, appoint the right decision-maker, and learn from the outcome. Yet her closing point extends beyond process design.
Organizations will make thousands of decisions, and many individual choices will fade in importance. The relationships, judgment, and shared wisdom built while making them can last much longer.
For Zhuo, the deeper aim is to create a workplace where solving difficult problems together feels constructive rather than combative. Influence is therefore not about repeatedly winning arguments. It is about helping a group think more clearly, decide more responsibly, and become better collaborators through the work itself.
