How LinkedIn became interesting: The inside story | Tomer Cohen (CPO at LinkedIn)
10 Sep 2024 (9 days ago)
Introduction to Tomer Cohen and his role at LinkedIn
- Tomer Cohen is the Chief Product Officer at LinkedIn, where he oversees all aspects of product development, design, business development, content creation, and customer operations.
- Cohen previously led the mobile team at LinkedIn and spearheaded the revamp of the LinkedIn feed, transforming it into a more engaging and social platform.
- He was instrumental in shifting LinkedIn's approach to an AI-first mindset, prioritizing the use of artificial intelligence in enhancing the platform's features and user experience.
The mantra “We might be wrong, but we are not confused”
- Clarity of thought and clarity of execution are equally important.
- Confusion within a system necessitates luck for success, while pulling in the same direction, even if wrong, provides a chance of success.
- The phrase "we might be wrong, but we are not confused" emphasizes clarity in thought and execution, originating from a startup founder facing imminent failure.
Clarity of thought and focus
- Clarity and focus are essential, and it's crucial to prioritize moving forward together over being right.
- When defining a problem, it's important to be specific and nuanced, envisioning the details of the challenge.
- Solutions should be based on first principles, with clear opinions and a willingness to make trade-offs.
Setting ambitious goals and overdelivering
- Ambitious goals should be set for products, even if they seem impossible based on current numbers.
- Instead of focusing on current limitations, product development should be driven by the potential impact and reach it could have.
- LinkedIn's feed is an example of this approach, where the initial goal was to create a platform used daily by millions of professionals, despite its initial limitations.
Transforming LinkedIn’s feed: strategy and execution
- LinkedIn’s feed was initially an activity feed that tracked network changes and became increasingly promotional.
- The transformation of LinkedIn’s feed involved shifting to an AI-first approach and redefining its purpose as a platform for knowledge exchange.
- The process involved unifying the AI team with the product team and focusing on objectives, algorithm features, and data training.
Running experiments at scale
- Running experiments at scale was initially challenging because any changes to the feed, which was the primary source of traffic, caused significant fluctuations in metrics, leading to anxiety among teams concerned about meeting their goals.
- To overcome this, a dedicated group of 2 million members was established as a testing ground, allowing for experimentation and iteration without impacting the metrics or causing disruptions for the broader user base.
- This approach proved successful, as significant positive behavioral changes were observed within the test group, providing compelling evidence to support the implementation of these changes across the entire platform.
Goal setting and identifying opportunities
- To avoid counting users who were not actively engaging with the LinkedIn feed, downstream engagement was used as a metric.
- Instead of focusing on past failures, the speaker believes in focusing on future possibilities and identifying opportunities for growth.
- The speaker believes in leading with conviction and identifying opportunities based on personal beliefs and aspirations, using evidence to support those beliefs.
AI’s role in LinkedIn’s evolution
- AI is crucial for matching user needs and content, acting as a sophisticated matchmaker to enhance engagement and value exchange on the platform.
- AI personalizes the feed, ensuring users see the most relevant and engaging content, leading to increased interaction and opportunities for both content creators and consumers.
- The use of AI has resulted in real-world benefits for users, such as increased economic opportunities, speaking engagements, and the ability to reach influential figures.
The AI-first mindset at LinkedIn
- An AI-first mindset is presented as a strategy where AI is considered from the initial planning stages of product development.
- Product leaders are encouraged to understand the core components of their AI systems, including the objectives, features, data collection strategies, and fine-tuning processes.
- The transition to an AI-first approach requires a shift in perspective for product leaders, moving from dictating specific user experiences to providing the necessary ingredients and guidelines for AI to create optimal outcomes.
Letting go of your roadmaps and allowing room for exploration
- Teams were initially instructed to disregard their existing roadmaps and instead, reconsider their objectives in light of the new technology.
- To foster creativity and exploration, teams were allowed to pursue similar ideas in parallel, even if it meant temporarily creating duplicates.
- After a period of open exploration, a top-down approach was implemented to focus resources on the most promising avenues, converging efforts based on the insights gained during the experimentation phase.
Career growth and personal insights
- It is important to find and learn from great mentors.
- It is crucial to identify personal passions and areas where one can make a meaningful impact.
- In product development, genuine excitement and conviction for the project are essential for significant impact.
Takeaways
- Prioritize work that excites and motivates you, even if it's not at the most prestigious company.
- When interviewing candidates, focus on their past accomplishments, what they learned, and the impact they made.
- When choosing between projects, prioritize those with greater potential impact, even if they seem less exciting or more challenging.
Lightning round and final thoughts
- There are no one-size-fits-all methods for building successful products or careers. Individuals should focus on their strengths and unique qualities rather than trying to excel in every area.
- The speaker recommends three fundamental books: "Mindset" by Carol Dweck, "Thinking, Fast and Slow" by Daniel Kahneman, and "High Output Management" by Andy Grove.
- LinkedIn is investing heavily in video content and encourages creators to consider their video strategy on the platform. The speaker also highlights LinkedIn's "coach experience," an AI-powered feature designed to support job seekers throughout their job search journey.