How LinkedIn became interesting: The inside story | Tomer Cohen (CPO at LinkedIn)

Episode Summary

Executive Summary: Tomer Cohen explains how LinkedIn shifted from a promotional, notification-style product into a high-value professional network by clarifying purpose, using AI as the core matching engine, and running focused experiments like a 2M-member cohort. He emphasizes clarity over confusion, conviction over hedging, and AI-first product thinking as essential to LinkedIn’s turnaround and future.

Main Topics: Clarity over confusion as a leadership principle (Priority: 5/5): Tomer’s mantra, "we might be wrong, but we are not confused," centers on aligning teams around a clear problem, clear direction, and clear execution rather than endlessly hedging or arguing. He ties this to better decisions, faster learning, and stronger organizational focus. LinkedIn feed turnaround and new product purpose (Priority: 5/5): He describes how LinkedIn redefined the feed from an activity/promotional stream into a destination for professional knowledge exchange—content from people that matter, on topics that matter professionally—driving a major product and culture shift. AI-first product leadership (Priority: 5/5): Tomer argues AI is a mindset and operating model, not just a technology. Product leaders should own algorithm objectives, features, data strategy, and fine-tuning, with AI integrated into strategy, talent, and roadmap decisions. Using focused cohorts and experiments to de-risk change (Priority: 4/5): To overcome internal resistance and metric instability, LinkedIn carved out a 2 million-member cohort to test the new feed experience, proving the model before scaling it to the broader product. Economic opportunity and professional value exchange (Priority: 4/5): LinkedIn is framed as an economic opportunity platform built on a social graph. The feed, search, jobs, and learning all work best when they enable high-quality value exchange between creators and the right audience. Career growth through conviction and strength-based development (Priority: 3/5): Tomer shares that his own advancement came from pursuing what he believed in, learning from great people, and focusing on impact rather than titles, logos, or conventional prestige. Future of AI and human-AI relationships (Priority: 3/5): He closes by stressing that AI will become deeply personal and intimate, especially in products like LinkedIn’s job-seeker coach, and suggests the relationship between humans and AI will become increasingly central and meaningful.

Key Arguments: Organizations fail when they are confused; clarity of thought and execution allows teams to pull in the same direction and learn from outcomes. A product team should define the exact problem, audience, and tradeoffs before building solutions, and solutions should reflect strong principles with real "teeth." The LinkedIn feed succeeded because it stopped being a traffic/upsell mechanism and became a destination for professional knowledge exchange and reputation-building. A turnaround product requires internally changing the system’s incentives and metrics, not just shipping a new UI. AI should be treated as the product engine and owned by product leadership, not delegated as an engineering-only black box. Product leaders need to ask algorithm questions directly: objective, features, data collection, fine-tuning, and infrastructure. LinkedIn scaled the feed by testing with a 2M-member cohort, proving the experience improved before expanding it. High-quality matching matters more than raw engagement volume; LinkedIn is willing to trade off bad engagement to improve professional relevance and safety. Career growth comes from conviction, impact, and learning, not from chasing titles or external prestige. AI-first product development means starting from the objective, then using AI to solve it better—not starting from the technology and inventing uses for it.

Data Points: LinkedIn hires per minute: 7 - Tomer cites this as evidence of LinkedIn’s core economic opportunity value proposition. LinkedIn learning activity per minute: 140 hours - He mentions this to show the scale of learning and knowledge exchange happening on LinkedIn. Cohort size used for feed experimentation: 2 million members - A carved-out group used to test the revamped feed experience before broader rollout. Career progression examples: Senior PM to CPO in not that many years - Describes his relatively fast internal rise at LinkedIn through multiple product leadership roles. AI-first rollout year: Early 2020 role; strategy shift started around 2016 and intensified in fall 2022 - He explains the timeline of AI-first adoption inside LinkedIn. Public awareness of ChatGPT: March 2023 - He says LinkedIn had already started its new AI wave in fall 2022 before broader market attention. LinkedIn member scale: 1 billion members - He uses this as a reminder that the platform still has substantial growth potential. Job seeker coaching availability: Top of the Jobs tab - He says the AI coach experience is now prominently accessible in LinkedIn Jobs.

Pivotal Quotes: "We might be wrong, but we are not confused." — Tomer Cohen: His core leadership mantra about clarity, alignment, and avoiding organizational hedging. "This is not a springboard for other products. This is not a traffic jumpstart. It’s really about people that matter, talking about things that I care about professionally." — Tomer Cohen: Defines the new purpose of the LinkedIn feed during its transformation. "AI is the ultimate matchmaker. It’s underutilized. It’s misunderstood." — Tomer Cohen: Summarizes his view of AI’s role in marketplaces and professional networking.

Implications: For PMs and leaders, the lesson is to define a sharp objective, own AI deeply, and use focused experiments to prove change. For LinkedIn, the future is higher-quality professional matching, stronger creator value, and more personalized AI assistance.