LinkedIn's product evolution: building complex marketplaces at scale

Executive overview

LinkedIn runs multiple interconnected marketplaces — hiring, learning, and content — that must work together without losing sight of one goal: connecting people to economic opportunity. Managing that complexity requires deliberate decision-making frameworks, a clear North Star that filters every trade-off, and product instincts tuned to second and third-order effects.

Hari Srinivasan, VP of Product for Talent Solutions, shares how LinkedIn approaches skills-first hiring, what job seekers can do to improve their odds, and how the company builds and maintains systems that shape behaviour at a billion-person scale.

The product's job is to amplify what people already want — not invent behaviour from scratch.

How LinkedIn's feed became more valuable

• Feed improvements trace back to one question: does this help someone connect to opportunity?
• Two signals drive what surfaces: in-network relationship content and out-of-network knowledge and advice
• Members consistently say knowledge and advice are what they seek — the algorithm is tuned around that
• GenAI-assisted prompts are an early-stage addition to help surface knowledge from nearly a billion members
• No single change drove the shift; small aligned decisions compound over time

Skills-first hiring

• COVID exposed a structural mismatch: hospitality workers losing jobs while customer service couldn't hire
• The blocker was title-based thinking — recruiters didn't see transferable skills across roles
• Skills-first hiring reframes the search: "can this person negotiate and de-escalate?" rather than "do they have this title?"
• Hospitality workers share roughly 70% of the skills needed for customer service roles
• 47% of recruiters now explicitly use skills when searching for candidates — and the figure is holding
• Job-seekers can follow the same logic: add credentials and work products against each listed skill on a job post

How values-based job search is emerging

• A growing segment wants to filter by purpose or interest, not just title
• LinkedIn launched collections and filters to support searching by topic area (e.g. AI) or values alignment
• The marketplace shift toward more candidates and fewer open roles makes differentiation more important
• For PMs specifically: industry experience (e.g. automotive background for automotive tech roles) is a strong differentiator when functional experience is common

What job seekers can do right now

• Turn on Open to Work — it is one of the highest-signal indicators recruiters look for
• Follow companies you want to work for and signal interest even before roles are posted; recruiters see that flag when a role opens
• Specify job types and preferences within Open to Work to reach the right searches
• Add skills with evidence — recruiters can scroll over a skill and see all supporting credentials
• Reach out directly to the hiring manager shown on a job post; LinkedIn surfaces this deliberately
• At senior levels (CPO and above), roles are usually filled through recruiter outreach, not applications — use LinkedIn to build long-term relationships

LinkedIn's operating model for complex systems

• Every decision filters through one North Star: "connecting people to economic opportunity"
• Members first is the explicit value that resolves conflicts when the ecosystem pulls in different directions
• RAPID framework assigns a single named decision-maker alongside recommenders, agreers, and input providers — clarity on the D (Decision) is the most important part
• Five-day escalation rule: if a PM can't resolve a disagreement in five days, it escalates to the next level; puts a clock on managers to unblock
• Second- and third-order effects must be modelled before shipping; interconnected systems mean a change in one marketplace ripples through others
• People who can see and simplify complex systems are rare — LinkedIn deliberately promotes for this skill

The PM skills triangle

• Great PMs don't need to be well-rounded; they need to live at an edge of the triangle
• Three vertices: creative/product instinct, data science and pattern recognition, general management
• Trying to compensate for weaknesses instead of doubling down on strengths leads PMs away from their best work
• Find a product people love early in your career — experiencing what that looks and feels like sets a permanent bar

LinkedIn Learning

• Entered the market in 2015 as the skill-development layer for the hiring marketplace
• Works with instructors to film, script, and produce professional-quality content — two full studio facilities
• Distributed primarily through enterprise; many individuals don't know it exists outside their employer
• Internal PM bootcamp (Product University) became the basis for Hari's public LinkedIn Learning course
• Key lesson from Product University: frameworks alone don't teach judgment — case studies, including failures, are required

Lessons from building products people love

• Products that feel invisible or embedded in a marketplace still drive outsized impact (skills infrastructure, Open to Work, LinkedIn Learning)
• Small process changes compound: 15-minute product reviews instead of hour-long ones; top-down "big rocks" planning before bottom-up team plans
• Building on the side keeps the builder muscle active — the habit of starting from a blank sheet is perishable
• Great product insight: a baby shark toothbrush turned tooth-brushing from a pain point into a moment of joy for a toddler — that's the standard