· PM Editorial · Product Sense · 6 min read
Improve LinkedIn Feed: Metrics and North Star
How to pick a defensible North Star metric and supporting metric tree for a LinkedIn-style feed, and defend it under interviewer pushback.
Why Metrics Questions Follow Feed Product Sense Questions
After you propose improvements to a feed product, interviewers almost always follow up with “How would you measure success?” or “What’s your North Star metric here?” This is a distinct skill from the initial product sense answer — it tests whether you can translate a product vision into a measurable, defensible framework. This article walks through choosing a North Star metric for a professional feed, building the supporting metric tree, and handling the trade-off questions that follow.
Step 1: Reject the Obvious Wrong Answer First
The naive answer is “Daily Active Users” or “Time Spent in Feed.” State clearly why you’re rejecting these as the primary North Star: raw DAU or time-spent metrics reward addictive, low-value content just as easily as high-value professional content, and for a network whose brand promise is professional relevance, optimizing purely for time-on-feed can actively damage the product by incentivizing engagement bait. Naming this trap and explicitly avoiding it is itself a strong signal to the interviewer.
Step 2: Propose a North Star — Meaningful Interactions Per User Per Week
A better North Star is Meaningful Interactions per Active User per Week, where “meaningful” is defined narrowly: a comment of more than a few words, a share with added commentary, a saved post, or a direct message initiated after viewing a post. This metric captures value delivered (the user found something worth engaging with substantively) without rewarding passive scrolling or one-tap reactions, which are cheap and easily gamed by bait content.
Defend this choice by contrasting it with the alternative of just “engagement rate,” which conflates a thoughtful comment with a drive-by emoji reaction. The North Star should be hard to game and correlated with long-term retention — meaningful interactions satisfy both conditions because they require actual reading and thought, and users who have them tend to come back.
Step 3: Build the Supporting Metric Tree
A North Star needs input metrics that explain why it moves. Break Meaningful Interactions per User per Week into three drivers:
- Session quality — average time spent per post viewed (a proxy for whether users are actually reading versus scroll-skimming), and the ratio of substantive comments to total comments.
- Content diversity — the number of distinct topics or creator clusters a user’s feed surfaces per week, guarding against filter-bubble stagnation that would eventually suppress meaningful interactions as users get bored.
- Creator retention — percentage of active creators (posted at least once in the last 30 days) still posting 90 days later, since a feed with no fresh supply cannot sustain meaningful interactions long-term.
Each driver maps back to a lever your product team controls: ranking algorithm changes affect session quality and diversity directly; creator tools (analytics, scheduling, guidance) affect creator retention.
Step 4: Add Guardrail Metrics
Every North Star needs guardrails to prevent optimizing it in a way that damages the business elsewhere. For this feed, guardrails include: total session time (should not collapse even as you shift from passive to meaningful engagement), ad revenue per session (monetization should not be starved by algorithm changes), and reported content rate (spam/abuse should not increase as a side effect of ranking changes). State explicitly that you’d watch these guardrails weekly alongside the North Star, and any experiment that improves Meaningful Interactions but tanks a guardrail by more than a defined threshold gets rolled back.
Step 5: Handle the Interviewer’s Pushback Scenarios
Interviewers will test whether you understand the metric’s limitations. Common pushback and strong responses:
“What if Meaningful Interactions goes up but users report the feed feels worse?” — Pair the quantitative North Star with a lightweight quarterly qualitative pulse survey (a single NPS-style question: “Does your LinkedIn feed feel professionally valuable?”) as a sanity check, since quant metrics alone can be gamed or drift from actual sentiment.
“Isn’t this metric slow to move, making it hard to run fast experiments?” — Use session quality and diversity as faster-moving proxy metrics for weekly experiment decisions, and only evaluate the full North Star on a monthly or quarterly cadence for strategic reviews, avoiding the trap of over-indexing on short-term proxy noise.
“How do you know ‘meaningful’ comments aren’t just longer engagement bait?” — Acknowledge the risk directly and propose a lightweight quality classifier (similar to the one described in the companion product sense article on improving the feed) to filter out comments that are long but still low-substance (e.g., generic “Great post!” variations padded with emojis).
Comparison Table: Candidate North Star Metrics
| Metric | What It Rewards | Gameability | Verdict |
|---|---|---|---|
| Daily Active Users | Any visit, regardless of value | High — easy to inflate with notifications/bait | Reject as primary North Star |
| Time Spent in Feed | Passive scrolling, addictive content | High — bait content maximizes this | Reject as primary North Star |
| Total Reactions (likes) | One-tap engagement | Very high — trivially gamed | Reject entirely |
| Meaningful Interactions per User per Week | Substantive engagement, real value | Low — requires actual reading/writing | Recommended North Star |
| Creator Retention Rate | Healthy supply side | Medium — can be gamed by lowering posting bar | Use as supporting driver, not North Star |
| Session Quality (time per post viewed) | Depth of attention | Medium | Use as fast-moving proxy metric |
Step 6: Tie It Back to Business Outcomes
Close the answer by connecting the North Star to the business: higher-quality meaningful interactions correlate with stronger member retention and higher willingness to pay for premium subscriptions (since engaged professional users are the core of LinkedIn’s subscription and recruiter-tool revenue), giving the metric a clear line to revenue even though it isn’t a direct revenue metric itself. This kind of metric-to-business-model connection is exactly the systems-level reasoning covered in the 100x Product Manager Interview Playbook (Amazon: https://www.amazon.com/dp/B0DBC1FQWH?tag=sirjohnnymai-20), which walks through North Star frameworks for several two-sided marketplace and social products.
Common Mistakes Candidates Make
The most common mistake is naming Daily Active Users or Time Spent as the North Star without acknowledging the gameability problem — this reads as surface-level thinking. The second most common mistake is proposing a North Star with no supporting metric tree, leaving the interviewer unable to see how you’d actually move the number. The third is forgetting guardrail metrics entirely, which signals you haven’t thought about unintended consequences of optimizing your own metric.
Final Answer Summary
For a LinkedIn-style feed, reject raw engagement metrics as the North Star and instead propose Meaningful Interactions per Active User per Week, supported by a metric tree of session quality, content diversity, and creator retention, protected by guardrails on total session time, ad revenue, and abuse rate. Be ready to defend the metric’s resistance to gaming, its connection to long-term retention and revenue, and how you’d use faster-moving proxy metrics for weekly experimentation while reserving the full North Star for quarterly strategic review.