LinkedIn · Product sense · Mid · 45 to 60 minutes

Expand My Network beyond People You May Know

LinkedIn is a professional networking platform where people build work identities, connect with others, and discover jobs and professional content. Answer this product management case and get it graded line by line against a weighted rubric.

Product sense
UX judgment
Metrics
Platform thinking
Stakeholder alignment

The situation

You are the PM for discovery in LinkedIn's My Network area on mobile and desktop. Today this surface mainly helps members connect with people through People You May Know. Leadership wants to use this high traffic area to also help members discover other parts of LinkedIn that can build communities around shared interests, such as groups, pages, hashtags, newsletters, and future recommendation types. Several partner teams are ready to supply recommendations, but each uses different ranking services, response formats, and tracking events. Design is worried that adding more choices could make the experience harder to scan, especially across languages with longer text. Engineering wants a plan that scales so new recommendation cohorts can launch quickly without heavy client work on web, iOS, and Android every time. Your GM asks you for a product proposal that defines the member problem, recommends an experience direction, explains how you would measure success, and calls out the minimum platform and instrumentation needed to support experimentation and future expansion without hurting the existing connection use case.

What to produce

A 1 to 2 page product strategy memo.

Your answer must

Define the target member problem and explain why My Network is the right surface for it.
Recommend a UI and interaction approach for exposing multiple recommendation types and letting members explore more, with key tradeoffs.
Specify one primary success metric, at least two guardrail metrics, and an experiment plan to validate the concept.
Describe the minimum platform requirements for aggregating and ranking recommendations from multiple backend systems over time.
Propose a shared tracking model across recommendation types that supports analysis, debugging, and future model training.

Constraints

The solution must work across desktop web, iOS, and Android.
People You May Know remains an important existing use case and cannot be materially degraded.
Recommendation sources are owned by different teams and return different schemas, rankings, and metadata.
Frequent client releases are slower than server side changes, so the plan should minimize dependency on app store timing.

Format

Use sections for member problem, proposed experience, metrics and experiment, platform and tracking, rollout, and risks.

How LinkedIn actually played it

LinkedIn tested tabs on top of its People You May Know feed to surface groups, pages and newsletters, and quickly saw tab-blindness where members did not notice them. The post explains how the team reworked the design for discovery across several recommendation types. A good case of a failed first design guiding a better one.

Early tests can reveal discoverability problems fast.
Design for international text length early.
Different recommendation types need a shared tracking model.

Built from Helping members discover communities around interests by Chiachi Lo, Bohong Zhao and Elina Lin.