Design an experimentation dashboard for Sales Navigator at global scale

Problem Statement Description

Product context: LinkedIn is Microsoft's professional network; its products include profiles, feed, jobs, recruiting, LinkedIn Learning, sales tools, messaging, and ads.

LinkedIn Sales Navigator teams run experiments to improve how sales professionals discover accounts and leads, learn prospecting workflows, engage with their network, and convert usage into measurable sales outcomes. At global scale, these experiments may affect individual sellers, sales teams, enterprise accounts, CRM-connected workflows, and LinkedIn member experiences, making measurement more complex than a simple consumer-product A/B test.

Design an experimentation dashboard that helps product managers, data scientists, engineers, and business stakeholders understand whether Sales Navigator experiments are healthy, trustworthy, and decision-ready. The dashboard should support experiments across regions, customer segments, account sizes, seat types, onboarding maturity, and usage patterns, including users who are still learning how to use Sales Navigator effectively.

The problem is not just to display experiment results, but to define what results should mean in a B2B product where value can appear through activation, retained usage, lead/account discovery, CRM actions, collaboration, and eventual monetization. The dashboard must also make it clear when an experiment result is invalid, inconclusive, risky, or not generalizable across global customer cohorts.

The experience should consider:

- Core success metrics for Sales Navigator experiments, including clear numerator, denominator, event source, and interpretation.

- Experiment units and exposure logic, such as member-level, seat-level, team-level, or account-level assignment, and how contamination should be detected.

- Cohort breakdowns by geography, company size, industry, sales role, subscription tier, onboarding stage, CRM integration status, and new versus experienced users.

- Instrumentation quality, including event coverage, logging delays, missing data, duplicate events, and consistency across web, mobile, email, and CRM-connected surfaces.

- Guardrail metrics for member trust, spam-like behavior, network quality, customer satisfaction, latency, unsubscribe/block actions, and enterprise admin concerns.

- Statistical and operational readiness, including sample size, duration, confidence, seasonality, ramp status, experiment overlap, and long-cycle outcome lag.

- Decision usefulness for stakeholders, including whether to ship, iterate, stop, extend, segment rollout, or investigate further.

- Global scalability needs, such as localization, privacy expectations, regional compliance, role-based access, and executive versus practitioner views.

Your goal is to describe the metrics framework and dashboard experience that would allow LinkedIn to evaluate Sales Navigator experiments reliably at global scale, while helping teams make faster and safer product decisions without over-reading noisy or incomplete data.