Tim McSween Work Authentication & Login Intelligence
Product Intelligence

Authentication & Login Intelligence

Turned dormant telemetry into actionable product decisions — surfacing MFA gaps, recovery drop-off, and filing-season load patterns across 65K+ preparers.

25M+ Login events analyzed
22% MFA adoption surfaced
32% Password recovery conversion
53K Weekly active users at peak
Context

Drake Tax's authentication module handles every preparer login — password validation, MFA, account lockout, and password recovery. Rich telemetry had been instrumented over several years, but none of it was being surfaced for decision-makers. The data existed; the questions existed; nobody had connected them.

Role

Mapped every instrumentation point in the authentication module to its Application Insights schema, designed stakeholder-driven analytics, and delivered a comprehensive report covering 25M+ login events across the full preparer base.

Challenge

Leadership had no visibility into MFA adoption trajectory, password recovery effectiveness, or how filing-season load patterns behaved — despite having instrumented data available. Without this, security initiatives couldn't be validated and the recovery UX had no evidence base.

Approach
  • Structured the analysis around questions stakeholders would actually ask — not around raw event taxonomy.
  • Reverse-engineered every telemetry instrumentation point in the module, creating a precise schema of available dimensions: login outcomes, MFA setup actions, recovery funnel steps, failure reasons.
  • Built targeted KQL queries for each analytical section across a 90-day window capturing the full filing season arc — ramp-up, April 15 peak, post-season decline.
  • Delivered an interactive stakeholder report with Chart.js visualizations, replacing ad-hoc data requests with a repeatable, shareable artifact.
  • Productized the reporting as a reusable pipeline — any engineer or AI agent can regenerate it on demand for any date range.
MFA was being disabled 2.3× more often than enabled — a net-negative trend invisible without telemetry. The finding directly informed a reassessment of the MFA rollout strategy.