Support Operations Data Analyst

User Operations runs on data — but right now, that data lives in too many places, speaks too many languages, and reaches the wrong people too late. This role exists to fix that.

As Harvey's first Support Operations Data Analyst, you'll own the analytics function for the User Operations org. You'll build and maintain the dashboards, reports, and feedback loops that tell us whether we're hitting our north stars — cSAT, TTR, QA scores, escalation rates — and surface the signal underneath the numbers so we can act on it. You'll sit within the Support Operations team, reporting to the Support Operations Manager, and work closely with User Operations leadership and Harvey's central data team to ensure the org is equipped with the right instrumentation as we scale.

This is a solo role. You won't have a team beneath you. You will need to be fluent enough in support analytics to hold the function independently, confident enough to push back on how metrics are framed, and fast enough to operate at Harvey's pace.

Own recurring reporting for User Operations — weekly, monthly, and QBR-ready — tailored to ops, leadership, and cross-functional audiences

Translate support data into clear narratives: what's happening, why, and what to do about it

Track and maintain north star metrics: cSAT, TTR by tier, QA scores, bug escalation rate to EPD, and First Response Time

Build and maintain self-serve dashboards that give the ops team and leadership real-time visibility into support performance

Partner with Support Systems to ensure Zendesk is instrumented to capture the data we need

Work with Harvey's central data team to connect support data to broader product and customer data sources

Identify and close data collection gaps — if we can't measure it, help define how we should

Design feedback loops that connect support signals to Product, Engineering, and Customer Success

Quantify the operational cost of product bugs, feature gaps, and onboarding failures

Contribute to QA analytics as the QA program matures

Track ticket deflection, AI/chatbot performance, and self-service effectiveness

Measure the impact of AI-driven support — containment rate, escalation rate from AI interactions, resolution quality — and surface findings that drive how we tune and invest in those tools

Support ad hoc analytical requests from the Support Operations Manager, User Operations leadership, and senior stakeholders

3–5 years of experience in analytics, with at least 2 years directly in support operations, customer success operations, or a closely adjacent function

Fluency in support platform data — you know how Zendesk (or equivalent) is structured, what data it produces, and what it doesn't

SQL proficiency — you can write complex queries against large datasets without hand-holding (CTEs, window functions, joins across schemas)

Dashboard experience — you've built and maintained operational dashboards in Looker, Tableau, Sigma, Omni, or equivalent

Reporting for multiple audiences — you know the difference between what a frontline manager needs and what a CFO needs, and you build accordingly

Strong data storytelling — you don't just present numbers, you write the narrative

Comfort operating solo — you don't need a team around you to deliver, and you don't need a ticket to tell you what to look at

Experience with Python for data manipulation or automation

Familiarity with dbt or similar data transformation tooling

Experience building or contributing to QA analytics programs

Background supporting enterprise SaaS or AI-native products

Experience working with Zendesk APIs or extracting data beyond standard reporting

AI-native: you use AI tooling actively in your analytical workflows — not as a novelty, but as a force multiplier

Pace: you move in hours and days, not weeks. You surface findings before anyone has to ask

Judgment: you know which metrics matter and which are vanity. You push back when framing is wrong

Clarity: your outputs are direct, jargon-free, and actionable. You write for the reader, not yourself

Ownership: you treat User Operations analytics as your problem to solve, not a ticket queue to process