Jake Heller

Product and strategy · 2020 to present

Track Record

Six years of product and strategy work at the messy, high-leverage intersection of monetization, growth, and things people actually pay for, now pointed at healthcare, where getting it right matters a great deal more than a conversion rate.

01

Roles, most recent first

Experience

Senior Product Manager

Citizen Health· San Francisco · 2025 to Present

  • Building Ari, an AI assistant that helps patients and caregivers in the rare disease community turn a mountain of medical records into clear answers.
  • Own AI quality and safety evaluation, the review frameworks, human grading rubrics, and automated post-release checks that catch regressions in a product where a confidently wrong answer has real consequences.
  • Run usage research on how the product is actually used, classifying what people ask an AI assistant to do rather than what we assumed they would. That analysis has redirected roadmap priorities more than once.
  • Own the unit economics of LLM features: what each capability costs to run, which ones quietly dominate spend, and when a prompt change beats adding infrastructure.
  • Shipped Citizen Health's mobile app, the first to bring appointment recording directly to patients, so nobody has to scribble notes while their doctor is still talking.

Senior Product Manager, Monetization & Billing

Rocket Money· San Francisco · 2023 to 2025

  • Ran 25+ experiments a year on price elasticity, funnel optimization, and billing, driving a 24% increase in average selling price, an 8% year-over-year reduction in churn, a 62.5% boost in reactivations, and a 34% uplift in post-onboarding conversions.
  • Simplified the in-house recurring subscription billing process, cutting quote-to-cash complexity and accelerating end-of-month close for accounting.
  • Automated 40% of billing-related support chats and improved in-app UI, lifting the app store rating from 4.2 to 4.4 and saving over $120K in support costs.
  • Expanded the monetization team's scope into partnerships, helping drive a 150% annual revenue increase.
  • Earned GitHub write access and opened 48 pull requests to fix copy and bugs, a PM shipping directly, which sped up cycles and made me much better company in code review.

Product Manager, Monetization & Billing

Rocket Money· Washington, DC · 2022 to 2023

  • Built and led a new team of full-stack engineers focused on post-onboarding premium conversion and billing, scaling operations as premium membership grew from 1.5M to over 4M paying members.
  • Led billing system improvements, resilience, tracking, and tech debt, cutting API costs by more than $660K annually while increasing collections.
  • Shipped a standout cost-and-revenue project that delivered a 6% increase in operating income.

Associate Product Manager

Rocket Money (formerly Truebill)· Washington, DC · 2021 to 2022

  • Led the Truebill → Rocket Money rebrand post-acquisition, aligning cross-functional teams to execute a unified rebrand across the app, the business, and advertising.
  • Owned Net Worth Tracking and Credit Score Monitoring, taking Net Worth Tracking from a 40K-user experiment to 2M users.

Analyst, Innovation & Strategy Consulting

Innosight· Boston, MA · 2020 to 2021

  • Conducted market-driven Jobs to be Done analysis to uncover high-impact growth opportunities and address unmet customer needs.
  • Designed and validated business cases with detailed financial modeling to assess feasibility and ROI for Fortune 500 investments.

03

How the work gets done

How I Work

Prototype First

Show, don't spec

I build quick, throwaway prototypes with tools like v0, Replit, and Lovable before writing a long document. High-fidelity visuals make conversations with design and engineering faster and more honest. They don't replace designers or engineers, they just mean I show up prepared.

In The Codebase

Ship the small stuff yourself

At Rocket Money I had write access and opened 48 PRs for copy fixes and small bugs. A PM who can clear their own paper cuts spends fewer meetings negotiating over them.

Jobs To Be Done

Start with the job

Trained on JTBD at Innosight and still use it. Understanding the job someone is trying to get done beats guessing at features, and it survives contact with real users better than a roadmap does.

Evaluation

Read the raw transcripts

Dashboards tell you a number moved. Reading the actual output tells you why, and it is the only way I have found to catch a model being confidently wrong. I work through evaluation transcripts myself rather than trusting a severity score I have not checked.

Own The Query

Pull your own data

I write my own SQL against the warehouse instead of queueing behind an analyst. It makes me faster, it makes my questions better, and it means I know exactly how a metric was defined before I quote it.

Unit Economics

Know what it costs to run

With LLM features, cost per user is a product decision, not a finance problem. I track spend per capability so we can tell the difference between a feature that's expensive and one that's broken.

Magna

Cum Laude · BS Business Management, Babson College

2023

Amplitude Pioneer Award · Revenue Monster winner

2017

Babson B.E.T.A. Challenge winner

2015

NFIB Young Entrepreneur of the Year finalist

04

Unpaid, ongoing

Volunteering

Search & Rescue

CALSAR field member

Volunteer with California Search and Rescue, responding to backcountry emergencies under Cal OES. I've also built small internal tools for the team, a callout dashboard that answers where, how long, what weather, and is there cell service.

CaliforniaOn call

Teaching

Stanford MS&E 265

Team mentor for Introduction to Product Management at Stanford, helping students work through their first real product problems.

StanfordMentor

Entrepreneurship

NFTE coach

Coach with the Network for Teaching Entrepreneurship, working with students building their first business plans.

NFTECoach

Let's Talk

Hiring, advising, or comparing notes on AI evaluation and monetization. Happy to chat.