Snapshot. This is less a single feature and more how I operate. At Talisman I ran prioritization, the specs, the release process, and the quality bar, and we shipped the Expense Management product on top of that operating system. I made trade-offs with data instead of opinion, wrote the documents the team built from, and blocked releases when they were not ready.

Why this matters

Most PM JDs ask for the same thing in different words: own the roadmap, decide with data and not stakeholder loudness, write clear briefs, ship incrementally, and instrument success. This is the case study that shows the operating system behind the features.

Prioritization with data, not opinion

I ran a RICE tracker so release decisions came from reach, impact, confidence, and effort rather than whoever argued hardest. I also cut scope on purpose. In one release I proposed removing several graphs and columns, login counts and users-per-month, because they did not give end users value, so the release stayed focused on what mattered.

Writing the product

I wrote the documents the team built from, and I codified the processes so they outlived any one project: a PRD process, a story-map process, and a release-tracker process. The specs were not after-the-fact paperwork. They defined the problem, the solution, and how we would know it worked. I also wrote in a working-backwards style, leading with the problem and the user before the solution, for example a brief on how Talisman should handle negative-value transactions, sent to the team with a clear review deadline.

Shipping incrementally and holding the line

I held the team to agreed release lists. When tickets got worked out of order, I named it and reset the expectation that we follow the plan we committed to.

I also owned the quality bar. I halted a production push when two real issues surfaced in morning testing, with a plan to resolve them before shipping. And when tickets reached QA with no user scenarios to test against, I flagged that we had no way to call something done and stopped them from sliding through.

Outcome

On top of this operating system we shipped the Expense Management product, then launched it.

External recognition:

  • #1 SaaS Product of the Week and #2 Product of the Day on Product Hunt (February 2024)
  • This Week in Fintech, Best New Startup / Fintech Startup of the Year (2024)

What shipped (a selection from the roadmap I prioritized and ran to release):

  • Insight surfaces across the App and Member profiles: summary, activity, team, and member-level insights
  • An Alert Center with email alerts, covering new transactions on existing subscriptions, newly detected subscriptions, renewal reminders, re-authentication needs, and new app and login detection
  • A redesigned Expense Overview with spend categories and spend projections
  • Subscription management: a guided create-subscription flow, missing-transaction capture, and highlighting of subscriptions with new activity
  • Onboarding connections and guided tours for Plaid, QuickBooks, and Google Workspace, including reworked loading screens
  • A Renewal Calendar and a Home dashboard
  • Amplitude event instrumentation, so we could measure what users actually did and decide what to build next
  • Responsive breakpoints across the product

Each of these came off the RICE-prioritized roadmap, was specced, shipped in releases, and held to the quality bar above.

Artifacts

Held in private records:

  • RICE prioritization tracker
  • PRD, story-map, and release-tracker process docs
  • QA and testing process doc
  • Release trackers

Reflection

Organization is, always has been, and always will be a PM’s best friend.