About
I help small software teams answer questions with their own data.
Product analytics and the software around it. One person, fixed scope, written deliverables.
What I do
I build the measurement layer for software products: the metric definitions, the event schema, the models that turn events into numbers, and the dashboards that keep those numbers honest. When the fastest path to an answer is a query or a script instead of a dashboard, I write the query.
I have run product analytics in-house at two software companies and built an internal data tool for a fintech advisory as an advisor. Those three are the case studies on this site, each labeled as what it was.
Gold Street Software is a solo practice that takes one or two clients at a time. That shapes the work: fixed scope, async by default, few meetings, and a written deliverable at the end of every engagement. It also means I turn down work I cannot do well in that shape.
What I work with
The list is deliberately specific. If your stack is not on it, I will say so on the fit call.
Analysis
- Product analytics
- Product management
- Dashboard and KPI design
- Event schema design
- Churn and cohort analysis
- Pricing-tier analysis
Data
- SQL
- Python
- dbt
- BigQuery
- Databricks
- Stripe data
Tools
- Mixpanel
- GA4
- Google Tag Manager
- Segment
- Looker
- Power BI
Software
- TypeScript / Next.js
- Playwright and test automation
- Claude and OpenAI APIs
- Data-first MVPs and internal tools
How I think about the work
Durable answers and answers now are different jobs.
Some questions need a dashboard. Most need a query, a short write-up, and a conversation. I will tell you which one you have before anything gets built.
AI only where it earns its keep.
I use language-model APIs when they save real time on a task that is already measured. I do not propose them as a strategy.
Systems that outlast me.
Everything I build is documented and handed off. You should not need me to keep it running.
What I do not do
Unbounded product-management retainers, build-anything software development, or AI strategy work. The Steward engagement is a fixed block of analytics hours, not a part-time product manager. If the data cannot support the analysis you want, I will say so and deliver a measurement plan instead.
Start with a fit call
Tell me what you are trying to understand or fix. I will tell you whether one of the fixed-scope offers fits, and if it does not, what I would do instead.