Projects · a quick tour
Different subjects, the same way of working: clarify the question, build the useful system, check the result, and explain the boundary.
You do not need a technical background to follow these. Each project starts with a clear problem, shows what changed, and keeps the deeper method optional. Some have a measured outcome; others show the build and judgment honestly.
Marketing measurement and the day job are the closest fit. The job-search and video systems show operational range; Options and Sports show research discipline; Delve shows product and creative range.
Some projects produced a measured result. Some did not, deliberately. Each one shows the question, the build, what failed, and the judgment that followed.
Want the professional context first? Start with the day job, then come back to the projects.
The closest match to the day job: reporting, attribution, and decisions.
Checks whether reports counted conversions that arrived after the budget decision.
I built a plain-language way to compare what a report says today with what a marketer could have known when the budget decision was made. This is the closest match to my day job and the easiest place to start.
Corrected cost per customer was about 4× higher in the selected 25.
The deeper evidence trail behind Marketing Measurement.
The planned checks, supporting charts, and source record for readers who want the technical layer.
Systems that reduce repetitive work while keeping the decision human.
Narrows job postings to a shortlist worth reading.
It reads many sources and narrows the list against a written profile. The useful output is a smaller shortlist that I can actually read and evaluate.
683 roles screened; 57 worth reading first.
A clip-sifting and editing pipeline with an owner release check.
I play the games and choose what is worth sharing. The workflow handles the repetitive work of finding, formatting, and remembering what the channel has already used.
33,365 clips entered the funnel; 150 reached owner review.
Paper-only systems built to reject weak ideas before they become action.
A paper-only testing process designed to make ideas fail safely.
The system runs experiments, checks them against controls, and keeps the results away from real money. The useful story is the discipline around saying ‘not enough evidence.’
15,412 backtest results; 11 of 446 cleared the gate.
A testing process that writes down how a sports idea could be rejected.
This project turns a hunch into a fair test, records the answer, and refuses to call something a win just because the first result looked interesting.
814 questions logged; none cleared the bar in the August 21 paper snapshot.
A small product built around a clear return loop and honest release boundary.
A small daily dungeon game built around returning.
I wanted a game I would actually open tomorrow: a short session, meaningful choices, different ways to play, and a collection that gives the next run a reason to exist.
Five floors, four classes, six themes, and a Friday boss.
Want the professional context? Start with Experience and Marketing measurement.
Want the clearest story? Read Marketing measurement.
Want to see a working system? Open the Job Search Workflow or YouTube Workflow.
Want the deeper method? Open Measurement Method or How I Check the Work.
Want something visual and playful? The Delve page is a 13-frame screenshot tour of the real build.