Oliver Chan

I was responsible for Scotiabank's piece of Scene+, the bank's loyalty rewards program. That meant hundreds of user stories and dozens of sprints, with a national launch date that wasn't moving no matter what came up. It shipped on time, which in banking is its own kind of achievement.

Full write-up (scope, the coordination problem, my specific role) still to come.

This one started as a blank page. I took Scotiabank's electronic FX platform from nothing to a live MVP in six months, then to the full vision within about a year. I did the database schema myself, all the wireframing and hi-fidelity UI, and ran the internal go-to-market, mostly because there wasn't anyone else lined up to do it. It's still pricing FX across every Scotiabank retail channel today.

Full write-up (zero-to-one process, design decisions, what I'd change) still to come.

The app had been quietly growing for years, and the thing that worried me wasn't disk space in the abstract, it was retention. People delete large apps to free up room well before they actually run out of it, so a ballooning install size is a churn risk wearing an engineering metric's clothes.

The harder question was what target actually mattered. "As small as possible" isn't a plan, so I went and looked at how people actually run into app size on iOS: open iPhone Storage settings and apps are listed largest first, but only the top 20 show, the rest hide behind a "Show More" button. Most people never see an app flagged as a space hog unless it's big enough to crack that list. So the target became staying under that fold, not chasing an arbitrary minimum.

To turn that into an actual number, I commissioned a statistical analysis, with some LLM help, on installed-app-size distributions: what size would a typical device's apps need to clear to land in that visible top 20 (roughly the 95th percentile, two standard deviations out). The answer came back around 440MB, so that became the number everyone argued about in planning meetings for the next while.

Full write-up (the size-reduction levers, how the target reconciles with the ~600MB delivered figure, what shipped vs. what's still in flight) still to come.

I owned product for Interac e-Transfer and Add Bill Payee across Scotiabank's whole retail base. The team's early instinct, understandably, was to optimize the single transaction: fewer taps, faster confirmation, a higher completion rate on any one transfer or payment.

I pushed back on that framing. A single successful transfer doesn't make a customer sticky, a recurring one does. Once you've set up a recurring e-Transfer or a scheduled bill payment through your bank, switching banks means rebuilding all of that somewhere else, which is real friction, and it's friction working in the bank's favour. So we shifted priority from "make one transfer easier" to "make setting up a recurring transfer at least as easy as a one-off," on the bet that recurring usage, not raw transaction count, is what actually keeps people around.

Full write-up (the recurring-vs-one-off product bet, what shipped, what the data showed) still to come.