Begin Learning promised parents one learning journey that would carry a child from age 2 to 10, sold as a premium bundle. Begin had built that lineup through acquisitions, including codeSpark, where I worked. The products had almost nothing in common except that they were for kids. They covered different ages and topics and ran on different tech.

Asking parents to buy all three apps and a physical subscription on their first purchase was too big an ask, and few did.

With our Head of Product, I planned a series of steps toward a single learning journey. I called it Project Secret Tunnel.

Step one was the Begin Learning app, a sampler with content from Learn with Sesame (built in Phaser), HOMER (Pixi, HTML5, and React), and codeSpark (C# and Unity), plus two new games that were the first digital content for Little Passports. Reusing existing content kept the cost low. It let us test whether a combined app was easier to sell, and it forced us to solve the hardest technical problem first: running all of that content inside one Unity layer. Parents could buy any of the full apps from inside the sampler.

Step two used that Unity framework to rebuild the legacy apps. Learn with Sesame went first because it was the smallest. HOMER followed with a full overhaul of everything outside its content. Both moved to shared login, purchasing, and entitlements, which meant major backend work on codeSpark as well.

The sampler launched with little support and still earned a 4.3-star rating. Parents who bought through it had 35% higher lifetime value than parents who bought a single app.

When the company later went bankrupt, the shared tech stack meant a skeleton crew could keep every product running.

Apple announced the end of IDFA, the identifier that let app companies follow an ad from impression to download to purchase. Every new release had to drop it, and older apps that still used it would eventually be pulled from the App Store.

codeSpark Academy was a live-ops subscription business. If we shipped an update, we lost our ability to measure acquisition. If we stopped shipping, subscribers stopped seeing anything new and canceled. Either way, we would lose users fast.

codeSpark Academy already let kids make their own games and stories, including stories with their own pictures and voices. I had tested a small campaign for an item pack using our Story Coder, with professional voice actors and complex code that showed what the tool could do. Kids loved it.

I proposed building on that with Fooville Legends: ongoing, official stories starring The Foos that taught coding and social-emotional lessons, with events kids could take part in. We surfaced them in the app’s news section and built contests and challenges around them. Everything shipped from the server instead of the App Store, so we could keep IDFA in the app as long as possible.

Meanwhile, the team built Adventure Game Maker, a third creative coding tool, so the release that finally dropped IDFA would also give players a big new reason to come back.

Fooville Legends ran for six seasons, with more than 120 stories and 8 million views. Players who engaged with it were 80% more likely than unengaged players to still be subscribers every month for the next 13 months.

We held our user base while other apps struggled, and Begin went on to acquire codeSpark in an eight-figure deal.

I worked for a free-to-play publisher that brought games proven in Eastern markets to Western players. 7 Dragons was our first big browser launch, and it arrived with translations that made no sense, repetitive gameplay, dated graphics, and no voice acting.

It was also a sequel to a game we didn’t have the rights to, Age of Seventh Dragon II. And the lore had only six dragons, each with its own faction. There was no seventh.

I couldn’t change the gameplay or the art, so I changed the story. On Luvinia Online and Dark Blood Online, I had rewritten the dialogue and brought in professional voice actors to give each game its own sound. 7 Dragons had no voice work, so the writing had to carry everything.

I rewrote the entire story bible and made the dialogue read naturally. Then I changed the premise. Players weren’t just fighting to put one of six dragons on top. They were fighting to become so powerful that they would become the seventh dragon themselves. I built repeatable, biweekly live events around that premise.

The game grew from zero to more than 1 million players in its first year.

One complaint kept showing up in parent feedback for codeSpark Academy: parents weren’t sure their kids were learning anything. It didn’t matter how much curriculum was in the app or how many teachers used it in class.

Parents asked us to explain how their child was learning, so we built exactly that. Parent Academy gave every parent a portal with context, lessons, and tips written, and sometimes read aloud, by our curriculum lead. Almost nobody used it. Adoption was close to zero.

I brought together customer service, product, design, and curriculum to find out why. Parents hadn’t been describing the real problem. Much of the learning happened while kids played other kids’ games, and parents who saw that assumed it was just play. Many also believed only puzzles counted as learning.

So we stopped explaining and changed what parents saw. We reworked the puzzles to build lessons step by step. We added badges for working through them, and each badge unlocked coins, codeable items, or avatar costumes, so what kids learned showed up in what they made. And I made Fooville Legends, our ongoing story series, part of the answer. Its contests gave away physical prizes, so kids talked to their parents about what they were entering.

Monthly active users grew 25%, and kids made 75% more games and stories year over year. Parents got involved in their kids’ learning without ever having to open the app.

codeSpark Academy was designed for kids ages 5 to 10, but we kept getting players younger and older than that. We wanted to pick one group and build for it.

Preschool was the riskier bet. It wasn’t clear that kids that young could benefit from computational thinking education at all, and building for them without proof would have been hard to justify.

We found a National Science Foundation opportunity to fund content for preschoolers, and we won the grant. We partnered with RAND to run a randomized controlled trial, so we would know whether the games actually taught anything, not just whether kids enjoyed them.

As Senior Producer, I produced three of the five games, including the title game, Pick That Pattern, which taught kids to recognize and create patterns, Paper Shaper, which taught shapes and spatial reasoning, and Doodle Looper. On the other two, Foo Rescue and Road Repair, I brought in a third party for the music and cast and directed the voice recording.

The study showed, with statistical significance, that kids in Head Start programs who used codeSpark Pre-K improved in computational thinking and math.

Paper Shaper, the final game and one I produced, won a Silver Collision Award.