The Stories Project

Mental-health need has been climbing worldwide, and the supply of clinicians hasn’t kept pace. Digital tools are the obvious lever, since they scale in a way therapists can’t, but they run into two walls. People don’t stick with them, and a lot of people won’t pick up something labeled as a mental-health intervention in the first place.

The Family History Technology Lab at BYU had a different angle. Family-history work is already a mass hobby. Millions of people do it voluntarily, for hours, with no stigma attached and no one to convince. And there is a body of research going back years suggesting that knowing your family’s stories is associated with better psychological well-being, especially in adolescents.

Nobody had tested the causal version of that directly with a digital product. Does reading a story about your own ancestor actually shift how you feel, right now, measurably, and more than reading someone else’s story would?

The answer needed a product before it could have an experiment. Stories signs in against FamilySearch and pulls together the memories and stories attached to a person’s ancestors, up to eight generations, into one place you can browse, instead of scattered across a tree you have to click through. People without an account can search an ancestor by name and read that family line’s stories without signing in at all.

I joined as a research assistant in the lab and led the engineering. The project had stalled years earlier with a half-finished site and no study; we rebuilt and finished it, got IRB approval, and ran the experiment. On the technical side the headline was performance. The original architecture made people wait, and waiting is fatal when the thing you’re measuring is how someone feels. Reworking the backend and the data access cut load times by 98%.

With the finding in hand, the question turned from “does this work” to “how do we get people to come back,” since a one-time mood lift isn’t much use if nobody opens it twice. That drove a round of features aimed squarely at engagement:

• On This Day — family events that fall on today’s date, which gives the site a reason to be worth opening daily.
• Favorite memories — flag the stories that land, and keep them somewhere you can find them again.
• Real search, sort, and filtering by topic across every story in a family line, including AI-generated topic tags, so the archive is navigable rather than just large.
• Featured ancestors — a surfaced person with stories to read, so an empty-handed visitor always has somewhere to start.
• A story-prompt card that suggests a close relative and asks you to write about them, turning a reader into a contributor, which is where the next generation’s stories come from.
• The published research, in the product, so the claim is inspectable rather than asserted.

Stories is live at stories.fhtl.org, the research is published and peer-reviewed at an international conference, and the lab is still building on it. I co-authored the paper and led the team’s engineering through the rebuild, the study, and the round of features that followed.

The thing I took from it doesn’t fit on a résumé line. It was the first project where I had to treat “is this actually working” as a question with a measurable answer rather than an opinion, and where a slow page wasn’t a polish issue but a threat to the result.