Case studies

The challenge

ZIYX had built a new version of their app. They needed to know if it worked for their target users. These were young people not in work, education or training, who hadn’t gone to university.

The app was built on guesses about how people would use it. That included a 3D metaverse, which was a big investment.

ZIYX needed three answers. Did the onboarding work? Did the core experience hold this audience? And was the metaverse worth building further?

What I did

I ran a survey with 1,200 people, all from the target group. They came through four email campaigns with partners. I found and secured two of those partners as part of the project.

I also interviewed 10 people about how they moved through the app.

The first round led to a full redesign of the onboarding. The 3D experience became a 2D questionnaire, laid out as a journey along a road. This gave users a sense of progress.

A second round then tested the new version.

The results

Users spent three times longer in the 2D version than the 3D version. Time spent and positive reviews both showed the simpler version was far more engaging.

The research led to a clear recommendation: drop the metaverse and focus on the core 2D experience. The metaverse was costly, and the basic experience hadn’t been tested yet.

The data made that case clearly. The work led to two new versions of the app.

The challenge

Researchers in trauma and adversity struggled to find work related to their own. Some were even repeating studies that already existed, because they hadn’t found them.

The same ideas were listed under different terms across the main academic databases. Someone searching “Adverse Childhood Experiences” on PubMed might miss thousands of papers. Those papers could be listed as “Early Life Stress” or “Childhood Maltreatment” on Google Scholar.

About 20 different terms were in use across the field. Nobody had mapped how they linked together.

What I did

I mapped 19 key terms across four major databases: Google Scholar, PubMed, PsycINFO and ERIC. For each term, I recorded the number of results.

Then I worked out the overlap between each pair of terms. This showed how much research someone would miss by searching with only one term.

The result was a grid showing where terms differed across databases, and how much unique research sat behind each one. ANSResearch then built this into their own database as a “related terms” feature.

The results

Some pairs of terms had zero overlap. So a researcher using one term would miss all the work filed under the other, even when both meant much the same thing.

This showed how big the field was, and how split up it had become.

The related terms feature helped a team of about 30 researchers, including partners at Harvard. They could search more fully without knowing every version of a term. This cut the risk of people repeating each other’s work.