Zya Store: Lessons learned by doing it wrong.
2010: I was the lone UI designer, tasked with designing the browser-based catalog from which people would purchase characters to create music in the Zya application.
We made educated guesses how many items would be available in the store. Execs were confident we needed to support hundreds of characters, representing all musical genres, but only a few had been developed so far. Would we also sell instruments? Outfits? Songs? Bundles? Also, the company hadn't come to conclusions about in-app currency, pricing, advertising, and so on.
I suggested interviewing our audiences, narrowing the target, and doing a detailed effort estimation for inventory. Execs weren't used to this kind of approach and were reluctant, so it was a difficult pitch. Marketing pressure drove the effort forward, but we didn't have customers yet, had dim awareness of competitors, and hadn't amassed inventory. It was too early to design a storefront, and I believe if research had been part of the process, we would have questioned the investment more.
The initial checkout process was direct. A character would be selected, details displayed, and purchase initiated. Rough concepts drove discussions about the in-app economy and back-end architecture, but it was squishy. I was processing a lot of input from upper management instead of looking at how real users wanted to browse or discover things.
There were numerous rounds of revisions—some interesting, some questionable, none driven by consumer behavior or testing. We added more detailed filtering and sorting, but weren't aware yet of how overwhelmed our users would be with a product like ours, and too many choices. Interviews and affinity mapping would have served us well.
I proposed displaying related items to cross-sell and upsell, something that's commonplace today. Management read that a blue Buy button would improve sales dramatically, but without analytics or multivariate testing I wasn't able to support the best color, shape, language, or placement.
Learnings
Key things from a UX toolkit would have saved effort going down this path. Initial needs assessment and problem definition was flawed, but might have halted development early.
Being more focused on customer behavior would have steered us toward a more useful target audience and smaller inventory (and later, it did). Research and analytics would have given me the information to make more durable design choices, and Management confidence in them.
This was a key project that led to proposing greater company focus on UX, and my promotion to VP of UX. Doing research in that role, I discovered a lot about our actual target audience that invalidated much of this approach. Even being in a desktop browser was a weak move when our audience started in our application. In fact, they were starting to move toward mobile phones, and we shifted development to that more successful direction.
