8 of 8 would recommend it. The SUS said 63.4.
5 of 8 below average. Nobody rated it “Excellent”. The concept was loved — the interface was in the way.
Version 1 was built from apps people already love.
Tinder's swipe for picking genres. Instagram's notes for reading moods. Spotify's library for shelves. Google's review chips.
It felt smart: borrow patterns people already know, skip the learning curve. In our first informal test (5 users), one person said they weren't sure where the Join Club button was. We fixed the logo someone else complained about. We left the button.
That button came back as the #1 failure in the formal study.
8 readers. 4 tasks. 1 moderator, 1 note-taker, no hints.
“Thought the genre selection screen was the homepage.”
“The homepage looks like Instagram.”
“Something closer to a modern library than a social media feed.”
Every v2 change answers a specific finding. Nothing was redesigned just for looks.
Testers fumbled the Tinder-style genre swipe. What do you do?
6 of 8 couldn't find the Join button in the corner. Fix?
5 of 8 said the home screen felt like Instagram. Fix?
Label every icon. Ambiguity isn't minimalism.
Attractiveness scored 3.6/7. A tester asked for a “modern library”. So I built one.
Kept the BookPod teal, moved everything else to paper and print: a bookish serif for titles, brass for ratings and shelf planks, one outline icon set. 32 screens, one component library bound to tokens — which directly answers the uneven consistency scores.
v2 hasn't been tested yet. Here's what would prove it worked.
Same 4 tasks, same SUS and NASA-TLX, a new group of ~8 readers — so the numbers compare directly.
Not addressed on purpose: a premium reading subscription (UT2). That's a business-model decision, not a usability fix, and it would have muddied the retest.