Case study · /01

Doone

Re-engaging digital readers through AI recaps, reading streaks and social reading

Role
Product designer & researcher
Timeline
4 months, 2024
Team
Solo, university project
Focus
UI design, UX research, AI integration
01 · The drop-off

Page 112. She folds the corner, puts the book down, and never picks it up again.

Digital reading keeps growing (£280m in the UK in 2023), but engagement doesn’t keep up. Readers start books and drift: distractions, too much choice, no sense of progress, no one to read with. For a reading app, a half-read book is a lost reader.

02 · Insights

Why readers drift

9 interviews · love and break-up letters to reading · affinity mapping

Affinity map of the interview research: five themes (users, reading habits and preferences, motivations, challenges, and app features and technology), each with codes and participant quotes.

Select a pin or an insight to zoom in.

  1. Dear Reading,

    Perhaps one day we can find our way back to each other, but for now, I need to let go.

    P1 · break-up letter
  2. Dear Reading,

    Whatever I do there are always more to it.

    P4 · break-up letter
  3. Dear Reading,

    I love talking about you with my friends and family.

    P5 · love letter
03 · The bet

How might I design an experience that reignites the joy of reading, and sustains it?

I didn’t want to guess what would bring readers back, so I designed with them. In Crazy 8s sessions, readers sketched ideas alongside me. Three kept coming back: a quick way into the story, a visible reward for progress, and company on the page.

Crazy 8s sketch sheets from two readers, covering onboarding, AI recommendations, gamification and group reading.

Select a pin to zoom into a sketch.

What’s out there, and the gap

  1. Kindle, GoodreadsHuge libraries and reviews, but nothing helps you back into a book you leftThe recap
  2. Bookly, DuolingoStreaks keep people coming back, but reading progress stays a numberThe garden and seeds
  3. TikTok, InstagramBookTok shows readers want company, but it lives off the pageRead together

From my competitor analysis of six direct and four indirect apps.

04 · The solution

One app, built around coming back

01 · First evening

Answers 01 · Coming back is hard

Promise the way back on day one

The first thing Doone shows is the recap, the answer to ‘coming back is hard’. Book Wizard comes second. Sign-up is a sheet that grows out of the garden, and email expands the same sheet.

02 · A Thursday evening

Answers 02 · Progress is invisible

A garden that grows as you read

Home shows your progress as a garden, with one quiet nudge to catch up. Seeds are the one currency: earn them by reading, spend them on rentals.

  1. Catch up in a minute
  2. Your garden
  3. Seeds, one currency
  4. Invite a friend
  5. Rent with seeds
  6. Finish a book

03 · Read together

Answers 03 · Reading feels lonely

Reading stays first, friends are a tap away

Every tool says what it does. Notes start private, friends’ comments sit on the exact passage, and anything from a friend further ahead stays hidden until you get there.

  1. Tools that say what they do
  2. Select a passage
  3. A note, private by default
  4. Friends on this passage
  5. Notes and highlights

04 · Find it, keep it, make it yours

One voice across the rest of the app

Book Wizard explains every match and lets you steer it. Shelves ask who can see them while you make them, and deleting one can be undone. Notifications stay calm: at most one a day, each with one clear action.

05 · Testing

Five readers, eight tasks: 32 of 40 felt good.

I tested the hi-fi prototype with five target readers. After each task, they marked how it felt: green or red. Most of Doone felt good. Where it didn’t, I redesigned it: the garden, the library and Book Wizard all changed because of what readers said.

How each of five readers felt after each of eight tasks
Reader12345678
Reader 1Task 1: felt goodTask 2: felt goodTask 3: felt goodTask 4: felt goodTask 5: felt goodTask 6: felt goodTask 7: felt goodTask 8: felt good
Reader 2Task 1: felt goodTask 2: felt goodTask 3: felt badTask 4: felt goodTask 5: felt goodTask 6: felt goodTask 7: felt goodTask 8: felt good
Reader 3Task 1: felt goodTask 2: felt goodTask 3: mixedTask 4: felt goodTask 5: felt goodTask 6: felt badTask 7: felt badTask 8: felt good
Reader 4Task 1: felt goodTask 2: felt goodTask 3: mixedTask 4: felt goodTask 5: mixedTask 6: felt goodTask 7: felt goodTask 8: felt good
Reader 5Task 1: felt goodTask 2: mixedTask 3: felt badTask 4: felt goodTask 5: felt goodTask 6: felt goodTask 7: felt goodTask 8: felt good
  • 32 felt good
  • 4 mixed
  • 4 felt bad

What I saw, what I changed

Drag to see each change

06 · Reflection

What I’d measure next, and what I’d do differently

If Doone shipped, I’d judge it on three numbers: how many started books get finished, how many readers return after a week away, and how many keep a seven-day streak. I’d test the recap first: it’s the cheapest way to win back a reader who has already left.

  • 01

    Finish rate

    Started books that get finished

  • 02

    Coming back

    Readers who return after a week away

  • 03

    Seven-day streaks

    Readers who read seven days in a row

What I’d do differently

I’d run a diary study over several weeks, because a one-hour session can’t show how motivation fades. And I’d design streaks that forgive a missed day, so the tree never becomes a source of guilt.