Fei Xiao
← All work

Yarn

Chat fiction: thriller stories told one text message at a time.

Designer → VP ProductMammoth Media2016–2019

What it is

Short fiction told as a text thread. You tap to see the next message, like you've picked up someone else's phone.

Problem

Teens like stories, but a wall of text on a phone asks for more attention than most people give it. Yarn made each text message the unit of reading. One tap, one line, and it feels like reading over someone's shoulder, not sitting down with a book.

My role

Yarn was the first app I helped Mammoth Media build, starting from nothing. I began as its designer and became VP of Product, and I stayed the principal designer the whole time. Every screen here is mine. I also helped come up with the tap-through video format and led monetization, including the in-app currency and the subscription.

Outcome

60 million stories and 8 billion messages read. The average reader opened Yarn four times a day, spending close to an hour in their first week. Yarn ranked consistently in the US Top 50, Mammoth's apps reached more than 30% of US Gen Z, and Yarn's revenue helped the company raise a $13M Series A.

Case study

Onboarding

The app used to drop you straight into a story, and it was always horror.

Four Yarn onboarding screens: the splash screen followed by three walkthrough steps ending in a text-based genre selector

A genre picker, in words

Testing showed that whether someone stuck around depended heavily on which genre they landed in, and horror isn't for everyone. So now the app starts by asking what you're in the mood for. We used words instead of cover art on purpose. At that point you don't know any of these stories, so a grid of images is just noise.

Case study

Discovery

Getting people to find their next story was critical to retention.

Five versions of the Yarn discovery catalog, ranging from large featured cover art to dense grids fitting many stories on screen

Five catalogs, A/B tested

We tested several catalog layouts against each other, plus different cover art for each story to see if the image alone changed clicks. The result was simple and a little humbling for a designer: the layouts that fit the most stories on screen won.

Case study

Paywall iterations

Yarn was free to start and paid to keep reading, so the paywall was where the business either worked or didn't. We tested a lot of them.

Three Yarn screens showing a blurred image inside a story with a prompt to pay to reveal it

The blurred image

The best performer wasn't a wall at all. It was a blurred image dropped into the story, with an offer to pay to see it. It worked because it showed up while you were already hooked, and wanting to see one specific image is a much smaller ask than a subscription.

Four Yarn paywall variants: a subscription wall, a calmer wall, and two special offer walls with timers and discounts

Walls, timers, trials and offers

We also tested the usual options: a plain subscription wall, a softer version, a countdown offer and a free trial. Each one is a different guess about why someone hasn't paid yet, and testing them against each other is how you find out which guess fits your audience.

Case study

New content formats

Two experiments with what a story could be besides a text thread.

Five Yarn screens: the discovery catalog with a gem balance, a chat story, a priced choice inside the story, and the gem purchase store

Gems and choose your own adventure

I spent a long time working out how a gem currency could fit into branching stories, with some paths free and the better ones costing gems. The choice shows up inside the conversation with a price on it, so you decide whether to spend right when you most want to know what happens next.

Two Yarn screens showing a story told as full screen short form video

Full-screen short-form video

I pushed hard to turn text stories into full-screen video. It drove millions of views and a lot of installs, and the format outlived the app. Readings of these stories still circulate on YouTube.

PreviousAttention