Surprising users with something delightful

For some background on how betaworks thinks about the intersection of media & software (and the increasingly blurred line between them), check out this intro from betaworks CEO John Borthwick.

Last Monday, betaworks launched our newest product: a silly and relentlessly fun trading cards app called Glitter. Note: if you haven’t tried Glitter yet, most of this will not make sense to you. Go get the app. I’ll wait here.

Now that you’ve played with the app, one thing you probably noticed is our unique onboarding. We thought we’d share some background on how this came to be, and some data showing how this approach has affected our usage so far. TL;DR: making onboarding fun and engaging — even though it’s much longer than normal app onboarding — has positively impacted engagement and retention in the app.

Product Hunt comments on launch day

“Will this scare people off? Who cares.”

Glitter has a lot of personality, but it started off as a pretty straightforward app with basic functionality. We wanted to keep things anonymous — no user accounts, no feeds — so people could keep each of their cards completely separate. (Imagine you have a business card in there and then also one you send to prospective dates you meet in bars; might not want those associated with each other.) Ultimately, though, we decided the cons of leaving out user accounts outweighed the pros, so we found ourselves contemplating that horrific and daunting part of mobile app development: onboarding.

Onboarding’s almost a dirty word now, among both app creators and app users. It makes you think of video tutorials, invasive (and mandatory) personal questions, and those ugly overlays with 20 arrows on them that try to teach you everything about the app all at once. And as a developer, you’re all too aware of the drop-off that inherently occurs when you insert a blocking account signup screen before users can get any value out of your app.

So we were all pretty upset about having to jam onboarding in front of our beautiful, ethereal app. Look how boring this is:

Original Glitter onboarding, sans fun

One day, though, our fearless leader Christian showed us something he’d been working on “for fun” and didn’t plan to actually include in the app because it would be too much work to implement, and because he thought it would scare people off. It was approximately 30 screens long, it made no sense, it went against all the advice we’d ever read about onboarding… and we all unanimously agreed that we had to include it in the app. The Glitter team is made up of 4 lifelong gamers, and we immediately loved that it made the app feel like a game, instead of a “service.” (And as you’ll see in the next few releases, we’ll be adding much more fun stuff to evoke that feeling…)

Tangent: A Freudian mascot

Christian had originally thrown in the deprecated Information Desk Person emoji as a placeholder for a mascot for the onboarding, and she worked well, but we always knew we couldn’t actually use her in production (yep, even emoji are bound by IP masters). So we set to work on finding a new mascot for Glitter.

As you can see, we started out inspired by Jason King (long story, but he’s a fictional crime-novelist-turned-detective from a 1970s British TV series who we became strangely transfixed by). We worked with the excellent emoji artist Julia Heffernan to create a few iterations of the child of Information Desk Person and Mr. King. To us, he was perfect. For everyone else around us, the results were… shockingly divisive.

We agonized over the mascot. Unlike the wacky onboarding, which we knew was risky but seemed worth it, including our new emoji (which some people at betaworks dubbed “creepy” and “dad-like”) seemed like an unnecessary risk which might scare people off for the wrong reasons. We ended up switching him out at the last minute for a less controversial mascot that Christian whipped up in an hour right before launch.

In this case, the important thing was that it became clear how much we needed a mascot to make the onboarding feel fun and personable. Going through the process of developing and testing out a completely ridiculous option helped us land on something a little more reasonable, but still unique and much more interesting than we started out with. The new mascot is more simple than our mustached man, but she meshes well with the gaming aesthetic of Glitter, and we’ve compensated for her relatively ordinary look by giving her an interesting personality. She has her own story to tell, which will add some dimension to Glitter and reveal itself slowly over the course of future releases. Here, you can even collect her card:

Ok, so does anyone like this app?

Keep reading if you care about the actual data, but the short answer is yes. Currently, although we have a modestly sized audience (we haven’t done any advertising yet), we have a 5 star rating in the App Store, some amazing engagement numbers, and a whole lot of happy tweets:

Even before we launched, we were pretty sure we had built something fun (everyone who had beta tested it loved it, so we knew there was *something* there), but we worried a bit about onboarding completion, engagement within the app, and retention over time.

In terms of onboarding, turns out we had nothing to worry about:

Here’s our insane completion rates over time for onboarding. You can see there’s a slight drop when we ask for email and password (as is to be expected), but even with that requirement our completion rate was at 90% on the first day, and has hovered between 60–80% on all other days. Since we launched the app on Product Hunt, we think that drop after the first day is due to two factors:

1. The first day was mostly tech early adopters, who are more likely to complete onboarding in any product so they can try it out.
2. The first day was mostly (if not all) English speaking users; since then we’ve had some adoption in non-English speaking countries, and we don’t have any localization in for other languages yet.

So that’s the hard data. Subjectively, people have also been clear about the fact that they really, really love the onboarding process:

Engagement in general has also been shockingly good. We left Glitter pretty open-ended; there isn’t a tutorial explaining what to do or how to do it. Even so, it’s clear that people are spending the time to figure it out and complete a number of important events in the app:

• 75% of people who completed onboarding created at least one card
• 40% of people shared at least one card
• 18% of people collected at least one card

Plus, average session time is 2 and a half minutes, right where it should be along with other social networks on mobile. Considering that there is no walk-through and we haven’t yet built an easy way to find people to share to or cards to collect, we think this level of engagement is due to the fun and engaging tone set in onboarding and continued throughout the app.

At betaworks, we’re particularly obsessed with D1 and D7 (day one and day 7) retention curves for apps; it’s something we learned from the Dots team and it’s a virus (a good one) that has spread around the studio. Right now, we’re hovering at around 25% D1 retention. This actually isn’t bad considering how insanely hard retention is on mobile, but because we’re building something inherently social, we’d obviously like it to be higher. We’re focusing on this now because we’re already seeing how important it is for people to stick around; if they do, they steadily create more cards, which is a good thing.

Cumulative cards per retained user over time

It’s too early to have a good sense of our D7 retention, but we’re monitoring that closely as well.

So that’s where we are so far. We’re thrilled that people have been enjoying Glitter, but there is so much more depth to come and we can’t wait to show you the improved social features that are coming next (and the data that will follow). You can grab Glitter in the App Store now, and email us if you’ve got any feedback or suggestions.

xoxo,

Team Glitter

PS: We’re running a contest this weekend to find the best Glitter card. Submit yours on Twitter with the hashtag #glitterpro by this Sunday (8/16) and an 11 year old will be choosing HER favorite card on Monday. The winner will be awarded a $500 Amazon gift card (and, of course, everlasting fame and really high card stats).

From L to R: Creative technologist Alex Baldwin; Co-founder Christian Rocha; Co-founder Maya Prohovnik; Mobile dev lead Jordan Carney