At the Web App Masters Tour in Seattle, WA Doug Bowman outlined Twitter’s approach to helping new and one-time users become loyal repeaters by discussing the company’s focus on A Simple Ladder of Engagement.
- Engagement is not a binary value. Once people get into Twitter, it’s a very fast ramp to active usage. But there a set of users that sign up and never come back. Twitter is really interested in getting these users engaged.
- Early adopters tend to show up and know what to share right away. As a result, Twitter originally thought it was equally about consumption and production. Over time they saw more consumers than producers and assumed their goal was to move consumers to producers.
- Twitter gets about 300 thousand sign-ups per day. So sign-up is a great opportunity to shape behavior, as it is people’s first time experience.
- Usually when people come to Twitter they see a big cliff. They don’t know what it is or how to use it. The people that have been using it for years know what is going on, new users don’t. There’s a big learning curve.
- In trying to understand how people “get” Twitter –the research team looked at a specific group of users: people that signed up, left, and then came back.
- Curious, Casual, Committed/Core: Twitter’s framework for managing a ladder of engagement. Users fall on only one of these steps.
- Curious users are hard to track. Twitter has many different entry points into their content and they can’t instrument many of them. Curious users may approach Twitter through many different means. Have an interest but have not signed up yet.
- Need to give curious users something to aspire to toward. Many developers focus on features that present “how?” information. But if people understand “why?” they should use their app, they’ll figure out “how”.
- There are three tools Twitter uses to manage their ladder of engagement. Magnets: intangible things that initially attract users to your app. Hooks: tangible things that users can grab on to and keep them coming back. Glue: something the service provides to their users so they create an experience they want to stick with.
- Twitter’s new homepage was designed to get people to tweets (through search) they may be interested in. But it didn’t work well to convert new users. The home page design lasted five months. Second redesign played down search and showed top tweets and users.
- Twitter wants people to be able to warm up to the service before they sign up. Ultimately they hope people can start using the service without signing up first.
- Blogger lost three quarters of their users (in sign-up process) at a two-radio button question asking where they wanted their blog hosted: on blogspot.com or connected via ftp.
- In the redesign of Blogger, the team took the old five step process and dropped it down to three step process. They hid the ftp option under an advanced options dialog. Number of users went up exponentially when they made the change.
- Twitter has a blank screen of death. Landing on a blank profile page is the worst case scenario for Twitter sign-ups. If people land on a blank screen, they don’t know what to do and may not become engaged.
- Twitter’s old sign-up process asked you to scan your address book to find friends already on Twitter. Second step was suggested follower list (480 people selected by Twitter employees). Every user that wasn’t paying attention would be following 20 users from the suggested user list when they went through sign-up flow. This also overinflated Twitter follower counts for people in the suggested user list.
- In their revamped the sign-up flow, Twitter now first asks people what they are interested in. All suggestions are algorithmically generated instead of employee selected. People can pick one of 18 topics they may be interested in. All follows are explicit actions (no more checkbox for all 20).
- New flow was three steps (instead of two) but conversions went up 29%. The process got people to relevant interests through categories of suggested followers.
- Sign-up flow improvements targeted curious users.
- Casual users: often make up the largest group of users. They are also the most volatile. They may move to core users or may fall off and no longer be engaged. Twitter casual users come back 1 to 6 times a month. Casual users really on hooks to keep coming back and discover the value of your app.
- If an app can help people achieve goals, connect to others, or become more efficient –this keeps people engaged. Entertaining or delighting people takes this even further.
- Twitter monitors behavior on the site, and turns these behaviors into features. These features are mostly for casual users. Three examples: @ replies, retweets, and saved searches. These features give casual users feedback that encourages them to return.
- Retweets were designed to simplify the process of sharing interesting content on the service with minimum input from users.
- Don’t forget about email notifications that bring users back to the service.
- There are over 100,000 applications using Twitter.
- People using Twitter.com and one over application that users Twitter’s APIs have +375% more logins and +867% more tweets a month. Most of this come from mobile clients.
- Core users: are you regular users. They log in more than 7 days per month. People who log in more than seven days per month have a higher likelihood of doing the same thing again. Higher number of log-ins per month doesn’t increase likelihood of repeat behavior substantially.
- Core users are not necessarily evangelists –hardcore users are. Business users really quickly jump to being core users. Other people take a more gradual path to get there. For core users, focus on the glue: tools for people to make their own experience.
- When users get to core level, they want to distribute their own tweets: embeddable tweet widgets.
- Embrace hoe core users use your service. Embrace lingo. One example: change from what re you doing to “what’s happening?” People using service in much broader way then Twitter intended.
- Build value: add features and remove features as well. It’s often really hard to remove features but if you want to create a great experience, you have to do it.