With approximately 20 billion page views a month and 30% page view growth year-over-year, Tumblr is growing and vibrant content community (this blog is hosted on Tumblr). Tumblr has invested heavily in its mobile product to drive this growth.
In the following post, I will be detailing Tumblr’s mobile registration strategy which is at the heart of its recent page view growth.
Mobile Landing Page
Opening a freshly downloaded Tumblr app presents a welcoming, clean stream of images. From the title, Tumblr communicates to a new visitor that its a community about creativity. The stream of images also includes a handful of gifs, which gives the stream a low-cost, vivid feel. The landing page is infinitely scrollable and explorable in a visitor experience.
Despite a explorable stream, the preeminence of the main two call-to-actions (“CTAs”), “Log in” and “Sign up”, probably drives most visitors into the logged-in user experience over exploring in the logged-out visitor experience. Both of these CTAs are close to the right thumb, the most typically used finger for mobile navigation.
There are some risks with a dynamic landing page. When the images fail to load quickly, a new visitor can have a sub-par first experience.
The first ask in the registration flow is selecting your interests. Tumblr avoids starting out with a high effort ask. Instead, it builds up several low effort asks as means of developing user investment to minimize drop off in the final steps. By starting with interest selection, more visitors are willing to go down the registration flow.
Suggested interests seem to be ordered randomly. Two interests tiles at the top are auto-selected, but at this point in the flow, Tumblr has no data about me to accurately select these interests. These are shots in the dark. Most likely, Tumblr is using this technique to demonstrate how to select interests and leveraging a user’s "sense of justice" to engage in the flow.
The title of the page frames selecting interests as means of finding Tumblr users rather than for personal edification (like a recommendation engine). Interest selection in Tumblr is designed to setup an interest/social network graph. As a mode of insurance, Tumblr auto-selects interests to try to assure that the next step in on-boarding (Following Tumblr users) is relevant and launches. Otherwise, selecting no interests skips this key step.
Suggested Tumblr Users
After selecting interests, Tumblr suggests relevant users to follow. These suggested users are probably very active Tumblers. Promoting active Tumblers to follow opens two opportunities: user-to-user notifications and priming desirable community social norms to a new user.
User-to-user notifications are deemed higher quality to a user over a message from the webmaster. These notifications are deemed authentic and relevant. You can only send so many product updates and weekly recommendations. By following active users, the number of possible email notifications is far greater.
Promoting active Tumblr users also primes a new user of the quality of the community and allows a new user to consume content. Most new users consume and don’t want to contribute to the community at sign up. They want to see how the party is going before they jump in.
In the header of the page, the CTA “Follow” efficiently includes the number of selected users (auto-followed of course!) that will be followed.
Interestingly, the progress bar/dots have not progressed even though one page has already passed. Obviously, there are more than three pages in Tumblr’s registration flow. Instead the progress bar is divided into three stages: follow interests, find friends, and high effort submission. This is an interesting technique to make the flow feel shorter.
Tumblr combines finding friends via address book and Facebook into one step. Most find friend flows separate these two data sources into different paths. In other mobile apps, finding friends from multiple data sources typically requires going back a step and repeating the process again and again. This is a nice simplification.
After connecting either your address book or Facebook account, Tumblr finds your contacts that are Tumblr users and auto-selects them to follow.
After Tumblr’s find friends flow, a modal appears over the landing page with the high effort ask of email and password. There is no header and the phrase “sign up” is minimized to the grey out completion button. The progress bar UI also disappears (I assume I am in the final step?).
Entering information is rather painless. The mobile keyboard automatically appears to fill in the space beneath the modal, so the modal feels supported. A new user’s username auto attaches their personalized Tumblr URL.
"Next" is dropped from the completion CTA. Intriguingly selecting "Close" dumps a registering visitor back to step one and the information invested into the new user profile is lost.
Age and Terms of Service
Submitting your age is painless and quite creative. Most apps use Apple’s native date picker UI. Tumblr only asks you to enter your age (the information that really matters to them for legal reason), not your birthday.
The “Next” success CTA rears its head again, but the “Sign up” success CTA disappears.
Uploading a profile picture is one of the most challenging aspects of completing a user profile. Drop offs are usually very high (50-70% are not uncommon); however, user generated profile photos are correlated with higher engagement and higher retention. For moving metrics, a user generated profile photo really makes a difference.
The profile picture UI is slightly intrusive but having a real photo is critical for quality user investment. No one wants to see a bunch of Tumblr avatar profile photos in their stream. It makes a user profile feel shallow and empty. Imagine seeing a LinkedIn profile with no user generated profile photo; you automatically question its authenticity.
Once new users finish the registration process, they land in their home feed.
After a minimum of 15 clicks, a visitor can turn into a new Tumblr user.
Tumblr’s registration flow is medium to high in terms of effort. There is a lot of up front investment. Rather than trying to leverage behavior later in the new user experience, Tumblr’s heavier upfront investment is more likely to produce a retained user. The opposite strategy would be to minimize the number of steps it takes to move a visitor to a new user. This strategy works well if your magic moment is not highly dependent on information unique to a user. Tumblr needs data from a user to create the magic moment.
Tumblr’s overall experience is slick but not ground breaking. There are some pitfalls: there is a lack of standards in button placement and CTA language. There is no activation flow after registration. My hunch is that Tumblr’s mobile app is most likely a retention play. Since Tumblr has position its product messaging as a creative, inspiration social network with a blogging spice over just a blogging platform, content consumption is a higher priority activity on mobile than contribution, though some brave ones fearless write posts on their mobile phone.
If you would like your product to be featured and reviewed, please send me an email with the details of your product to aginnt @ gmail.