BuddyBuzz (Master's Research)


As part of my master's program at Stanford, I performed research the Stanford Persuasive Technology Lab. While there, I designed, developed, and ran experiments on a mobile reading platform that we called, “BuddyBuzz.”



BuddyBuzz was a reading application for mobile phones that used a technology called Rapid Serial Visual Presentation, or RSVP. RSVP involves quickly flashing single words in a fixed position on the screen such that the reader’s eye stays focused on one point. It is used as a reading aid for the visually impaired, but can also be used to assist in speed reading.

At the time of our research, most mobile phones used 128 by 128 pixel screens, resulting in a poor reading experience for blocks of text longer than SMS or short emails. We believed that RSVP could provide a better reading experience for longer reading sessions on mobile devices. In addition, we wondered if we could apply any persuasive techniques being studied in the lab to affect readers’ behavior.


Demonstration Video

Here’s a video of BuddyBuzz in action. Users could change their reading speed by hitting up and down, and select content from a variety of news outlets and blogs.

View the video here.

Design Considerations

The two most important design decisions in the application came from early iterations on the project and careful consideration of how we read naturally.

First, we found that at a constant rate of flashing, it was hard to establish a rhythm of reading. It was difficult to get into the flow of the text. Furthermore, it was easy to miss commas, periods, or long words. To address this, we adjusted the flashing rate to display words for longer periods of time based proportionally on their length, and added extra time after commas and periods. The result was a sequence that lined up surprisingly closely with spoken word.

Second, we noticed that it was easy to get lost in lengthy quotes or parentheses. In news articles that quoted people, knowing when quotes began and ended was particularly important. To help address that, we added contextual icons, separate from the text stream, that were present whenever the current words were inside of quotes or parentheses.



The focus of my research was studying the effects of periodic praise displayed while using the BuddyBuzz application. We took advantage of the load times to display short statements of praise. During these wait times, which lasted 15 to 30 seconds, the application displayed sentences like, “You are a very fast reader,” or “You learn quickly.” We created a database of 50 praise statements to attempt to avoid repetition.

Using volunteers from the larger BuddyBuzz user base, we split 50 users into two groups of 25: one that saw the praise and one that saw a neutral “Loading…” screen. Over a period of two weeks, we measured, among other things, the total number of words read by participants. The results were striking and statistically significant. Over the course of the two week study, participants who received the praise read twice as much as those who did not.

In previous studies on computers praising humans, people in the praise group tended to believe that the computer spoke the truth, reported more positive feeling, and judged the computer’s performance on tasks more favorably. While we didn’t measure these in our study, we theorized that positive feelings towards the application may have resulted in longer reading sessions.


Interest from Outside of the Lab

BuddyBuzz attracted some attention from a variety of people outside of the lab. One article, originally published in the San Jose Mercury News, was syndicated in several newspapers around the country.

In addition, we presented BuddyBuzz at several conferences, including IBM’s New Paradigms in Using Computers 2005, where we presented a poster and showed a demo of BuddyBuzz to conference attendees. Our presentation of BuddyBuzz at this conference was covered by CNET.