When browsing message streams, users primarily scan the latest messages from top to bottom, scroll a little, and identify sources by names and logos, rather than catch up on every item they missed. These observations come from research described in Jakob Nielsen’s 2009 article. They apply to the reverse-chronological stream interfaces common at the time, which placed the newest messages at the top, and cannot be directly generalized to all platforms or to streams that use different sorting methods today.
Starting with the latest content, rather than catching up on older posts
All the media formats in the study presented posts in a single stream, with the newest messages at the top. Although some RSS users sorted by source, browsing posts in time order was still the main approach in the study.
Users appreciated the simplicity of this interface: they could look down the list and perhaps scroll a little further, without using additional commands. They were content to read the latest information and generally did not seek out older posts they might have missed. This does not mean that every message was read in full. Rather, users were not trying to catch up on everything.
Messages that leave the main page rarely regain attention
The article states that once a message dropped off a user’s main page, it effectively ceased to exist, because very few people continued to the second page. This describes how users in the study allocated their attention. It does not mean that messages were actually deleted or that all streams divide content into pages in the same way.
The author contrasted this with email newsletters and other email notifications: emails require users to delete them manually, while social updates move downward over time and fade from view on their own. Participants liked not having to clear away social updates manually. On this basis, the author also judged stream-based media to be less effective than email newsletters at maintaining customer relationships. However, this was the author’s judgment based on the research at the time, not a general rule for comparing media effectiveness today.
Identifying sources through names and small logos
Browsing a stream involves more than looking at text. The study noted that users also looked for logos and usernames to identify sources, making it important for both to be easy to recognize.
The article cited the U.S. Department of Education’s Twitter account name at the time, “usedgov,” as an example: users associated it with “used government,” which created a poor impression. The example illustrates how a name can be misread; it does not describe current account-naming rules.
Logos presented a similar problem: some services displayed them at sizes so small that the original design might no longer be recognizable. The design lesson from this research is that names should be easy to understand and logos should remain clear when scaled down, so users can identify sources while scanning quickly.
Questions about message stream browsing
Can this study calculate the probability that a message will be seen?
No. The 2009 study described participants’ tendency to read newer information, but the supplied material contains no impression rates, readership rates, or data sufficient to calculate the probability that an individual message will be seen. The browsing tendencies observed in the study therefore cannot be converted into a percentage or used to compare reach across different platforms today.
Does recognizing a message’s source mean that users have read its content?
Not necessarily. Recognizing a source does not mean users have read the content. The study noted that users scanned names and logos to identify sources, but the material does not establish that they then necessarily opened, finished reading, or remembered a message. This observation alone therefore cannot support a claim that easily recognizable names and logos will ensure the content is read.
References
- Writing for Social Media: Usability of Corporate Content Distributed Through Facebook, Twitter & LinkedInwww.nngroup.com · 2009-10-12T05:00:00+0000 — Discusses: Can this study calculate the probability that a message will be seen?; Does recognizing a message’s source mean that users have read its content?






