by YUNNUO CHENG and JAKOB NIELSEN on August 21, 2016
Our user testing revealed plenty of usability problems in WeChat and the various official accounts we tested. This is no surprise, since the perfect user interface hasn’t been built yet, but should come as a warning to those technology enthusiasts who believe that simply moving to a new platform or embracing a new interaction style will result in great user experience. On the contrary: any design will still need polishing and adaptation to users’ real needs, as opposed to companies’ internal hopes for what their customers will do. We’ll just give two examples here.
Team Maker allows users to define group events. The setup screen requires users to enter a name for the new event, but 10 of 14 study participants overlooked this field and went straight for the second field (denoted by a clock icon) to enter the time of the event.
Several design issues combined to cause this problem:
Banner blindness possibly made users overlook the top field because of the graphic.
The prompt was shown as a placeholder text within the field, which is less noticeable than a separate field label.
That field was not aligned with the other fields in the form, and thus looked as if it was not part of a form, but rather a static title.
The activity title prompt was shown in low-contrast text, especially compared with the brighter text used for the time field.
Users may have had a selfish-action bias, preferring fields related to their own needs (the time of the group meeting) over fields related to system needs (the title).
City Heatmap is a WeChat feature that indicates how crowded a given location was. This feature is available under *Wallet → Public Services → City Heatmap. Our users had great difficulty finding it. *This navigational path was simply not expected, and suffered from a convoluted information architecture andproblematic naming with poor information scent.
Such issues are not special to WeChat and we’ve seen them many times in usability studies. New platforms have old usability issues, because the main determinant of usability is the human mind and people’s needs, not technology.
Text-Based Interaction Gets Limited Use
WeChat supports two different types of interaction with an account:
Text-based: Users can text a number or a keyword to a service account and receive an IM reply from that account. (The reply can be automatic or can be generated by a customer-service representative in charge with responding to WeChat queries).
Menu/link based: Users can select a link embedded in a text message or they can use the menu buttons available on official-account interfaces and in the chat window.
Sometimes all these interaction methods were available for an account, but in other situations only the more basic text interface was supported.
In our diary study and in the usability-testing sessions, we noticed only limited use of the text interface. In particular, some users texted back a number in response to a first message received after subscribing to a service account. They also occasionally texted back a keyword hoping to get back matching search results from that company. However, most users much preferred a menu-based interaction over the more effortful text-based interaction. Whenever the bottom menu was available for a service account, users relied on it, as a more expedient way to interact with the company. They also attempted to tap on the numerical options displayed in the chat message as an even faster way to make a choice.
Our initial interest in WeChat was to better understand conversational text interfaces. The appeal of such natural-language interfaces is that they supposedly allow users to simply express their goal and then sit back, while letting the site do all the work for them. No clicks or taps are involved once the initial query has been formulated, so in a sense, such interfaces have the potential of getting closer to the holy grail of usability — zero interaction cost. (Such interfaces do assume that users will be able to formulate a goal — an assumption that does not always hold, because users do not always know the search space well enough.)
We did not find evidence for sophisticated natural-language understanding in WeChat. (Those human-staffed official accounts that reply to individual user queries are definitely not scalable in the long run and were not the norm in our study.) Instead of a true conversational text interface, we discovered a system that warrants the interest of an evolutionary web scientist for the way in which it mimics the evolution of the mobile web — a world in which historic, simple interaction such as the numeric-menu selection and keyword-input coexist with more sophisticated menu-based interfaces or GUIs. These latter methods are newer and often preferred by WeChat users for their lower interaction cost, yet, at this point, they have not yet completely erased the chat-box interface.
