Background

Modern web applications can be complex for novice users to learn, and often include only a limited set of help topics provided by the application developers. Widely-used applications sometimes attract a broader community of help givers who post supplementary help resources (e.g., step-by-step tutorials) to various blogs, Q&A websites, and help forums. But for end users with a goal in mind, searching for external help resources and trying to assess their relevance can be an arduous chore that distracts from the task at hand.

CheatSheet provides an alternative for user communities to build comprehensive in-application help resources that can be intelligently recommended to help-seekers as they learn various tasks.

Simplifying the search for community-curated help

The vocabulary problem imposes limits on many users' ability to use traditional search engines for learning about application features they are unfamiliar with. In contrast, CheatSheet's non-intrusive browser extension lets users find contextually-relevant help for their current task with the click of a button, without ever leaving the application they're learning. Moreover, its nature as a client-side addon means it operates entirely in the hands of the user community, without the need for developers to integrate it into their application.

Unlike older intelligent help recommenders like the much-maligned Clippy, CheatSheet doesn't interrupt tasks in progress, and it gives control back to user communities to curate the range of help offered for each application.

Automatically generating step-by-step tutorials

In addition to tools for quickly annotating and sharing application screenshots, CheatSheet goes further in streamlining the creation of multi-step tutorials. Users can turn on CheatSheet's tutorial recording feature when they want to document a longer task, during which CheatSheet will automatically capture annotated screenshots upon certain page interactions, building a tutorial template in the background. The author can then completely customize their tutorial by adding or removing steps, editing the generated annotations and text descriptions, or flagging sections for community input.

Once they're satisfied, they can publish their tutorial on CheatSheet for others to learn from in-context, or share it as a standalone link for anyone to view on the web (e.g., to share a quick guide with a friend who doesn't use CheatSheet).

User-driven personalization

Unlike many other help solutions, CheatSheet gives users several ways to personalize what kinds of help are prioritized for them. First, they can identify their level of expertise with an application to prioritize help that is likely to be more helpful for someone in their position (first-timers, novices, or more experienced users). Many users also exhibit strong preferences for help that is either highly visual in nature (like screenshots and videos) or more text-based (like written instructions). CheatSheet can take these preferences into account to surface help resources that cater more toward either end of this spectrum.

Other tools in CheatSheet give users more options to quickly find the help they need, such as its task-oriented filtering and community-driven metrics to highlight what others have found useful.

CheatSheet showed promise with real users

Usability testing highlighted that users appreciate how CheatSheet is seamlessly available within their application when they run into problems and learning opportunities. When I deployed CheatSheet to 15 users for a week to understand how they would use it in certain real-world contexts, several users also independently authored tips and tutorials for various web applications they use, even beyond the more limited deployment goals.

CheatSheet was the subject of published papers at GI 2015 (the original incarnation as a visual memory aid) and CSCW 2017 (the newer community-based help features, under the name Social CheatSheet).