In October 2019, Portland State University hosted the annual conference of the Association for Computing Machinery (ACM) Special Interest Group on the Design of Communication (SIGDOC). Several local academicians and practitioners attended, including your STC Carolina vice president, Art Berger. In this recap, he shares what he learned and hopes to take forward into the new year, about three broad areas: designing for your audience, usability testing, and chatbots. What industry trends and practices are you excited to try out in 2020?

Audience and design thinking

In an increasingly personalized world of technology that both targets and distracts us, how do we design for our audiences? Or even figure out who they are? Many panels addressed such questions, with discussions centering around reaching customers through methods such as UX, personas, and network analysis.

Gustav Verhulsdonck spoke about how different types of UX impact design choices. Behavioral UX, stems from fields of psychology and human factors. It exams how techniques such as nudges, cue-action-reward, and bias prediction are used to steer users towards action.

  • C: Context, environment, and focus of the task as designed
  • H: Habit, mental shortcuts that people take
  • O: Other people, social peer pressure
  • I: Incentives, short-term rewards usually chosen over long-term
  • C: Congruence, or the desire to maintain a positive persona in public
  • E: Emotions are remembered and felt more than details
  • S: Salience, the principle that we attend to easy tasks before hard ones

When such steering is motivated by questionable ethics, the design might be considered dark or hostile UX. Often such design falls on a spectrum. It can be useful to plot out this spectrum in a quadrant, such as the following figure developed by Gustav.

Mindful UX is when you seek to understand the motivations, constraints, and other situatedness of your design practices and communication models, in order to identify how to shift these towards more benevolent, user-empowered experiences. Clair Lauer, for example, gave a talk about how mindful UX can be applied towards scientific communication.

Emma Rose and Jenny Greeve led a workshop on “demystifying” user personas, which included a history and discussion of the controversies. It was noted that some people don’t think personas are useful, for reasons such as:

  • Personas are often not used at all, left forgotten
  • Personas usually benefit the designers who create them, not the populations they’re meant to represent
  • Personas often include information such as demographics that is controversial and might not have any bearing on user action, which is what we’re supposed to design for

Nonetheless, personas can be helpful tools, and if your organization uses them, you might as well learn some techniques for developing ones that align both to your users’ and business’s goals. Three notable techniques for developing personas are as follows.

  1. Alignment persona: Gather a representative group of internal stakeholders, about 8-12 from docs, development, project management, customer support, and marketing. Facilitate a top-down exercise to brainstorm who the users are; their goals, actions, and problems; grouping and prioritization; and next steps for data collection.
  2. User-generated persona: Ask a group of actual users to brainstorm their responses, emotions, and motivations to a particular scenario, from a bottom-up approach. You might then use alignment techniques like grouping to organize the information.
  3. Full-fledged persona: Based upon as much data as you can get, and potentially using the persona groups that were the outputs of the alignment or user-generated personas, flesh out your user as much as possible. This type of persona often produces a sort of bio card, such as the example. Each portion of the card should be mapped back to specific data you have, not impressions or creative imaginings.

The following figure is an example card that you might use to capture the results of your persona research.

Philip Gallagher gave a talk about redesigning audiences based on analysis. As technology changes, the users of our products might be pieces of technology itself, such as a microservice that consumes your product’s API. In such cases, traditional ideas of task-based writing become harder for writers to frame. Human users might have a variety of tasks that must be integrated across products not just your own, whereas technology might have very discreet tasks that your documentation as words on a page cannot inform.

With distributed network analysis (DNA), you might be able to better map out the relationships among your human and non-human users, as well as use artificial intelligence (AI) to predict potential use cases down the line. A framework for DNA includes 4 main points:

  1. Situational analysis for design
  2. All users make, change, and need resources for design
  3. Usability and access to the nodes in the network
  4. Design can be both the problem and solution; focus on the outcome

Using the framework to think about the network of users, you can begin to map the rhetorical appeals that you use to design your communication, such as in the following figure. Rhetorical appeals of DNA include:

  • Genre appeals, such as a wiki layout
  • Responsiveness of the system, including both the system responding to users and allowing users to respond in a timely manner (performance)
  • Alignment to a usability metric, such as Quesenbery’s 5 Es (effective, efficient, engaging, error tolerant, easy to learn)
  • Connection amongst nodes, such as by linking out to other topics
  • Reference materials, tagging, indexing, and other metadata particularly for non-human users

Finally, you might wonder how to get audience members to participate in your research. Ashley Hardin at Red Hat shared some practices that her team uses to write upstream open source and product documentation. Because access to external users can be hard to get, start internally, drawing especially from “boots on the ground” teams like customer support, solution architects, and technical sales. Seek out internal teams that use your product, and embed it in as many of your own workflows as possible. Push back on documenting everything, try to work on customer tickets yourself, and frame your own documentation issues in terms of user stories. A common template for user-story issues is as follows.

As a <type> user, I want to <do this task> so that <this goal happens>.

Usability testing

Content analysis for usability testing

Candice Welhausen showed how she performed a content analysis to study the usability of an app based on its publicly available user reviews in the app store. From this, she categorized the content and used it to determine usability along three main areas:

  • Functional: How the product works
  • Productive: What people report doing with the product
  • Speculative: The context of use and how the app interacts with the user environment and other actions

Conditional usability testing

The student research competition winner, Nupoor Ranade from the local NC State Communication, Rhetoric, and Digital Media (CRDM) Ph.D. program, developed an open source algorithm for assigning conditional values to usability testing steps. This approach maximizes efficiency by allowing you to spend less time testing user behaviors. All users get a certain amount of time (say 5 minutes) and the algorithm shows them steps that fit within their capabilities. For example, if they answer a harder question, it will keep asking them harder questions to see how they do.

Search and chatbots

Alexandra Catá-Ross and Nupoor Ranade (both from NC State’s CRDM local program) gave what was probably my favorite talk among many great talks about the “Boundary of Content Ecology: Chatbots, User Experience, Heuristics, and Pedagogy.”

In the first part of this talk, Nupoor provided a background for the intelligent algorithms that chatbots are based off of, and how they relate to technical communication. For example, chatbots revolve around:

  • Restructuring technical communication information to relate to specific, user-generated applications
  • Producing new knowledge that’s actionable within a particular context
  • Increasing UX satisfaction and performance
  • Reducing support volume
  • Delivering accurate and accessible content

While all this sounds great, a word of warning is that chatbots are only as good as their design and delivery. Many companies roll out a chatbot just because the competition has one, and this is not an ideal experience.

Another point of discussion is how chatbots differ from search, and whether your content findability problem is actually a chatbot problem or a search problem. The following figure contrasts how search differs from chatbots.

If you decide to implement a chatbot, consider these points when designing it.

  • Data and taxonomy
  • Decision-making algorithm: rules-based (going away) vs. probability
  • Audience and algorithm response variance, including bot personality, identity, and credibility
  • Storage and delivery (you can’t just spit out 1000 results like search)

Alexandra discussed usability qualities of a good chatbot, and an heuristic that you can use rank a chatbot as good, bad, or average. It included qualities such as the following.

Objective Goal Good Bad Average
Input The way a user wants to access information
Role Clear that it’s a bot
Limits Ease of transferring to a human
Context Answers are within the pane
Link Directs you straight to the right place, not a general landing
Responsiveness Drills down to get information quickly
Knowable To all senses, accessible
Operable Supports actions such as enter, skip, upload, complete tasks


Bonus takeaway

Finally, a big trend that emerged, perhaps as part of the participatory research emphasis, was increasing collaborations between academia and industry. Heather Turner and Laura Gonzales, for example, reported on programs they worked on to increase underrepresented undergrad’s internship placement. Focusing on asset-based connections, their program included these features:

  • Coaching students through developing a UX portfolio piece
  • Providing weekly mentorship meetings following methods of
  • Hosting a final convention with big employers like Google in a big city

Local STC Carolina members and NC State professor Dr. Stacey Pigg and MSTC alumni Laura Zdanski, Rachael Graham, and Art Berger shared an academic-industry partnership research into ways that early career technical communicators professionally develop and learn on the job. Some of the key takeaways are that learning how to learn helps better than learning a particular topic, and learning together in a group can help you learn more than you could have individually. All the more reason to attend STC Carolina’s monthly educational events!

Save the date!

Wish you could attend a professional conference, but don’t want (or have the budget) to travel? Your local STC Carolina chapter is excited to announce that we are hosting the DITA Oak City Summit (DOCS) conference on structured authoring in a changing technical communication landscape, right here in the Triangle area, June 18-19, 2020. More details to come, but save the date!


Submit a proposal

Interested in getting some speaking experience or meeting with industry and academic researchers? Consider submitting a proposal to SIGDOC 2020 conference, October 3-4, in Denton, TX. 

Art Berger

Art Berger

Technical Writer at IBM and VP of STC Carolina

Recap contributed by Art Berger.