“Is an egg a fruit or a vegetable?”
“Will my laptop get heavier if I put more files on it?”
“Is pepperoni pizza vegetarian?”
Yahoo Answers, the long-running question-and-answer platform that hosted bad questions, worse answers, and everything in-between, is shutting down in May 2021. While it may be tempting to conclude that its closure is a sign that the time for question-and-answer sites has passed, displaced by Facebook, Google, and Wikipedia, the opposite is true. As we discuss in our essay “Top 100: The most popular social media platforms and what they can teach us,” Q&A is a popular social media logic, occupying a middle space between niche logics like crypto logic and the dominant logics that house platforms like Facebook, YouTube, and WhatsApp.
Who thought asking strangers on the internet to answer their questions was a good idea? And what makes Q&A logic different from other logics we’ve discussed?
Let’s start with a bit of background. People have been using the internet for Q&A since its earliest days. The WELL, an influential early bulletin board system (BBS), had a dedicated “Experts On The WELL” topic, which was a place to ask questions that experts in the community could respond to. Similarly, Usenet, a decentralized precursor to internet forums, had a number of groups dedicated to Q&A, many of them about technical topics; for example, “comp.soft-sys.matlab” focused on the programming language Matlab. Within active Usenet groups, .help groups formed explicitly to answer user questions and help with debugging and troubleshooting.
As Usenet faded in importance and Web 2.0 emerged as a dominant internet paradigm, a lot of Q&A moved to dedicated Q&A platforms like Yahoo Answers, Quora, and Stack Overflow. Today, Quora, Stack Overflow, and homework-focused Brainly are some of the most popular sites in the world—though they aren’t usually what comes to mind when we hear “social media.”
Users of social media platforms like Twitter or Facebook may not know what kind of content to expect when they log on—“status updates” could involve anything from a joke, a selfie, or a rant designed to start an argument. But on Q&A platforms, the rules are much simpler—you can ask a question or answer one. Marc Smith, a sociologist, and director of the Social Media Research Foundation, has done a number of studies focused on Q&A communities, particularly the different social roles that exist in them. He’s found that users on Q&A sites largely fall into one of two roles: Questioners or Answerers. Questioners mostly ask questions, and many ask only a single question. Answerers, on the other hand, mostly answer questions and typically return to the site fairly consistently to answer questions.
Smith has found that the distribution of Questioners and Answerers is quite skewed. The vast majority of users are Questioners—a study estimated that 83 percent of users on Stack Overflow, a popular Q&A platform for software developers, were Questioners. Because there are so many Questioners, so few Answerers, and little overlap between the two groups, it’s common for tensions to arise. Answerers can feel like they aren’t receiving proper recognition for their contributions, or that Questioners are wasting their time by asking repetitive/basic questions and not following community rules. Meanwhile, Questioners can feel intimidated and discouraged by the rude and unforgiving responses they receive from miffed Answerers, especially when they are new to the community.
A good example of this dynamic can be found on Stack Overflow where in recent years, a number of users have complained that many Answerers are mean and discouraging, creating an environment that is annoying for most people and harmful for novice and marginalized software developers. Unhappy Questioners point to responses like these to illustrate their complaints: “if you don’t get this … you have no business making a portfolio as a web developer” and “I’m not sure I know how to spell it out in plainer English … not much to digest, even if you have to read it ten times.”
Some Answerers have responded to this criticism by pointing out that they face a deluge of questions, many of which are repetitive, basic, or don’t fit the norms of the community—the snarky responses are both a deterrent to bad questions and an outlet for their frustration. As one Answerer put it: “[Stack Overflow] forces us to constantly interact with a stream of garbage; that will inevitably create hostility.”
Why do people bother answering random people’s questions on the internet in the first place? They aren’t being paid, so is it altruism? Smith observed that most Answerers on the WELL were motivated by status and recognition: “Being knowledgeable in the WELL and being free with your knowledge is a sure way to gain status, friends, and visibility.” In a broader study of online cooperation, Peter Kollock argues that, in addition to status and recognition, contributors to public goods online are motivated by reciprocity (someone else will answer their question), attachment (they want to see the community succeed), and a sense of efficacy (pride in their work and observable impact). Kollock’s broader frame of online cooperation in the service of public goods can help us understand that people contribute to Q&A for many of the same reasons that people contribute to open-source software, Wikipedia, and online reviews.
It is useful to think about Q&A platforms, especially those that are archived and easily searchable, as community constructed knowledge bases that are closer to Wikipedia than they are to Facebook. When people can’t find information from their typical information sources, they turn to Q&A platforms.
One implication of understanding Q&A platforms as archives of knowledge is that people can use them to find their preferred answers to questions. For example, HuffPost India revealed that Quora India is home to a right-wing ecosystem of contributors who answer questions about controversial political and social topics like “love jihad” (a conspiracy theory that alleges Muslim men trick Hindu women into converting to Islam) and the Ayodhya dispute (a dispute over a Hindu holy site, previously home to a mosque that was destroyed in Hindu nationalist riots). A BJP political strategist (BJP is the right-wing political party of India’s current prime minister Narendra Modi) told HuffPost India she spends over an hour a day on Quora to clarify her party’s stance on various issues. For people looking to express opinions as facts or to oversimplify complex issues, Quora’s question-and-answer format can be attractive—it can give a false veneer of truth to top answers, which are largely selected on the basis of popularity, not accuracy or quality, and ignores whether a question is resistant to a single “correct” answer. As a PR executive working with CEOs on their Quora presence explained to HuffPost India: “opinions should be expressed as ‘scientifically’ as possible, they shouldn’t seem bigoted.”
The veneer of authority offered by Q&A sites points to another issue: even if a question is well-suited to the question-and-answer format, the crowd-sourced answers can be wrong. For example, a study exposed major security flaws in some of the most popular C++ code snippets posted on Stack Overflow. The snippets were included in over 2500 GitHub projects, a sign of how widely the dangerous code spread. It’s worth recalling that before Wikipedia became the arbiter of truth for major web platforms, it was widely criticized for flaws in quality. Our understanding of the quality of the content on Q&A platforms will likely evolve, but just as with Wikipedia, robust rules, tools, and norms must be developed to improve the quality of the information on them.
The business model of most Q&A platforms is a combination of advertising and subscriptions. Many Q&A platforms struggle to make a profit thanks to competition from Facebook, Google, and Amazon for advertisers and a weak value proposition for subscribers. For example, China’s largest Q&A platform, Zhihu, recently explored an IPO but is facing questions about its valuation due to its weak revenues. Additionally, as Zhihu has scaled, users have complained that content quality has decreased, a common problem for large Q&A platforms.
Since Q&A platforms facilitate the creation of public goods, it’s worth considering whether an alternative business/ownership model would be more appropriate. Both Wikimedia and Q&A platforms rely on volunteer communities to co-create a knowledge base that is broadly available. Freed of conventional venture capital success metrics, Wikimedia has been able to focus on maintaining and improving the quality of its platform even as it scales. Quora may find it easier to live up to its motto as “a place to share knowledge and better understand the world,” if it were free of the obligation to justify its $2 billion valuation to investors.
As with any social platform, the experience of a Q&A platform is related to its governance and affordances. In a paper comparing the quality of content on LiveQnA (a mid-2000s Q&A platform) and Q&A groups on Usenet, Smith et al. argue that affordances, norms, and governance explain much of the differences in quality. For example, LiveQnA’s open tagging system made it easier for users to create new, sensational tags and thus attract attention to off-topic posts. On the other hand, Usenet’s balkanized and byzantine system made it “harder to attract widespread attention but easier to reach specialized audiences.” Additionally, decades-old Usenet norms dictated that you were more likely to receive a reply if you referenced your role and involvement with the group, encouraging members to invest time and energy towards quality contributions. In contrast, LiveQnA lacked norms or affordances for people to document their contributions, making it difficult to establish and maintain standards for the community.
The importance of strong community norms becomes even clearer when we recognize that most Q&A platforms rely on volunteers to do much of the governance, using a model similar to Reddit’s. Community moderators work to verify answers, prune questions, and set and enforce the rules for the community. Governance also occurs through upvotes and downvotes, where users vote on whether a question or an answer is helpful and relevant. Users with a strong “reputation” have more influence with their votes. Reputation is built up through a variety of actions that the platform thinks indicate that the user is reliable and deserves more sway in the community. Reputation systems can be gamed if they aren’t designed correctly, and can also result in a small group holding outsize influence on the platform. For example, April Wensel, founder of Compassionate Coding, described the influence high reputation users have on Stack Overflow as a “self-reinforcing toxic power structure.”
The distinct “social roles'' on Q&A platforms may hold some lessons for other social media. Benjamin Mako Hill has proposed that Wikipedia’s success came in part from its familiarity—people knew what an encyclopedia entry was, which provided a rough script for them to follow in contributing, rather than figuring out what a new tool or community was for. Similarly, Q&A logic communities provide a simple script for participation and may benefit from the resulting social scaffolding. It’s worth considering whether other social media logics have well-defined roles, and what this might tell us about them: YouTube’s roles of “creator” and “audience” suggest another popular model with clearly defined ways to participate.
Tech companies sometimes imagine their users as interchangeable parts, when in fact, just like in the offline world, they form different parts of a community with distinct roles and responsibilities which require tailored approaches. Much as smart Q&A communities will design for the special needs of Questioners and Answerers, Facebook, Twitter, and YouTube likely need to design around elites and influencers as well as less experienced users.
For example, someone with 1 million followers should not be treated the same as someone with 100 followers. (And, in truth, they are likely not handled the same way by platforms’ moderation and content policies, with platforms reluctant to take action against their most popular users.) For a person with 1 million followers, Twitter is essentially a printing press on steroids, handling logistics and distribution for a one-person publication. For a person with 100 followers, Twitter is more like a book club, somewhere to discuss things they’ve read and maybe interact with the people who wrote what they’re reading. An approach informed by social roles could make it so people with 1 million followers are subject to stricter rules around harassment and misinformation, while people with 100 followers are freer to interact without platform intervention. Moving towards a better understanding of social roles on platforms might help platforms escape from the impossible task of designing one-size-fits-all content policies.
Clearer roles might also help us better understand the scale and concentration of platforms. YouTube has billions of users who upload hundreds of hours of video each minute. But only 3 percent of the site’s over 50 million creators account for 90 percent of views on YouTube. That changes the equations around YouTube’s difficulties with content moderation—a small group of popular and controversial creators likely account for many of the site’s most challenging moderation problems. Is the best way to handle those users with automated tools and overworked/underpaid contract moderators, or should YouTube invest significant effort in a separate, robust, and nuanced moderation process for those users? In any community, people take on different roles with different responsibilities. Social media should embrace that reality and implement policies and affordances that reflect it.
Q&A platforms can serve as a reminder of the narrow focus of much of the contemporary debate around social media. The hundreds of millions of people gathering on Quora, Stack Overflow, and Brainly to ask and answer questions are left out when we allow Facebook and Twitter to become synonymous with social media. Though Q&A platforms have their imperfections, we should marvel that millions of people donate their time and effort to help strangers online. Just as we celebrate and learn from Wikipedia, we should celebrate and learn from Q&A platforms, taking cues from their emphasis on social roles and how norms and affordances that respect those roles can lead to healthier social media spaces.
Chand Rajendra-Nicolucci is a research fellow at the Knight Institute.
Ethan Zuckerman is associate professor of public policy, information and communication at the University of Massachusetts at Amherst, director of the Initiative on Digital Public Infrastructure, and was the 2020-2021 visiting research scholar at the Knight Institute.