Current Issue
This Month's Issue

Follow Fast Company

We’ll come to you.

5 minute read

Q: Who's Dominating Q&A Sites? A: No One

Q&A websites are plagued with problems that few have yet to solve.

Social Media Q&A is all the rage: Facebook just upgraded its questions feature, TED launched a website for its own community, Quora has culled a suprising response from experts, and firebrand Congressman Anthony Weiner held a marathon Twitter session on the anniversary of the new healthcare law. The sheer diversity of platforms shows that a single site is yet to fully satiate the complex demands of the question-and-answer appetite. In part, this is because the most popular platforms suffer from trolls, half-baked soapboxing, and poor data on user's background expertise. Some niche websites have found success, such as Help a Reporter Out, but everyone, from Google to the White House, is still seeking a champion to lead the space.

Negative response bias

Professional pollsters learned early on that voluntary surveys were plagued with angry, emotional respondents. Classic Statistics texts love to regale undergrads with the story of famed writer Ann Landers who solicited responses to the question, "If you had to do it over again, would you have children?" A whopping 70% answered "no," ostensibly upending the long held belief that humans, by and large, love their children (and want more). The culprit, was, of course, negative response bias, since feeling angry boosted the motivation to participate and warn others from experiencing their same fate—flooding the survey with regretful respondents.

Yet, Q&A platforms commit the same methodological sin as Landers: soliciting voluntary answers (as opposed to randomly contacting participants). The negative response bias would ultimately become the death knell for America Speaking Out, an opinion solicitation website for congressional Republicans to canvass their consituents. Almost immediately after it launched, a stampede of radical conservatives and snarky liberals overran the website with recommendations like "How about if Congress actually do [sic] thier job and VET or Usurper in Chief, Obama is NOT a Natural Born Citizen in any way." Since then, the much touted experiment has been left to fade away. While ASO may be an extreme example, any voluntary survey, from a Facebook poll to a Twitter question, is haunted with doubts that the answers come from the least rational observers.

Personal soapbox for attention seekers

Voluntary surveys can pander to the most verbose users in the crowd (think about the student in a college class who feels the need to answer nearly every question). Congressman John Garamendi's office felt the difficulty in sifting through the chaff during his experiment using the mega-popular user-submitted news site, Reddit.com. "Ask me anything" he courageously wrote, and 2,670 comments later, his office updated the post with a diplomatic apology about not being able to "answer all of your questions".

In one case, a user went on a 700+ word diatribe about state education, leaving the office to respond:

"Rep Garamendi is in the US House of Representatives. Your question, while well thought out and detailed would be better addressed to your governor and state legislature."

Other questions were pithy, but just as unlikely to yield a functional response, "Now, would you like to give us a meaningful answer that is not 90% political fluff?"

About a week earlier, Fast Company tried its own Q&A with Congressman Eric Cantor's office on Quora. To pre-empt the chaos seen on Reddit, Cantor asked a specific question (how social media could be used to enhance legislation), and we pleaded with readers to be brief and cite examples. In the end, we got ⅔ of our wishlist, lots of through-provoking, pointed opinions, but little in the way of actionable evidence.

Bystander intervention

Aware of the irrelevancy issue facing public forums, Facebook retooled its "questions" feature to favor a user's inner circle of friends. However, this still leaves Facebook saddled with the fact that fishing for answers can often yield a depressingly low number of responses, leading one to falsely believe that they have ignorant, uncaring friends.

The root cause is that broadcasting a signal for help triggers what psychologists call "responsibility diffusion," where everyone in an audience assumes that someone must have already responded, and therefore no one ends up helping. The phenomon was first studied to understand how, in 1964, a woman could have been violently murdered in plan view in the crowded city of New York. Psychologists were able to successful replicate the behavior, leading otherwise well-intentioned citizens to ignore cries for help once they were aware that others had also heard the problem.

Facebook questions can be one big responsibility diffusion mess. In the tiny screen real estate afforded to questions, there's no indication of urgency, how many friends responded, if the answers were satisfactory, or if it even reached the right people. In short, Facebook questions may be a convenient feature, but a solid Q&A website might need a dedicated platform with an ethos of attentiveness.

Pockets of Success

Smaller niche sites have stumbled upon pockets of success with targeted audiences and specific questions. Help a Reporter Out, for instance, is a popular community for journalists to seek story leads, post interview requests, and scout for quotes. HARO has received rave reviews from countless A-list media outlets, despite minor issues with self-aggrandizing users spamming writers in the hopes of getting attention.

Innocentive is an X Prize-like website for farming out complex scientific problems to the public. Problems include everything from a $20,000 reward for inventing a safe way to drop food and water from a plane to $100,000 for an RNA sequencing techniques that can shield citrus trees from ravenous diseases. While not a typical Q&A website, Innocentive has been remarkable for its ability to uncover nooks of talented, amateur scientists that would have otherwise never been tasked with a major research project.

Innocentive's success with hyper-niche, expert audiences raises the expectations for quality crowdsourcing and illustrates just how far Q&A sites have yet to go.

And, there is demand for such a solution at the highest levels of industry and government.

Google payed a cool 50 million for Aardvark, a platform that ranks user expertise based on characteristics, such as location and web acitivity, and then snakes through the social graph for the right person to pose questions to.

The White House has an ongoing experiment with an expertise finder wiki, Expertnet. As Chief Technology Officer of the United States, Aneesh Chopra's, recent blog alludes, even the White house is experiencing the same challenges with getting specific, actionable answers.

Thus, the opportunities for a solid Q&A platform are enormous, and the first person to design a product that hits the entrepreneurial sweet spot of utility, marketing, and design, may likely become the next web industry titan.

Follow Fast Company on Twitter. Also, Follow Greg Ferenstein on Twitter.

[Image by [F]oxymoron]

Read More: Most Innovative Companies