Quora: The gap that Q&A sites need to cross

Last month, Tawanda Sibanda and I decided to conduct a study on Quora’s community engagement to understand whether Quora had the potential to become a mainstream resource.

Basic Views

Quora's product design is destined to be a niche website for technical people. We will consider whether the following three factors continue Quora's philosophy for technical people:

1. Newsfeed : Unlike Facebook, Quora's feed is not linear, that is, each possible answer to a popular question forms a separate feed entry, so that the entire feed can revolve around the same topic;
2. Voting: Our data shows that more views lead to more votes (not surprisingly). Therefore, for a newsfeed design, the only way to get more votes or community attention is to do trending topics, and for Quora, this means asking more questions related to technology and business.
3. Tech Celebrities: Questions from well-known users like Reed Hastings and Dustin Moskovitz easily get 50+ votes and tons of responses. If they respond to your thread, it will be a major Newsfeed hit.

In Quora, these three elements are self-reinforcing; if you are a non-technical person visiting Quora for the first time, Quora may not be of much value to you. Let's look at the data:

What we did

From a sample of 30 answers, we analyzed the factors that contribute to high engagement (votes and thanks on Quora). Although this is only a small sample, it can serve as a test of the waters.

The sample input data (data related to whether the answer will be read) includes the number of answers/votes for the current question (including all answers), the number of previous readings and attention of the question, and whether the question is related to our area of ​​expertise;

Output metrics (community response to our answers) include the number of votes/comments on the question, the number of reads after answering the question, and the growth in the number of followers.

Our findings

1: Choosing the right questions is half the battle
Using the 2×2 table below, we found that if we answer popular questions, we will get more votes and comments. On average, replies in popular topics with high views get 3 times more votes than replies in popular topics with low views (2.6 vs. 0.9), 4 times more comments than replies in popular topics with low views (0.4 vs. 0.1), and 14 times more traffic than replies in popular topics with low views. In short, people on Quora care more about technology.

Question popularity matrix reflecting community engagement

Definition of coordinate axes in the figure:
Reading : The number of views of a question before we answer it. The median is 130, and views above 130 are high, and views below 130 are low.
Write : The number of responses to the question before we answer it. The median is 3, and numbers above 3 are high, and numbers below 3 are low.

To further reveal how this difference biases Quora responses, we generated the following table from 30 sample answers. In short, 87% of the views and 57% of the votes are concentrated in 20% of the questions.

Question/Views/Votes Concentration

2: Question quality is key, except if you choose a "sexy" question <br /> We calculated the correlation coefficient between the number of votes and the number of answers from our sample. We found that the following two indicators have a statistically significant correlation with getting votes, namely the length of the answer and the number of views of the answer.

The importance of answer length is self-evident. More thoughtful and lengthy answers are more likely to receive more votes, and the correlation between views and votes is consistent with the view that "hot topics receive more votes".

3: Respondent reputation/identity matters <br /> In this experiment, our top answer was about venture capital, which happens to be my job. Across all samples, we averaged 1.7 votes per answer. Reed Hastings’ answers averaged only 19 words, but his average vote was 109.8, and we averaged 220 words per answer.

The chart below provides a survey of some Quora "celebrities" whose words and actions have a significant impact on Quora.

what does that mean

Although our 30 samples are not a lot, it raises some interesting questions about the future development of Quora. Quora currently gives people the impression of being just a niche market for technical users (including Zhihu in China, nothing else, if Zhihu wants to serve all users, it will be no different from Baidu Knows.), and its logic is as follows:

• The existing user base is mainly technology users, and their concerns are often related to technology trends and entrepreneurship;
• As a result, these topics have a higher priority in the Newsfeed algorithm;
• Therefore, new users will also engage with the same content, vote more on these topics, and make more replies;
• This encourages the questioner to ask more similar questions on Quora, while questions from non-technical fields and non-celebrity topics are ignored.
• This cycle repeats

Some interesting thoughts:

• Could Quora provide more visibility/status indicators to incentivize users to start new topics?
• Can we select those responses that are actually useful to the question, such as hiding the respondent's name from being visible to prevent fans of celebrities from voting for their idols?

Reprinted from: //www.leiphone.com/can-quora-cross-the-chasm.html