What is meta? ×
Meta Stack Exchange is where users like you discuss bugs, features, and support issues that affect the software powering all 128 Stack Exchange communities.

Background:

Moderators can, for some time now, access limited analytics data in the moderator tools for a Stack Exchange site. (Refer to my previous feature request.) While the analytics tools aren't full Google Analytics, they do help provide insight about site traffic and traffic sources.

I find the Search Keywords section particularly interesting, as most traffic comes from search engines. I like to look at keywords responsible for the most traffic, and the long-tail (infrequent) keywords. (Single-day date ranges help uncover the long tail, but results displayed are limited to ~150 rows. Can pagination be added, or length increased?)

The main question:

What do you think of the idea of moderators posting occasional as-yet-unasked questions on a Stack Exchange site, drawn from long-tail search keywords?

While most long-tail queries are simply variations of the more popular queries, sometimes there's a rare nugget: an unanswered question.

I actually tried this idea at Personal Finance and Money, back when it was still my Stack Exchange 1.0 site, BasicallyMoney.com. Look at Web Search User - Personal Finance and Money for example questions actually drawn from Google Analytics long-tail queries. Those are questions a real human being asked (of a search engine), which led to a site visit, but didn't yield an exact match.

The goals in posting some of those questions were:

  1. to provide the community with additional good questions to answer, and
  2. to drive more traffic to the site by converting what began as a long-tail search into a more frequent search, due to better matching quality content.

Yes, it's Search Engine Optimization (SEO) – and I know the idea of SEO turns some of you off – but I considered this "white hat" SEO, i.e. "involves no deception [...and...] making efforts to deliver quality content to an audience that has requested the quality content."

I haven't practiced this since the site was re-launched as a Stack Exchange 2.0 site, but I have considered it. Of course, since Personal Finance and Money is no longer "my site", I would like to ask the community what you think of this practice? What are your thoughts? Thank you.

share|improve this question

2 Answers 2

I enjoy the concept of having the content that you know people will look for. I have often times thought, on Gaming, to ask a few questions which I know are questions people have searched for answers, and will continue to search for as time passes.

When it comes to random analytics, though, I'm slightly concerned from the standpoint of the question owner's investment. Barring your drive-by users and other anomalies, the question owner is interested in getting a solution to the problem. They have a certain investment in having the problem solved, which means clarifications when necessary as well as acceptance at the conclusion. At the very least, in having the problem the author has some measure of ability to measure how correct or useful questions are.

This isn't guaranteed when simply going through the motions of question asking off a script. If the words mean something to you, but the actual problem doesn't, it strains the good faith of asking the question. Questions are more than just words, they are real problems that people understand. And if you don't know the first thing about asking the particular question, you can make quite a fool of yourself. This is especially the case for the sites with wide scopes.

Consider these horrid examples where users had real questions in part derived from confusion but otherwise had real problems beneath them. Even the subject matter experts were baffled between the question as it probably might have been meant, versus the question the author intended and literally asked. And because the author is invested in the solution to the intended question and made proper correspondence, they got resolved properly. A third-party user asking the same question may not know the real question, and may stray away by answerer influence from the literal writings even if they were the intended content. In these two scenarios, it would have ended up with two legitimate questions being closed.

Those are extreme examples, but muted examples would still be bad. Investment in a question is necessary for good questions. Plant seeds which you plan to watch as a loving gardener, and which you actually can harvest the yield from. If the third-party author doesn't comprehend the answers in good faith, then the question is meaningless. And, worse, you are stuck with something that would close the same question as asked by someone who does dabble in that subject matter. I think at the very least, this investment needs to be required of the moderators before we even attempt such a thing. It really improves the quality of our content, and quality is where we deliver best.

share|improve this answer
4  
Certainly - I agree 100%. Not having seen the question on the site before is necessary, but not sufficient: For this to work, moderators undertaking the practice would need to (1) understand, (2) appreciate, and (3) be curious about a potential question drawn from analytics when considering it for posting. The approach should neither be random, nor following a how-to script. –  Chris W. Rea Dec 23 '10 at 17:06

My first reaction was that it feels dirty, but after reading your paragraph claiming that it is White Hat, I have changed my mind. While I agree that it isn't a bad thing to do, I'm still not sure that it should be encouraged.

First, this seems like a ton of work to do, and how helpful is it? Would the site benefit more from simply editing the relevant questions to make them a better match for the search queries?

Also, duplicate content, while not explicitly discouraged, shouldn't be generated on purpose. At best, those duplicate questions should be created, and then immediately be closed as duplicates.

share|improve this answer
1  
There isn't always relevant existing content. Sometimes, people stumble onto the site with searches that are rather fruitless, pinging off of merely related questions that won't do at all to solve their actual concern. Looking over some of the Web Search User's questions, many aren't really duplicates and are just their own independent questions. The rare nuggets of unanswered questions, to quote Chris. –  Grace Note Dec 23 '10 at 16:16
    
@GraceNote Exactly- the questions weren't duplicate; merely containing some matching keywords to something related (but not the same) already on the site. –  Chris W. Rea Dec 23 '10 at 16:44
    
It is some work, but not a ton. I tried it when the site was otherwise not very active; e.g. Sundays, holidays, or just slow days. As for worth it, or not: the votes, answers, and number of views for the Web Search User questions seem respectable... although popularity of a question is sometimes hit-and-miss. Not a science. –  Chris W. Rea Dec 23 '10 at 16:55

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .