7

I would like for high rep users to be able to recommend posts for tweeting through the site's twitter account. Call this a "positive recommendation" if you will.

I would also like for high rep users to "flag" a question to prevent it from being tweeted. And we could call this a "negative recommendation."

Why?
This question on Programmers led to some follow up commentary within The Whiteboard. The question itself is a "meh" question that could have been tolerated, but then it got way too much publicity from being tweeted. The problem is that this isn't an exception. This pattern has repeated itself before, and this request is to help control the damage meh questions do to a site's reputation when they get a lot of visibility.

How it might work...

What I'm envisioning is two more icons within the share panel.
enter image description here

  1. The first would recommend the question for tweeting.

  2. The second would count as a recommendation against tweeting the question.

If there aren't any positive recommendations, then the [existing tweet selection algorithm] would apply but including a check to skip questions with recommendations against them.

Once a positively recommended question is tweeted, it would be removed from further tweet consideration. Just like it is handled currently.

Benefits...

  • Improve quality of what is tweeted - active communities can curate for the best and make sure it gets highlighted.

  • Make selection more human and less automated (therefore more valuable or less spam like). Higher quality tweets may lead to higher numbers of followers.


Footnotes:

This wouldn't apply to SO or MSO as they don't have twitter accounts.

My request is related to this MSO request to allow moderators to retweet with the site account but is different in that a) high rep users could participate and b) only recommendations would be made and users would not directly send the tweet.

35
  • We already have buttons for users to provide feedback on whether or not they feel a post is of high or low quality; it's up/down votes.
    – Servy
    Commented Jan 10, 2014 at 16:32
  • 2
    @Servy it's known for ages that one of these buttons (vote down) doesn't work in hot questions
    – gnat
    Commented Jan 10, 2014 at 16:37
  • 4
    @Servy - this is a bit different. The questions are merely meh but not close worthy. And they could be tolerated, but you certainly don't want to advertise them for the site. The problem is that they'll gather up votes from the "popular / fun questions are great" camp.
    – user194162
    Commented Jan 10, 2014 at 16:38
  • @GlenH7 It sounds like such a question should still be downvoted then. And if the problem is nobody is doing that, then do you really think the people upvoting these questions are going to actively suggest they not be tweeted? These seem like the kinds of questions that, if people are going out of their way to upvote, that they'll go out of their way to tweet/suggest tweeting.
    – Servy
    Commented Jan 10, 2014 at 16:41
  • 3
    Having the questions hit the hot questions and tweet drastically increases the number of people who want to see it and upvote it on "interesting" criteria rather than the "useful" criteria. People who see it on the hot network and go "meh" to the question rarely visit the question, and those that do don't always have the rep to downvote if they were so inclined.
    – user213963
    Commented Jan 10, 2014 at 16:41
  • @MichaelT So you think those same people who upvote are going to not suggest tweeting, and those who don't have the rep and/or interest to downvote are going to go out of their way to suggest that it not be tweeted?
    – Servy
    Commented Jan 10, 2014 at 16:46
  • 3
    @Servy people who can and do downvote in an attempt to limit the hotness formula impacts are often hopelessly unable to have an effect once a question reaches a certain critical rep. Then it gets tweeted and the floodgates open and you get 5k views on a question in a day where the normal views on a question are ~500... and the rep explodes. Unless a mod is active enough to protect (10k can't protect within a time period) and put notices on it often gets a dozen of meh answers that ultimately leads to a poor impression for new visitors about the site's quality and experience.
    – user213963
    Commented Jan 10, 2014 at 16:53
  • @Servy per my reading, protect feature is proposed for "high rep users" - not to those who don't have the rep...
    – gnat
    Commented Jan 10, 2014 at 16:54
  • 1
    What this is asking is the ability to put on delay the tweet, let things that shouldn't get tweeted from getting tweeted, and allow the site to prepare (protecting the question, notices, deleting existing poor answers to avoid setting a poor example for others). Significant focused viewership can do significant damage to the site (mod flags, meta posts, "why did my answer get deleted", etc...). This is further increased when the mod count is lower on the site... (and we haven't been able to persuade Yannis to carry a pager hooked up to the twitter feed yet.)
    – user213963
    Commented Jan 10, 2014 at 16:57
  • 1
    @MichaelT The fact remains that if people are going out of their way to upvote questions like this, and they're not being downvoted, then what makes you think there are going to be dramatically more votes to not tweet than votes to tweet?
    – Servy
    Commented Jan 10, 2014 at 17:02
  • 2
    @Servy if this was something that was limited to people of a sufficiently high rep, one tends to be under the impression that they have a better idea of the site and how they want it presented to the populous at large. At worse, its no worse than it currently is. At best, we can work to avoid the belief that the twitter feed should be hooked to the close queue by presenting quality questions rather than popular meh questions.
    – user213963
    Commented Jan 10, 2014 at 17:09
  • 3
    @Servy That would be a happily accepted alternative, but due to the hotness formula algorithm improvements that have been suggested multiple times being rejected, this is an attempt to improve something without requesting the algorithm change as the sounds from SE are that the technicals of how the algorithms work makes changing them extremely onerous. Commented Jan 10, 2014 at 17:14
  • 2
    @Servy I don't propose to know the technicals on the SE side of what is easier or harder. If someone from SE comes in and says this is far more onerous and not technically feasible, that is a completely sufficient response. It's worth asking though to see if it is possible because it would be a good benefit to people being able to advertise their sites more effectively. Commented Jan 10, 2014 at 17:33
  • 2
    @Servy - I think you have a solid suggestion in filtering the high-rep users' votes to adjust the tweeting algorithm. Would you mind capturing those comments as an alternative answer? If you don't want to do so, would you mind if I put them into an answer? I think it would answer the intent of the question quite well along with some additional benefits such as keeping the UI simplified.
    – user194162
    Commented Jan 10, 2014 at 19:17
  • 1
    @GlenH7 You're free to write up an answer if you want, I'm not sufficiently interested to write a detailed answer.
    – Servy
    Commented Jan 10, 2014 at 19:31

3 Answers 3

3

An alternative approach...1

Short version

just change the tweeting algorithm to only incorporate votes from high rep users

Long version

The proposed approach has at least a few problems with it:

  1. Not everyone who can use it will use it because it requires additional steps outside of normal site usage.

  2. It requires additional UI work to implement.

  3. StackExchange already has a better reflection of the high-rep user community's thoughts on questions through up | down votes.

The alternative approach would simply filter out votes from users without sufficient rep when determining what should be used for the tweeing algorithm.

Advantages:

  • The information is already there and this approach avoids creating another data field.
  • The scoring mechanism is already in place.
  • It doesn't change how you vote, it's simply changing the algorithm to leverage the information already provided, namely the score from users above a given rep.

Voting anonymity would still be preserved - the filtered value would never be displayed.


1 This answer was originally proposed by Servy in the comments to this question. With Servy's permission, I wrote the answer up as I thought it was worthwhile and it would be an alternative approach to meeting the spirit of the request. Please see this comment and this comment

2

Given the community's record with bikeshed votes, I think recommending questions for tweeting is going to have an adverse effect on the site overall.

Every question that is tweeted will get views and votes out of proportion to the actual quality of the answer. I can see how this could be gamed to give a question attention it doesn't really deserve.

All in all, not an idea I would endorse.

11
  • recommend/protect feature is proposed for "high rep users" (10/15/20Kers?) - do you expect these guys ride on promoting bikeshed stuff?
    – gnat
    Commented Jan 10, 2014 at 17:03
  • 4
    @gnat Yes. ;)
    – yannis
    Commented Jan 10, 2014 at 17:06
  • @Yannis content of the link you refer says "No" to me :)
    – gnat
    Commented Jan 10, 2014 at 17:08
  • 1
    @gnat: No, but all the lemmings that visit the site from the tweet will vote on the post, without regard for the votes that are already there. It's the same thing as Redditting.
    – user102937
    Commented Jan 10, 2014 at 17:08
  • @RobertHarvey as far as I can tell, additional links aren't expected to be displayed to low-rep passers by ("lemmings")
    – gnat
    Commented Jan 10, 2014 at 17:16
  • @gnat: Then why would you bother tweeting?
    – user102937
    Commented Jan 10, 2014 at 17:23
  • 1
    for the same reasons as Yannis, why else. If we're talking about 15Kers, it could go especially easy - 1. protect the question 2. tweet 3. profit
    – gnat
    Commented Jan 10, 2014 at 17:32
  • 1
    @Yannis FWIW, you've only got two closed questions out of 25. I suspect this is a better ratio than the tweet bot. (Alternate suggestion: give Yannis the twitter feed? ... but then, this is kind of what is being suggested)
    – user213963
    Commented Jan 10, 2014 at 17:35
  • @gnat: I think I might be in favor of letting people tweet questions if we could also remove the hot list tweeting in exchange.
    – user102937
    Commented Jan 10, 2014 at 17:36
  • well, that "would be a happily accepted alternative" (Jimmy Hoffa)
    – gnat
    Commented Jan 10, 2014 at 17:38
  • 4
    Actually I'm probably not articulating myself properly. Providing a manual lever where trusted members of the community can endorse high-quality questions actually sounds like a good idea in principle, but I can't think clearly due to my revulsion for Twitter.
    – user102937
    Commented Jan 10, 2014 at 17:43
-2

but then it got way too much publicity from being tweeted.

I'm extremely skeptical StackExchange is going to implement policies specifically meant to discourage people visiting SE sites.

It's good if people visit questions. If it's low quality, downvote, close, do whatever.

Secondly, this is silly manual work that is redundant with downvoting and closing questions. If an answer receives many downvotes, it goes to the bottom of the page, is displayed in lighter grey, the user many delete the answer, etc. If anything I would automate from that, but if someone wants to tweet, even, "Hey look at this silly wrong answer," that's their right and discretion.

15
  • This is more to help control the "damage" from questions that are meh, but not necessarily close worthy.
    – user194162
    Commented Jan 10, 2014 at 16:39
  • @GlenH7 if it's not closeworthy it's not doing any damage. And people visiting the site is certainly not damage.
    – djechlin
    Commented Jan 10, 2014 at 16:39
  • question asked has nothing to do with close votes Twitter bots don't share posts with closevotes on them
    – gnat
    Commented Jan 10, 2014 at 16:39
  • "if it's not closeworthy it's not doing any damage" -- study I did about a year ago suggests that this assumption is flawed
    – gnat
    Commented Jan 10, 2014 at 16:51
  • @gnat that's not damage. new users is good. even if new users' first answers are bad.
    – djechlin
    Commented Jan 10, 2014 at 16:53
  • 5
    This isn't about discouraging people from visiting the site; it's about stopping the site being advertised as something it's not.
    – yannis
    Commented Jan 10, 2014 at 16:55
  • 1
    'my favorite pattern is from MetaFilter, which is: When we start seeing effects of scale, we shut off the new user page. "Someone mentions us in the press and how great we are? Bye!" That's a way of raising the bar, that's creating a threshold of participation. And anyone who bookmarks that page and says "You know, I really want to be in there; maybe I'll go back later," that's the kind of user MeFi wants to have.' (A Group Is Its Own Worst Enemy)
    – gnat
    Commented Jan 10, 2014 at 17:00
  • 2
    How would this feature discourage readership? The way I see it, if the quality of the posts that are tweeted goes up, that will increase engagement from people with the twitter account- and by proxy increase engagement with the sites that get tweeted. Imagine if more of what the twitter account spits out are posts that are of a sincerely high quality, that will attract people who have interest in that content and are more likely going to generate similarly high quality content. There's no way this could be bad other than possibly the technical implementation which I know nothing about Commented Jan 10, 2014 at 17:00
  • 1
    Also in your answer here you're missing one fact in your thoughts about just downvoting fixes this - it doesn't. Getting downvotes respected by the hotness algorithm has been requested but rejected, as it stands they hardly have any effect on the ranking of a question that gets many low/no-quality/me-too answers. If simply down voting balanced against the algorithmic ranking of these posts then this feature wouldn't be requested because the automated tweets would likely choose quality content. This feature speaks about giving a manual lever to allow choosing quality content as well. Commented Jan 10, 2014 at 17:19
  • @JimmyHoffa there's no problem. Visitors are good. You've supplied exactly one opinion that visitors are bad, and zero evidence. For instance, visitors see ads. Some become users. I don't know where this idea that we don't want the cruft of human beings who look at bad questions or answers and continue their existence thinking SE is full of such answers and questions has gained traction, but apparently you and gnat are really into it.
    – djechlin
    Commented Jan 10, 2014 at 17:22
  • 1
    When did I say visitors are bad? I said if we increase the quality of our twitter feed we'll get more visitors. Where in that opinion is "visitors are bad" ? Commented Jan 10, 2014 at 17:22
  • @JimmyHoffa are you claiming that a proper subset of current SO tweets will lead to more visitors?
    – djechlin
    Commented Jan 10, 2014 at 17:24
  • @JimmyHoffa I see... take some effort to curate tweets we think will attract more people so it's easier to follow SO tweets. I'm still against this and don't think it will work this way but admit it's more subtle. For one I think it's perfectly fine that ridiculously bad questions and answers grab visitors. I think probably half of my own visits to SO are to look at stupid questions someone linked me.
    – djechlin
    Commented Jan 10, 2014 at 17:26
  • 1
    @djechlin the difficulty is that no other site is SO. No other site has the number of active moderators to protect questions quickly, 20k+ users to delete bad answers, and the shear volume of other posts to cover up some of the damage on the front page. On smaller sites, we may be lucky if there's a mod on at that time, or a 20k+ user (much less three) and the front page moves at a glacial pace compared to SO. What works for SO doesn't necessarily work for the other, smaller sites.
    – user213963
    Commented Jan 10, 2014 at 17:28
  • analysis in this post suggests that stable increase in visits happens only when site strongly focuses on providing high quality content
    – gnat
    Commented Jan 10, 2014 at 17:30

You must log in to answer this question.