Don't let the trolls win!

Instead of killing comments, we should be trying to fix them

Stay on Top of Enterprise Technology Trends

Get updates impacting your industry from our GigaOm Research Community
Join the Community!

Every month or so, it seems, a media outlet decides to get rid of their comments. The latest is The Week, which follows Reuters and Re/code, both of whom shut down their comments recently. Every outlet that does this says the same thing: conversation has moved to social media, etc. But as New York Times staffer Mat Yurow argues in a post at Medium, this argument is essentially a cop-out. Comments need to be fixed, not killed.

In its post about shutting down comments, The Week says that “in the age of social media, the smartest and most vibrant reader conversations have moved off of news sites and onto Facebook and Twitter.” But even if this is the case — which I’m not disputing — whose fault is that? Most sites have done virtually nothing to try and make their comment sections a more hospitable place for smart and vibrant discussion, so why wouldn’t it go elsewhere?

The Week piece also says that the site has “a deep respect for the intelligence and opinions of our readers.” But not enough respect, apparently, to allow those readers to post their thoughts about its articles on the same page where those articles appear.

Instead, readers are forced to try and track down the writers and editors of the magazine on various social networks and then do their best to find the conversation about whatever article they are interested in, and then convince someone on the staff to engage with them. What many news outlets seem to mean by the discussion “moving to Twitter and Facebook” is that it’s much easier to ignore.

Engagement equals value

Yurow, by contrast — who works on the audience development team at the New York Times, and before that worked for Huffington Post and Bloomberg — believes as I do that handing over a key component of your relationship with readers to Twitter and Facebook is a mistake. Not only does it give up something valuable, but it suggests to readers that their comments and interaction aren’t worth the trouble:

“To simply give up, and hand our most engaged users over to Facebook and Twitter is a major loss to a industry that is in dire need of loyalty. We need to come up with real, sustainable solutions?—?solutions that view community through the lens of modern culture, technology and business. It is imperative that we save comments. We owe it to our readers, we owe it to our writers, and we owe it to ourselves.”

Comments are broken, Yurow argues, because most publications have not put the time or resources into trying to make them work, and so they have become troll and spam-filled backwaters that everyone tries to avoid. But that’s not the fault of readers — it’s the fault of publishers for not seeing their relationship with their readers as being of value. So how can this perception of comments be turned around? Yurow outlines several ways in his post.

For one thing, media sites could look at the actual return on investment that they get from engaging with readers — which is real, and can be measured. It’s easier for sites with subscriptions to do this, since they can track how many commenters eventually “convert” into being subscribers. But it’s not that hard to tie reader time spent with things that matter to your business, whether it’s advertising or something else. As Yurow points out:

“Commenters do (at least) two things most site visitors do not: they explicitly demonstrate interest in your product, and they willingly hand over their email address. In any other business, we’d call these people ‘warm leads.’ In media, we call them trolls.”

Readers deserve our time

The other key point I agree with Yurow on is that many sites are to blame for their own troll-filled comments, because their writers and editors fail to engage even with the intelligent commenters, and so the predictable happens — flame-wars and offensive behavior take over. As blogger Anil Dash pointed out in a post in 2011, if there is bad behavior in your comments then you as the site owner are partially to blame. Yurow notes:

“It is important that some action is taken to remind readers that their voice is being heard. This can come in the form of a featured comment, a short response, or even a strategic email or tweet. Will this completely stop belligerence at the bottom of the page? Absolutely not. But it will help set the expectation of civil discourse and conversation.”

Among other things, Yurow also suggests that publishers try to figure out some way to make comments more relevant for more readers — whether it’s by having editors and writers highlight or point out interesting comments (something Forbes and other sites such as Gigaom already do), or by using algorithms and other tools such as reader votes to surface the best, something the New York Times itself has experimented with in the past.

41 Responses to “Instead of killing comments, we should be trying to fix them”

  1. Stephanie Hoover

    You said “readers are forced to try and track down the writers and editors of the magazine on various social networks” – by whom are they “forced”? And why do readers feel they have a duty/right/obligation to comment on a journalist’s work…? I laugh thinking of what these aggressive wanna-be columnists must do when they watch TV. Do they talk to the screen, because it’s their “right” to comment on what they view…? While everyone is entitled to an opinion, from where did this philosophy arise that everyone is entitled to sharing their opinion in some public forum? This is why the line between true journalism and crap opinion columns has been dangerously blurred. Let the journalists do their job. Wanna’ talk about what professionals research and write? That’s what friends, family, and the other folks living in their parents’ basements are for.

  2. There are multiple comments here about the high cost of moderating comments, but why does it need to cost anything at all? Why not crowdsource this function? I have seen multiple entertainment fan websites where the comments are expertly policed by nominated members of the commenting community who are passionate about the subject matter. And they cost nothing.

  3. One way that some (BuzzFeed, Huffington Post for example) are trying to deal with this is by forcing you to log in to Facebook to comment. Such sites leave my bookmarked reading list as soon as I encounter that. I don’t Facebook, don’t want to Facebook and never will Facebook.

  4. dieseltaylor

    Absolutely agree. I have used bulletin Board for gmaes and history, forums for investing in companies, and for wrting on op eds and stories.

    SOm eof htese have been very enlightening and yes you do get trolls on some sites but overall to cop out is ridiculous. I would argue strongly that the thoughtfil person with te added value comments is to smart to be on Facebook or Twitter commenting on something elsewhere.

  5. Enrique Lavin

    Matt, another terrific piece of evangelizing the merits of comments under news stories.
    Here’s a question that came up recently when I was part of a panel that had both national and local news organizations discussing the merits of comments.
    The national sites were trending toward tamping down on keeping the comments open. The cost of moderating daily stories seemed to outweigh the benefits. It made me rethink position on newsroom engagement. My newsroom, which is focussed on one state, has
    a narrower audience than a national outlet. Hot-button issues, which are immense time-sucks on moderators (such as myself), force us to focus on negative comments, policing the space to make it less toxic for other who want to participate civilly.
    We value our readers because they are our leads to stories and they provide a good glimpse of what our community cares about.
    I could see how a national desk would have its hands full.
    In any case, in theme with this conversation, here’s a little something we tried this year:
    Top 10 commenters of 2014

  6. christoph j. walther

    An interesting example of how to deal with comments is the German online magazine Krautreporter which was launched end of October. There subscribers do not pay for being able to read articles, but for belonging to the community and the right to read and to post comments. The founders there argue that many readers are eager to read the comments because knowing what others think about an article is important for the formation of ones own opinion.

  7. Key reason we prefer on-site comments – Disqus allows pre-moderation. I wish Facebook did – but shudder at the time that would take. Latest abuse: Obscene graphics that get by the FB profanity filters. Sigh. Are these people REALLY like this (I feel sorry for their friends and family if so) – or is it more that this is their awful alter-ego that allows them to be more “normal”/nice in real life? (“Nicebook,” anyone?)

    • Mat Yurow

      How did he solve it, Michael? Seth’s post suggest he’s done what many other publishers have done since: removed comments altogether instead of finding a solution that works for his site.

    • Dan Mitchell

      Does that include the people who don’t even look at the headline, but decide to declare that Obama is a socialist Muslim under stories about the bond market or river pollution? I guess that disturbs the “echo chamber,” but I’m not sure what good it does.

      • Mat Yurow

        Dan, there’s no doubt there’s a lot of junk in the comment sections. But I think it’s on the publishers to find a way to surface the best of content to the top. If we can effectively do that, the vitriol will be buried to the bottom, and bad actors (hopefully) will lose the incentive to post.

        • Dan Mitchell

          Sure it’s on the publishers. Which by definition means that they’re directing resources away from their core function — reporting the news — and toward managing a platform for people to yammer. On general interest or otherwise broadly focused sites, the overwhelming majority of comments are at best pointless (“I wish we could do something about this crime problem.”) and at worst vile.

          The proportion of comments that are actually valuable in some way (like, they add information or truly unique perspective) is tiny. And even when comments are well managed, you have to wade through a lot of pointless yammering to ever see those few worthy comments. That takes time, effort, and money. Meanwhile, reporters and editors are being laid off left and right, and we see the predictable resulting awfulness piling up all around us.

            • Mat Yurow

              Yes, moderating comments takes time and money. Each business needs to assess its unique situation to determine if the arbitrage would be ROI positive.

              And just to clarify, when I say value I mean monetary value. Each commenter can be assigned a tangible LTV (see my response to your other thread). In this regard, I’m “valuing” all commenters (good and bad) the same.

  8. David P Church

    I’ve often thought that if people had to pay 5 cents for the privilege of using media websites to broadcast each of their opinions and rants to the world, comment sections would immediately improve. Because there would be a record of the payment transaction, people posting under pseudonyms would be reminded that there is no such thing as a truly anonymous comment. (The 3 cent markup on 2 cents is for costs and profit. Commenting privileges would be included with subscriptions.)

    One of the reasons publishers (particularly daily newspapers, I imagine) are reluctant to do away with their comment sections, or crack down on the trolls and ideologically fixated, is that comment sections bring eyeballs and page views. People who would be loath to actually buy a “conservative” or “left wing” newspaper, are quick to seize on any news story with with a ideological angle and write disparaging comments on all the enemy newspapers.

    The people who are reluctant to put their real name to their opinion are probably equally unlikely to put up 5 cents. But regular readers of a news media site, might be willing to pay the 5 cents if they have a insight or comment relevant to the article.

    • Mat Yurow

      David, I think this is an interesting idea. Certainly solves the revenue issue too!

      My biggest concern for a pay-to-play (or say!) model, is that it might hinder otherwise good actors from contributing. If you were a frequent commenter of a site you might begin to self censor yourself to keep expenses at bay.

      Would you pay 5 cents to post this comment?

      • Dan Mitchell

        Are “frequent commenters” on sites like Reuters or The Week, or those run by regional newspapers, worth courting? And wouldn’t more self-censorship be better?

        • Mat Yurow

          Certainly. A commenter typically needs to hand over their email address first. This opens the door to new ways to reach those users. Smart publishers will be able to find a correlation between commenting/registering and some revenue generating action (subscription, additional page views, newsletter ad rev).

          • Dan Mitchell

            “This opens the door to new ways to reach those users.”

            Unless, of course, they opt out or ignore emails, as I assume do many commenters who are mostly interested in saying racist stuff or calling Obama a socialist libtard.

            Basically, your theory that all the effort and time expended on maintaining comments sections — and the overwhelming tide of awfulness that results — is worth it in order to get people’s email addresses, some (vanishingly small?) fraction of which will result in some kind of revenue-generation.

            Is it possible that this time and effort be better spent peddling newsletters, which have the advantage of presenting news to the public (the core function of the journalistic enterprise), rather than giving a platform the world’s most emotionally damaged people to say terrible things?

            • Mat Yurow

              Yes. This isn’t at all to say that commenting makes sense for every company (you need to make the math work). But I find that most aren’t doing any sort of value assessment before dismissing them.

        • charlesarthur

          Metafilter has had the “pay to comment” model since forever, and it worked really well until May, when a dramatic drop in traffic (possibly due to a Google ranking update – meant they had to fire staff, including some moderators.

          I continue to think that most comments add virtually nothing to the understanding of the story; it’s more like a Power law or Poisson curve than some neat aggregation of value per comment, where the comment that really adds understanding to a general news story is so rare that you could pretty much hide everything else.

          As to the “warm leads” – yes, excellent point Mat, and you’ll know that news orgs sell more valuable ads when users are signed in to comment. Trouble is, if moderation was properly aggressive, and aimed to really elevate the useful comments and dismiss the useful, most peoples’ comments would never make electrons dance on your screen.

          And that’s before you look at the question of what the purpose of comments is at all. How should the story above them change if new information appears in the comments? We haven’t tackled that question at all, but it’s actually fundamental if you’re suggesting that (even some) comments have value. It’s absolutely essential if you think that *all* comments have some sort of value once you’ve stripped out the spam/porn/racism/ranting ones.

          • Mat Yurow

            Value can be assessed in different ways to different people. To an author, yes, I think most non-spam/ranting comments *could* have value. However, in the point you’re referencing, I’m using value as a monetary term – the probability that segment (commenters) will perform a high value action (subscribe, consume multiple pvs) on your site. In this regard, I’m aggregating the good and the bad together.

  9. I think it would eliminate most of the trolls if comments were not anonymous. I would love to engage in some positive discussions instead of the profanity and downright stupid responses we in so many comment sections. Free speech is abused much to often by people who are destructive rather than constructive.

    • Mat Yurow

      Unfortunately, I don’t think anonymity is the culprit. Check out the comments on any Facebook post by a news brand, and you will likely find much of the same harassment and clutter.

    • A broken fish

      Really? Because there are studies that show this effectively changes nothing, and if anything, silences people who may have “dangerous” opinions.

      If my political opinions were to reach my work, I could probably lose my job. Does that mean I should tie my real name to this comment, allow people to reach my job, and get fired? Simply because I don’t agree with the SJW hugbox going on the past 4~ years? All you do is silence people like me by forcing names. And for what purpose?

      Names serve no purpose other than to silence opinions. My name isn’t me. THIS is me. Unless you want to hear a fake, outward experience “me” that self-censors to avoid ridicule.

  10. John Dineen

    Read Yarow’s Medium post earlier and now this piece. Thanks for highlighting this issue. We spend our days talking to businesses about how much of a bad idea it is to outsource their community engagement to the social platforms. “Got a question for us…. Message us on Facebook or send us a Tweet”. Reading that more media sites are turning off their primary option for onsite community engagement is just incredible. How much of their cake are publishers going to let Facebook and Twitter eat before they wise up? Your fans / visitors / customers are your business. You should be increasing the opportunities for them to engage not removing them.

      • Mat Yurow

        Sure – We (NYT) did cookie analysis to determine what the probability of a new registered user (and/or commenter) was to subscribe. I also know that the life time value of a subscriber is to my company.

        Now, I can take the the number of registered users added, multiple by the probability that they subscribe, and then multiple that by the value of subscription. Subtract and divide the total by the costs for acquiring said users and you have a tangible ROI.

        • Dan Mitchell

          I’m not positive I understand this entirely, but it seems to apply to registered users as a whole, not “commenters,” as such. But if commenters, as such, are part of this metric, are you in any way measuring their levels of “engagement?” or whether this engagement adds in any way to readership or revenues?

          • Mat Yurow

            Same deal. Because commenters are registered users, most sites should be able to track their actions (even in aggregate). If you know the average # pages/mo that commenters consume, you can determine value by backing out of your RPMs.

            Let’s say a commenter of GigaOm reads visits 5 pages per month, and they have 10,000 active commenters. The RPMs of an ad are $5, each article has 3 ads (roughly $15/1,000 pvs).

            Your equation would be (10,000*5*15)/1000. In this case GigaOm’s comments are worth a collective $750/mo. to the business. This is an over simplification, with made up numbers, but you get my point.

            As Mathew stated, it’s easier to make this math work with a subscription supported businesses (higher LTVs).

            • Dan Mitchell

              I’m still not sure I’m following your formulas, but, from what I can tell, it seems to presume that every page viewed by commenter is viewed *because* he’s a commenter. I’m not sure how you get from the metric you put forth here to “comments are worth…” The pages *read by* commenters are worth that, right? If there were no comments section, some (possibly large) proportion of those views would still happen, right

              Also, it doesn’t take into account the costs associated with managing the comments. They’re probably very low for Gigaom, but much higher for, say, Reuters, or the Chicago Tribune, or other sites that are inundated with terribleness.

          • Mat Yurow

            Can’t reply to your below comment, but you’re right. You really should be taking the difference between your commenters and the avg. user (I would bet that commenters read more pages than the average user).

            Plugging that number in (instead of total PVs), would give you the incremental value of those users. Again, I’m not arguing that the math needs to work in order for an investment to be justified.