The Washington Post’s Greg Barber goes in-depth on online commenting

The Washington Post’s Greg Barber goes in-depth on online commenting

In June, The New York Times, The Washington Post and Mozilla announced that they would be collaborating on a project to develop an open-source comment and user contribution platform.

The platform is funded by a $3.89 million investment from The John S. and James L. Knight Foundation, a philanthropic organisation that issues grants to fund innovative journalistic ideas.

As The Washington Post wrote in an article reporting the project, the platform will be build with the aim to categorise and highlight the most pertinent user comments and “standardise the many different ‘community engagement’ systems that Web sites now use to collect and publish outside contributions, especially reader comments and photos”.

Project leads from the three organisations involved in the collaboration are Dan Sinker (director of the Knight-Mozilla OpenNews project), Marc Lavallee (editor of interactive news, The New York Times), Sasha Koren (deputy editor of interactive news, The New York Times) and Greg Barber (director of digital news projects at The Washington Post).

The Newspaper Works had a discussion with Mr Barber about the state of comment platforms online, the ambitions of the project and his own involvement in its development.

Here is the full transcript of our conversation.

*             *             *

Considering that this platform will be open-source, how do you plan on building something that is adaptable to different publications and audiences?

That’s our challenge, the way we’re going to start is by talking to people and we’ve already done that to an extent. The grant that we received from Knight earlier this month, the $3.89 million grant, is actually the second grant that they’ve given us. The first grant we got back in March, which was $75,000 to fund a research portion of the project. So we worked with a group called Machine, based out of New York, and they helped us do some audience research by talking with users, readers and also with publishers.

There were a couple of ideas in play there. One was just to kind of help us firm up our ideas, especially since there were three major organisations coming to the table on this, The Post, The Times and Mozilla. But then also to test our assumptions and make sure that the solutions we thought were going to work for The Washington Post and The New York Times would also work for publishers of different sizes, many of them smaller than us, and also broadcasters. So I think it’s research like that and conversations that we’re having, which are obviously much easier to have now that the project is out in the open, that will make sure we’re gathering the kind of input we need from other publishers.

Was the grant given to an already developed partnership between the three companies? Or were Mozilla brought on later in the process to help develop the platform?

So actually it started out just by editors talking socially, so some of my colleagues here at The Post were talking with friends that had that worked at The Times about the challenges that they encountered in community building and in using community tools and it started striking folks on both sides that, ‘hey, actually we have really similar goals and really similar challenges even though we go about commenting in very different ways’.

Aron Pilhofer, who used to be an assistant managing editor at The New York Times and is now the executive editor for digital at The Guardian, was instrumental in getting us together. We met here in Washington in November, that was our first official meeting on this idea, and it was just to kind of get our staffs together and talk, and what wound up coming out of that meeting was that ‘yeah, we really should work together on this’. Which is obviously not typical for The Washington Post and The New York Times. We just saw that our goals and our challenges were aligning so well and we came out this meeting really enthused about it.

We also thought that it meshed really well with the mission of The Knight Foundation. So we talked with The Knight Foundation about it, we had a meeting in December at The New York Times’ office in New York, with Alberto [Ibargüen] and Michael Maness at the Knight Foundation. It was during that meeting that they suggested, ‘well if you’re thinking about doing this and you’re thinking about open-sourcing it, then Mozilla is an obvious group to bring in because they would bring so much to the table, especially in connecting with the open-source community’. So that’s when Dan Sinker came into the process and Dan of course comes in with a terrific perspective and a ton of ideas. So far, at every step in this process, it’s been just sort of folks going ‘yeah, absolutely we should do this, and here’s how I can add to it’. So it’s really meshed tremendously well.

Do you know the kinds of functions that you’re trying to build into the platform and are there consistencies among publishers about what they would like to see from an enhanced commenting system?

There are a couple of things. One, and this is has been covered very well over the past few years and it’s a perennial concern, is the level of discourse in comment threads. Now we think this project is bigger than just comments, we want to build a system that allows for all sorts of different interactions between publishers and users – so that’s everything from comments, to full blog posts, full articles even, or smaller content in the form of annotations. We want to be able to take in reader contributions of different sizes as far as text is concerned, so annotations on the small side, comments in the middle, blog posts and other longer stuff at the high end. But then also, things like photos and video. One thing that we heard from the publishers, is that they were looking for ways to do that in a way that raised the level of discourse on their websites. Also, in ways that were easy for them to integrate and that were easy for them to moderate. So what we want to create is a system that allows for, at its most basic, a streamlined stream of content, that is easy in, easy storage, easy moderation, easy out. That’s to help on the publisher’s side.

On the reader side, what we heard was that folks are actually, despite what you may read elsewhere, really interested in reading comments. A couple of them told us that they used comments as a way to gauge what real people thought of the news, as opposed to just what journalists or opinion makers thought of the news. So they saw comments as pieces of writing themselves that are valuable. And then as far as a feedback mechanism, the readers that we talked to were very interested in being able to do that, but wanted a way to do that that was an easier conduit without the noise that trolls and other things can bring. So we’re hoping to be able to cover both sides of that spectrum.

What are your views on the standard ‘comment at the bottom of article’ model that has typically been the most common style of comment organisation?

The ‘comment at the bottom of the article’ model is a perfectly fine tool and it certainly has it’s uses, it’s one way that users can respond to a piece of writing or a story that a publisher puts out. But it’s not the only way and it shouldn’t be the only way, what we want to create is a system that allows publishers to choose the engagement method that sits best with their content. So maybe in some cases that’s a comment stream, maybe in some cases it’s annotations. We have some bloggers on that use our comment streams basically as main well content. So they’ll say, you know, ‘here’s an open stream’ and they’ll help direct the conversation, or they’ll direct the conversation with a short blog post that basically just poses a question.

There are probably means of interaction that I haven’t thought of yet, that other publishers might have. We want to create a flexible platform that allows publishers to present their call-outs in creative ways and then to innovate in the ways that they display the contributions they get from users based on those call-outs.

How do you address the issue of potentially having too many avenues for engagement? Can there be too many forms of interaction on an article to the point that it almost becomes impenetrable?

I think that that’s going to be up to the publisher. From the perspective of The Post, we’ll want to choose the avenue that works best for the particular story that we’re trying to tell, but from the perspective of folks who are creating a platform, we’re going to create the toolbox and also some of the tools inside and then it’ll be up to publishers to choose which tools they use and how many.

Workflow wise, how is the whole timeline of the project being coordinated and mapped out?

So it’s going to be a collaboration among all of us, so it’s not that The Times and The Post have brought Mozilla in and they’re going to do the work for us. There are going to be people at all three organisations involved in pretty much every piece of this. So there are project leads from each of the three organisations – Dan Sinker at Mozilla, who’s in charge of the overall project, myself from The Washington Post and Marc Lavallee and Sasha Koren from The New York Times. So we’re involved in charting the full product, defining it and making decisions throughout the process. But then there are also going to be people from each of the organisations that are going to be playing different roles within the project, for example my colleagues here at The Post, Greg Franczyk [principal architect] and Sarah Sampsel [director of digital strategy], are going to be taking lead roles in designing the products architecture and its design. Meanwhile, we’ve got a team of developers from The New York Times that are going to be doing a lot of the hands-on coding work. Then Mozilla is going to help us with quite a bit of the strategy and then there are some Mozilla fellows that are going to help out in other ways. So it really is going to be a full team effort.

As far as coordination goes, we’ve been communicating a lot via different electronic media – we talk a lot over email, we talk a lot over the phone of course and then we’ve had a number of meetings already in New York. The project will be based in New York. We have a team that’s based there from The Washington Post and that’s where Greg and Sarah are based, obviously The Times is also up there, and those of us who aren’t based in New York will be travelling there a decent about over the next few months and years.

Do you know where this thing will be six months from now, do you know where it’ll be in a year?

Sure, I mean it’s an iterative project so our plan is to continue to talk with publisher’s and readers, to scope this out, as seems to best fit the platform, and then my hope is we’ll be able to engage with the open-source community too, to help mould and refine it. Hopefully, as they usually do, once we put pieces out there they’ll be able to improve upon it. So that’s a really exciting notion for us.

Will the platform work as a plug-in to a publisher’s CMS (content management system), or will it operate as its own entity that is accessed separately from the publisher’s back-end?

You’ve got it right actually, it’s going to be a core piece of software that we’ll then build plug-ins onto and that publishers and open source developers and that sort of thing will also be able to build plug-ins onto. So the idea is that it will be like a set of building blocks, with a foundation in the core product and then a series of plug-ins that publishers can decide to use, not use, plug-ins that they could decide to make open source. So hopefully this is the project that never ends, in the sense that hopefully the core will continue to be improved upon, new plug-ins will continue to be made and the platform can continue to be updated and iterated upon to meet our future needs.

To be more specific; say you’re a comment moderator for The New York Times and you’re currently moderating comments through Scoop [The Times’ CMS], would you still be moderating through Scoop with this platform plugged-in to the CMS, or would this operate separately at an external site?

Those decisions will probably need to be made on the publisher level, because there is the potential that there could be some overlap between what we’re creating and what the publishers are creating. Our goal is to create a platform that will be easy to integrate into CMSs and if there are duplicate features, where someone has a system that they’ve built and then there’s the system that we’ve built, our plan is to make the system flexible enough that they can make decisions about what they would keep of their own and what they would use of ours.

Obviously this is a significant financial commitment from Knight, what convinced them that this was an urgent and important area for investment?

I can’t really speak for Knight on that, but what I can say from The Post’s perspective is that our management has been 100 per cent behind this project from the beginning and that we saw it as a very worthy investment of our resources and we’re really thankful for the support that the Knight Foundation’s been giving us.

A lot of post-read conversations now happen on social media, Reddit etc., do you feel like there is some kind of divide between social media platforms and comment threads on online publisher sites that you can bring together through this platform?

I don’t think that conversations on publishers sites and on social media need to be mutually exclusive, and certainly they’re not. The Post alone, I know, brings in Tweets, brings in posts from Instagram, we had a project that used Snapchat earlier this year or last year. So we do lots of integrations with different types of social media. But for me, I think, it’s about ensuring that the conversation about the news happens where the news is reported. It’s publishers who start the conversations, so it makes sense to me that on publisher’s sites is where those conversations should continue. Not that they can’t go onto social media, in fact publishers will reach out on social media to interact with users. I don’t think it’s an either/or, I think the conversation can happen in all sorts of places, but I do think that our hope is to enhance the existing forums for civil discussion and civic debate on publisher’s sites and then also to enhance those, to broaden them, to allow publishers to bring in user contributions in new and different ways. To give them the tools that they need to make moderating those conversations more efficient.

One of the things that we’ve seen that improves discussion and debate in website comment sections, is participation by content creators – by reporters  and editors and that sort of thing. One of the difficulties that we have in newsrooms, in getting reporters to spend time in comment sections, is helping point them toward the places where that time will do them the most good. As you know reporters are busy people and they’ve usually got a couple of irons in the fire, so you get to the point where you say, ‘OK, I have ten minutes, my last story has 2,000 comments on it, where are the good ones?’ And right now I can’t answer that question, because I’ve got plenty of tools that can tell me where the bad ones are, I can suss those out for you, but as far as finding the good ones, that’s not something that we can easily do. So what we’re hoping is to create a form of smart moderation that will use different elements to help us to highlight contributions from our most thoughtful users. Then to be able to present those to reporters so that they can interact. Also, present those to other users, because I think one of the biggest things we can do to change the tenor, and also the types of conversation throughout the site, is to lead by example and show some of the great content that we’re already getting that’s just difficult to find because of the way our means for conversations are structured now.

Is valuing those comments something that you hope to do in an algorithmic way, through word recognition and what not? Or is this something that will be managed directly through users building reputations as thoughtful commenters and moderators and other commenters having the ability to judge that?

I think reputation scoring is more the way that we’re thinking about it and I think you can measure reputations based on a few criteria. By looking at the actual words that have been submitted, so you’re running those through dirty word filters, you can do some semantic analysis on that and that’s one thing that can help you determine comment quality. But I think one of the other things, and I think perhaps the more powerful thing, is to manage by the reader, by the commenter. Because, taking a page out of typical human behaviour, that thoughtful contributors will tend to be thoughtful contributors, folks that are coming to troll are more likely to come back to troll, that the idea of getting an understanding of which of our contributors are more likely to be thoughtful and allowing the contributions of those folks to flow to the top will be one way of starting in this direction.

Is there a concern that this kind of model creates a barrier to entry? 

That’s the thing, do you start new entrants at the bottom of the pile? I don’t think you do. The difficulty that we’ve got right now, is that everyone’s comments are being lost in the mix, we’re starting at a place where there is already chaos. So how do you do get order out of that? In the past, for example, comment threads have used badges, so you reach a certain level and you earn a badge. The difficulty of that is that quality of contributions is not a fixed thing, it depends on the contribution, it can change as more contributions come in. So I think that designing a system that allows for constant re-examination of user reputation scores and, as you suggest, makes entry easy and makes it easy to be noticed if you come in and you’re first comment is a thoughtful one, that is important. My hope is that we have lots of really thoughtful commenters and that reading through comment threads and other means of contributions is an easy and pleasurable experience and that users can take advantage of other means of sorting contributions to discover the kinds of contributions that most appeal to them. We’re hoping to do all sorts of things to allow users, readers and commenters to better connect with each other, notifications that would allow commenters to know when someone has replied to a contribution that they’ve left on the site. We want to introduce features that would allow for lots of different levels of connectivity and would allow for lots of different ways for contributions to be surfaced. So I think the barrier to entry question is one that’s important, but I think our approach is one that I think would not create a barrier like that.

Last thing, the $3.89 million investment, is that allocated for the two years of product development, or also for the ongoing existence of the platform?

The money is allocated for the next two years, so that will fund the first two years of the project, which is the initial creation of the core and the base set of plug-ins and all of that – the initial project. And then we’ll see what happens from there.

For more news from The Newspaper Works, click here.

Leave a comment