The Global Editors Network (GEN) Summit was held in Portugal and Dataminr had the opportunity to host a Masterclass for the delegates. The following is the presentation Dataminr’s Director of News Meghan Plambeck gave at GEN Summit, and a transcript follows.
Dataminr’s early breaking news alerts give journalists a head start on covering the stories that matter most to their audience. Using proprietary algorithms to analyse all public Tweets in real-time, Dataminr delivers the earliest tips to events as they are just emerging.
The following transcript has been edited for length.
Thanks so much for coming today. I'm Meghan Plambeck, Director of News at Dataminr, and today we're going to talk about separating news from the noise. Before I joined Dataminr, I was a journalist. I spent almost 10 years in newsrooms in both New York and Washington. I worked as an assignment editor, managing editor, a web producer, a booker, a control room producer. So I've seen a lot of what happens in newsrooms. I am very familiar with some of the challenges that you're up against. I love hanging out with journalists, so please come and say hi at some point today if we haven't already met.
When I joined Dataminr, the product was brand new, it was in its infancy. It was an idea that we had with Twitter about how important it was for journalists to get information faster from Twitter. As a lot of you know, the Twitter dataset is massive. There's about 500 million Tweets every single day. It is very difficult, if not impossible, to find the most relevant information on Twitter without a tool like Dataminr, and we'll get into that in a second. But I saw the product start at CNN, they were our alpha partner, and I've had the chance to watch it grow from one newsroom into what's become an indispensable tool for more than 400 newsrooms around the world.
For those of you who are not as familiar with Dataminr, we provide journalists with the earliest indication of breaking news, often significantly ahead of other sources. Many, many journalists tell me the first time they hear about an event happening, they're getting a tip from Dataminr.
Just a quick agenda setting item here so we know what you are in for today. We're going to start by talking about discovering newsworthy content on social. I know that that is not a brand new concept to a lot of you, but it's still has challenges and I think that Dataminr offers a really unique and interesting solution here. From there, we'll talk about how some leading newsrooms are taking advantage of the early information that they receive from Dataminr. That will take us into a bit about workflow and how important good workflow is when it comes to news gathering, especially around breaking news. Part of that will be a discussion about how Dataminr integrates into the news tech stack. For those of you who aren't familiar with the news tech stack concept, essentially what it is, is all the different technologies and platforms that you are using in your newsroom to make your workflow more efficient and to gather news more efficiently.
Finally, we'll take a look at how one digital newsroom has been relying on Dataminr alerts to significantly drive traffic to their site, which I know is a challenge for many digital properties. Then we'll take questions and we'll have a little bit of breakout session with the representatives from Dataminr in the room, and we can meet with you one on one and talk to you about how newsrooms in your markets are taking advantage of these alerts.
Let's get started. Like I said, there are 500 million Tweets every day. That's a lot of tweets and it can be very difficult to find the ones that contain breaking news. If you're like me, you may have tried to set up a TweetDeck feed or a Twitter feed that said #breaking, and then hope that you were going to find something that was helpful. But the problem with that is that if you type in #breaking on Twitter, you know that the information is so overwhelming that it's almost useless because you can't find what you're looking for. And the truth is, you never know who's going to break news. I remember setting up TweetDeck lists and I put in all the reporters in my region, and I put in all the news outlets, and I thought I was going to get all this great information, and sometimes you do. Sometimes news is broken by journalists like you who have scoops, but sometimes it's broken by random people that you would never follow on Twitter because you don't care what they tweet about until they happen to tweet about something that's relevant and of interest to you.
So the way that we like to think about Twitter -- and I should note that Dataminr for News was co-developed with Twitter -- it's a great dataset for us to work with because that's where news breaks first and that's where news breaks fastest. Twitter is essentially a global tripwire made up of this really, really diverse set of sources. Like I said, it could be politicians, reporters, local reporters, celebrities, citizen journalists, and I emphasize the citizen journalists because I think that's where a lot of times the richest, most interesting content can come from, but it's also the hardest to get your hands on as it's happening.
I'm going to go through some examples here. What you're seeing is the text of Tweets that Dataminr detected in the overall Twitter dataset and then delivered to its clients. Some of these stories you may remember, some of them you may have even covered. What you'll notice is that it doesn't sound like a news outlet [published it - which is the point. None of these tweets, with the exception of this one, contain the word news, and this is just news in reference to SABC.
You'll also see some that come through that have emojis. Lots of people Tweet with emojis these days, different languages. You can't speak every language. If you do, that's great, but Dataminr automatically translates everything. So even if a Tweet comes through from someone who's speaking a language that you don't speak, you can still receive that content. Oftentimes, these Tweets will often be accompanied by rich media, photos from the scene of an event. What's nice about the images that Dataminr pulls out of these Tweets is that you don't necessarily ... it's telling you that something is happening, but also giving you these really powerful visuals. Because sometimes it's hard to know, is this actually something that's a big story or not? Having those photos delivered directly to you really makes an impact and lets you know, hey, something is going on here. You don't have to go searching for any photos or videos. Oftentimes, that information will show up directly for you.
So we've seen how powerful these eyewitness reports can be. So let's talk about how some major newsrooms are operationalizing them, how are they using them in their workflow? We'll start with the New York Times where they rely on Dataminr alerts to turn breaking stories around as quickly as possible. One of their breaking reporters, Jonah Bromwich, and the way he describes it, Dataminr is a tool that they actually use most prominently to stay on top of breaking stuff. So he says that the alerts he gets from Dataminr are "often tweets that we might not otherwise have seen because they come from, for example, a cloud computing consultant in Houston, who has 43 followers on Twitter." I wouldn't see that on Twitter for hours, but Dataminr would bring it to my attention right away. I think that's an important distinction. It's because there's so much information out there, it could be hours that go by before you realize there's a major news event happening because you weren't following the people that are tweeting about it in real time, and that's sort of the power that Dataminr delivers.
You can see on Jonah's computer there, that's what Dataminr looks like, and it's delivering to you tweets, it's a live Twitter card in his screen, but also provides contextualizing information and we'll talk about that in a second. It's important to note that Dataminr does not provide verified information. It's the earliest tripwire of an event that's taking place.
As another example, CNN, they were our alpha partner, like I said. We built Dataminr for News with their input, and we've worked really closely with our teams ahead of our launch and to market, and we're still really deeply embedded into all of their newsrooms. At the end of last year, I spoke with Richard Greene, who is a senior editor in their London Bureau. And one of the things he told me was that CNN allows them to be very competitive in places on stories where they might not have resources on the ground. I know that's a challenge for a lot of newsrooms who are up against shrinking budgets, they can't have reporters in every place, they can't have a reporter on the scene at every event. But because Dataminr delivers eyewitness reports directly from the scene, it gives them the opportunity to sort of be on the same playing field as those people who might have a reporter there, who might have a camera person there. And instead of having to follow other reporting, they're starting at the same point, essentially.
I've also overseen our relationship with the EBU Social Newswire, and their use case is very interesting. They're very lean team, small group of them, and they rely on Dataminr alerts for tip offs, so that they can start looking for user-generated content that they can clear for usage for their partners to use on air.
This example here is the Mandalay Bay shooting in Las Vegas not too long ago. They were able to mobilize very quickly because of the Dataminr alert and acquire their first piece of content in the first four minutes of the shooting happening. So Derek Bowler, who's the Head of Social Newsgathering there, said that by the time that this story was alerted by the wires, 18 minutes later, they were not chasing the story, they were ahead of the story. They were well ahead of the story because in the first hour, they had already put out more than 10 pieces of content for use by their partners, and by the time the story was four hours old, and some groups were just catching up, they had already put out 70 original pieces of content. So that extra time, that extra breathing room gives them the opportunity to mobilize faster, to get their resources in order, and to deliver better for their teams.
Here we are talking about how Dataminr integrates and how, I think, workflow in general is very important. For the team at the Social Newswire, the key way that they rely on Dataminr is through a Slack integration. But as we've been developing this tool, it's understanding how important it is for journalists for the information to find you. Instead of you having to manually look for content, or to spend valuable resources looking for information, it's supposed to flow into your existing workflow so that you don't have to do an additional step, you don't have to go looking. This is the full suite you see here, desktop, web, email, mobile pushes, Slack, TweetDeck, if any of you use TweetDeck, and also an API.
I want to stick with Slack for a second here. How many of you in this room are using Slack and your newsroom? Okay, awesome, a bunch of you. I'm seeing more newsrooms are starting to embrace Slack. I would say more than half of our clients are using a Slack integration to receive their Dataminr alerts. One of the challenges that I hear lot from news organizations is that during breaking news, it can be difficult to collaborate because people get very excited and everyone wants to shout out, "Oh, did you see this? Did you see that? Do we know about this? Do we know about that?" And ends up with a lot of duplication of effort with the Slack channel. And what HuffPost does is they actually put their Dataminr alerts into the Slack channel, you can see them flowing in here. And then all the reporters join the channel. So it makes for really easy collaboration because everyone is on the same page already. There's no one that's left out, everyone's seeing the same information. They can all stick together and not duplicate their efforts because everyone's seeing the same stuff at the same time.
Behind Slack, I would say that TweetDeck is another very popular way of using this tool. How many of you are still TweetDeck users?. I love TweetDeck. You can see here that you can integrate Dataminr, it's the middle column there in your TweetDeck. And you'll notice that it can deliver again, it exists next to your existing TweetDeck. It's not meant to replace TweetDeck. Your TweetDeck is awesome if you know what you're looking for, if you have sources that you're following already, if you know what keyword you want. Dataminr is meant to be a supplemental tool to that. It will, again, pull in those images, and photos, and videos, content from sources that you don't have any idea exist, sources that you don't follow, but who might be providing valuable information.
Then as a comparison, you can see this particular example is taken on the day there was a large explosion in Benghazi. If you look on the far right column, that's a keyword search of Benghazi, and you can see how many Tweets there are. Well, I don't know if you can see, but it's like 35 a second or something crazy, so an overwhelming amount of information.
Going back to this idea of delivering content directly into your workflow, I think it's very helpful during breaking situations. When a new story is breaking, people are coming to your website. They want to see an update, they want to see what's going on, and they trust you, they're coming directly to your site. It's unfortunate if you don't have something up for them to look at when they get there, because they'll go someplace else. They'll see who is reporting on a particular breaking story. These Context emails from Dataminr are really valuable here. A lot of newsrooms tell me that they rely on these to get something up on their website, for example.
So the Context emails deliver, again, eyewitness sources around a particular event directly into your inbox, and with one click, you can put them into your CMS and have them appear on your website headline, so you can grab that SEO traffic, and then you can continue to update, but at the very least, you have something published. And again, it's content you don't have to look for, its content that shows up for you, and you can customize from there. So on the left you have, looks like bad weather in South Africa. And then on the right, you have people responding or reacting to the death of an athlete, I think that's the example there. But the point is, is that it's information about the event that shows up for you.
And lastly, mobile. So again, when news is breaking, you may need to be dispatched out to the scene. With this mobile app, what you can do, is a story comes through as a push alert, you can, with one click, opt in to receiving all the updates about it. So here we go, click, track story. Off I go to the scene. While I'm en route, Dataminr is pulling out all of the relevant Tweets about that story and putting them in a chronological timeline. So by the time I show up at the event, I haven't missed anything, I know exactly what's already happened. It will include obviously other eyewitness sources, but official sources and handles as well, whether it's emergency responders, government officials, anyone that might be reacting. So you get kind of this full picture as you're on the scene, and you can really hit the ground running from there.
Now we've talked about separating the news from the noise, how these alerts can integrate into your workflow. What does all of this mean? We know that when it comes to bringing traffic to your site, you can't really rely on social distribution of content alone. A lot of leading newsrooms rely on search rankings as a core part of their growth strategy. So how can Dataminr help with that? We spoke with Inverse.com, which is a digital media startup based in the U.S. They have a monthly readership of about 50 million, and they focus on Millennial men as their audience. So for them, early and original content really helps to fuel their growth strategy, which is focused on those top rankings and search on Google and Reddit. Their challenge was finding more original stories that they could be fast on that weren't out there yet, that they could package up in a way that really resonated with their audience.
So what is the solution? The solution is that they took Dataminr, they customized their Dataminr feed based on the topics that they care about. So for them, the stories that do well, our science stories, technology, culture, and entertainment. They filtered their Dataminr feed for just those topics and waited until the alert showed up in their Slack channel, and then they were able to collaborate around the types of stories that they wanted to package with their audience in mind and pushed out quickly. So what did they get as a result? Executive Editor says that, again, Dataminr surfaces stories that they might have missed, or stories that have a national angle that they can then repurpose for their audience. And they said it's really powerful because they have Twitter lists with news makers, but Dataminr services voices from parts of the country or cities that they're not plugged in to.
They used to rely a lot on local news reporters. I don't know how many of you follow local news reporters. I love local news. But the local news reporters that I follow, they tweet constantly, all the time. If you look, they have like 35,000 tweets. It's a lot of volume. With Dataminr, it will flag up those ones that are considered algorithmically significant, and it gives them a narrowed down version of stories that they might find interesting.
In this particular example here, Inverse found a story that they would not have been able to cover themselves. It was what happened at a conference. It was the Bloomberg Global Business Forum last year that they weren't invited to, so they didn't have a reporter there. But someone who was in the audience tweeted a picture of Bill Gates sitting on stage and he was re-sharing an anecdote he had about the one thing that he could have changed about Windows, and inverse said, "Oh, that's a great story. That will definitely drive traffic." They took the story, wrote it up quickly, embedded the tweet, pushed it out, and it drove a significant amount of traffic their way. So what you have there is the early Dataminr alert that gives them that extra time to react, the savvy headline writing that's going to draw people in, and then editorial team that understands how to optimize those to drive search to their site. And for those of you who are wondering, Bill Gates wanted to redo Control-Alt-Delete.
We all acknowledge that it's important to be early news, and it's getting harder and harder and harder, so this is an example of a model that is working for Inverse, but it's also something I hear in a lot of newsrooms. This early awareness so they can then jump on SEO search has been very, very valuable.
Dataminr team, can you raise your hands? Rebecca, Nick, Jonathan and Kirsten work really closely with a lot of our news partners in various parts of the world. So if any of you have any questions about how newsrooms are using this tool, and if you want some better examples, or some additional examples of how newsrooms are using the tool, feel free to grab one of us, we'll be here. We also have a booth, number four across the street, and we're giving away a pair of headphones today. So make sure you stop by and drop off your business card, and maybe you'll go home with some new headphones.