Using Analytics: A Round Table Discussion
Moderator: Gene Allen
Marissa Nelson, Senior Director of Digital Media, CBC News
Erika Tustin, Digital Editor, Toronto Star
Kenny Yum, Managing Editor, AOL Canada, Huffington Post Canada
The rapid and profound changes affecting journalism make it difficult for journalism educators – many of whom have not worked regularly in newsrooms for several years – to keep up with new developments. In the continuing digital transformation of journalism, one of the most important recent changes has been the ability to receive rapid and essentially constant information about audience response. Using measures known as “analytics” or “Web metrics,” journalists (and others communicating information and images in digital space) can learn a great deal about such things as where website visitors come from, which articles or other forms of content they consume, when and how often they return, and so on.
The growing reliance on analytics has elicited considerable interest among scholars, who see it as representing a significant change in how journalists go about their work (see, for example, Tandoc 2014; Schlesinger 2014; Vu 2013; and the paper by Paul Benedetti in this volume). Issues such as the role of journalistic standards; concerns about dumbing down content to attract larger, if perhaps less committed, audiences; privacy; and the erosion of the boundary between journalistic and advertising or circulation decisions have all properly been raised, and continue to be raised.
This round table discussion was intended simply to provide a primer on analytics as the practice was being used in Canada as of mid-2014. It is more descriptive than evaluative and represents the point of view of practitioners who are deeply engaged in the use of analytics at three different journalistic organizations. One is a relatively recently established, digital-native outlet, The Huffington Post Canada (www.huffingtonpost.ca). The second is Canada’s long-established public broadcaster, the Canadian Broadcasting Corporation, which has been aggressively expanding its digital presence (www.cbc.ca); the third is Canada’s largest-circulation newspaper, the Toronto Star (www.thestar.com). My hope is that this will give journalism educators a better idea of what exactly is involved in the use of analytics, so that they can make better-informed decisions about how and to what extent this practice should form part of their evolving curricula.
Gene Allen: The first question is a very simple, descriptive, factual one: What is analytics? More specifically, what sorts of information about your readers, listeners, and viewers are you able to collect? And what do you do with it?
Marissa Nelson: We collect quite a bit of information. It depends whether you are logged in as a member or not. If you’re logged in as a member, then we obviously have a lot more information, but generally we track device use, browser use, operating systems. We can tell what product you’re using to consume our content. We also track referrals; we can also see internal traffic patterns. So we can see where you came from and how you go through our site – you went from this page to that page, to this page and then back to the homepage, for instance. So it’s internal traffic patterns, and if it’s a member, we collect much more information obviously; things like e-mails, and sometimes there’s demographic information, though the bar [the amount of personal information required] is relatively low in terms of signing up for a membership.
Erika Tustin: That’s very similar to the Toronto Star. We have a member sign-in as well, My Star. When we use that option, we can collect a lot more information than we typically do just based on IP addresses. Also, within our commenting system we have a social sign-in aspect as well – if you use that, that also allows us to collect specific information on social [media] from you, based on what your privacy settings are.
GA: Does that mean, if you come in from Twitter or Facebook or something like that?
ET: Correct. You can use your Facebook login to sign into our commenting system. Generally along the same lines as the CBC, we collect data on what we know our readers like to read, when they like to read it, what platforms they read it on, how they’re accessing our journalism. From what country, from what website, and their internal traffic patterns within our website as well.
MN: These are all very typical. They’re tracked in Omniture, which is the tracking tool. It isn’t journalism-specific, it’s a Web-specific thing.
Kenny Yum: Just the fact that you have three journalists talking about analytics represents a massive sea change: businesses, newspapers, broadcasters, digital plays are putting their frontline journalists at something that, twenty years ago, you would say journalists would not be involved with, which is that we’re in tune with our audience. A lot of the discussion we have about analytics sounds really geeky, because it’s all about numbers, but the flip side for me as an administrator and from a strategy perspective is that for the first time you have journalists actually taking a look at these numbers. Twenty years ago, you would judge a great front page by how many newspaper stand pickups there were. That’s exactly how they used to calculate those metrics. We have way more data now, and not only can we act on it on a day-to-day basis, but we can actually make strategic decisions on how we track readers and understand their patterns, because it helps us serve them better. When we say tracking, it’s based on everything you do. The way I describe it is, how do you interact with our brands? I run many brands: one of them is Huffington Post Canada, and one way you can interact with us is on our homepage and on our story pages, but a lot of our readers interact with us on social platforms. So when we talk about analytics, we’re also talking about your engagement with us on social platforms, like Facebook. For example, right now, Saturday morning [May 31, 2014], I’d say there’s a lot more Facebook traffic happening on mobile with our brand than on our website. Why? Because, except for my esteemed colleagues here, everyone is outside enjoying their day, looking at their phones and using platforms like Facebook to consume information. It’s not just about what is the top-read story on your site, and that’s the end of the discussion. It’s a lot deeper than that, and we spend a lot of time thinking about combining the data to give us actionable items to either act on now or to say, where are we going to be a year from now? Where were we a year ago? And to find those trends.
MN: There’s even a Facebook bump late at night when everybody comes home from the bar. And the kinds of content people are going to consume then are very different. Weekends are, for sure, a mobile play because nobody’s on desktop. So then your focus, in terms of what you’re doing and what your staff are doing, is going to be very different.
GA: When people are coming in through Facebook, does that indicate a higher level of engagement? How many people sign in as members? Is somebody coming in through Facebook different from somebody just happening on something and seeing it once?
KY: You’re asking basically an audience question: What is the difference between a person who subscribes to the Globe and Mail and gets it six days a week, versus one who may pass by a newsstand and pick it up? The way you’re asking about Facebook is the way people like us think about our audiences in different slices. And when it comes to analytics, we actually have a good view of our audiences in different contexts. So I have a reader who may come to my homepage every day for a month and they’re really loyal and their engagement on our site we can track; we understand what that reader is like. The Facebook audience is about how the social platforms, in general, drive interaction of our brands with the consumers. So they may not be as loyal, but, you know, a reader changes. A reader may enter your realm once through Facebook and within five times they become loyal; the same reader is now a bookmark reader. It gives us a big opportunity because we spend a lot of time creating our gardens and we measure who comes to us. The era right now we’re in is not a passive one. We are actively finding the readers, and the analytics is about understanding the Facebook behaviour. It helps us understand, do we have the right strategies to serve these audiences?
MN: On Facebook I think it’s really interesting too. There’s the brand-loyal person – who I would call lovers – and then there are the people who are just breezing by, who I would say are the snackers. The interesting thing about Facebook is it that it plays off both. I’m exaggerating the polarity of those two types of users, but Facebook actually plays to the lovers and the snackers. So, Facebook is great because you can make stuff viral; you can get it out and share it beyond your own influence. You can get people who don’t know you, who don’t know your brand, who don’t even care about your brand, to actually see your content. So you can hook in the snackers that way. But then, once they’ve actually liked your brand, are engaged with you, or asked questions, had a back-and-forth, then they’re becoming lovers and they’re starting to develop a relationship with the brand. So it plays off both of those two halves.
GA: How do you use the information about analytics specifically, in terms of your journalism? Does it help you decide which stories to do? How to do the stories? How to play them? All those newsroom journalistic decisions – how does analytics play into that? And does this help serve readers better?
ET: For us it’s a little bit of all of that, but first and foremost the Star very much is guided by its Atkinson Principles [http://www.thestar.com/about/atkinson.html]. We never make decisions on stories based on analytics. Where we would use the analytics is [in terms of] presentation and platform, and the type of content that would go on one particular platform over another. So, for example, a story that we would particularly conceive for a mobile audience wouldn’t ever see the light of day on desktop or in print. And that’s where we’re focusing our efforts; it’s mainly in the presentation aspect. Also, the analytics allow us to understand the particular type of content we can serve to those readers at specific times on those platforms. For instance, on the presentation side, we struggle very much with our long-form. We do a lot of long-form pieces, a lot of investigative pieces; they resonate quite well in print. Desktop, mobile, tablet, it’s a struggle. That’s when we use the analytics to determine presentation, placement, time of publication, all those things that help us ensure that that piece of important journalism will get into the largest possible number of hands …
GA: Can you give an example of a story that would be on mobile but not another platform?
ET: Our mobile usage is quite high in early morning commute, around six a.m. We cover Toronto transit because there are a lot of subway shutdowns; they shut down for half an hour and then all of a sudden they’re back up again. That would generally not be something we would cover in a digital newsroom, in terms of putting it on the desktop. This story is basically gone by the time anyone would be reading it. That’s definitely a social/mobile story that we would quickly put together in order to keep the public updated as they’re moving throughout their day, and then all of a sudden it’s gone. Unless it turned into a massive water main break, that then flooded the entire city, then we might see it in the paper the next day. But the shelf life of that story is probably about a half an hour.
KY: And you’re driving, I’m assuming, the success of an early morning commute story based on what you’ve seen happen when you put it up. So if you just put it up and say we’re serving a market need, in theory, the analytics gives you that layer to say, how many people were reading it? Is seven o’clock the sweet hour or is it eight o’clock? And to ignore all that, I think, is being blind to ways of better serving your audience. The fact is that more people come from Twitter or Facebook than are coming to your homepage. Using the data, using analytics to prove that what you’ve done has worked, will help you hone that strategy …
MN: I think Kenny was alluding to this earlier too – having data drive audience strategies isn’t a new thing. In TV we have diaries and BBMs [Bureau of Broadcast Measurement results], there’s NADbank [Newspaper Audience Databank Inc.], these things – we’ve always wanted this kind of data in every other medium. It’s just we have it much more, so we can use it more strategically. And BBMs and diaries and NADbanks come out so infrequently that it’s very hard to use them in a day-to-day way. In my newspaper days, you were going after a particular audience, whether it’s male or female, you were going after a particular demographic and you were going to use research to judge whether you reach that demographic and you would assign against that demographic. So if you’re going after women twenty-five to thirty-four you’re going to find stories based on that demographic. We just have the ability to see these numbers in real time. We haven’t even talked about things like Chartbeat – live numbers that actually tell you when a story is spiking. So now, with something like Chartbeat, you can tell when you’ve got a Reddit bomb and suddenly you’ve got this huge peak of traffic. We had one last week that was this year-old story about a stranded orca. You can see some of these strange anomalies in real time.
ET: When we see those spikes at the Toronto Star we dig into it, we investigate why the story might be resurfacing after two or three years of being hidden away. Sometimes it’s a bit of a fluke; most times, though, there’s something that is on the periphery of becoming news at some point, so we take that opportunity to then drive that story forward in whichever way – creating a new piece of content while also acknowledging what is happening with the old one.
GA: I’d like to ask you to talk more about how analytics affects the way the journalism is done. You were saying, Marissa, that this is nothing new, but differences in quantity become differences in quality, and the radical changes in the time structure, the way you can time information, are quite striking. So how does this inform the journalistic decisions broadly?
MN: If you’re going after a particular demographic, then you can get feedback immediately about whether you’re actually reaching that demographic. I think we’ve been really successful, for instance, in engaging our audience more in things like politics, that we feel are an important part of the civic vertical. We want people to engage with political coverage. So a simple setup piece to a debate in the House of Commons would traditionally get very little audience. Well, now we can try to package that in a variety of ways and it can actually hook a lot more audience. So then you realize, oh, Canadians really do care. Politics for CBC is a huge vertical, it drives a ton of audience. And the conception that Canadians don’t care about politics, from my perspective, isn’t true. But we can find ways of reaching a broader and bigger audience to get people engaged with important political stories in how you package it. For those of us who were in newspapers a decade ago, it was the alternative story format, lists or Q&A’s or pro–con lists or charticles or any of those things. Those are all things newspapers have tried. It’s just that we get instantaneous feedback so we can figure out that a story that you post on Facebook with a photo is way more likely to be picked up than if you just post a headline. So you can start to find ways to actually hook people in.
GA: In the case of a setup story for a parliamentary debate, how would you set it up to work more effectively?
MN: I would probably do the old tried-and-true list. You know, the five important things we need to know about this bill. Or a Q&A with the minister who is presenting the bill. Those would be the two easiest, but we tend to do either lists or charts or alternative story formats in setup pieces in particular, because setup pieces traditionally don’t engage audience at all. But sometimes setup pieces are particularly important when it’s a complicated topic.
KY: I’ve been in newsrooms long enough to say that I’ve been in newsrooms long enough. You think about veteran editors-in-chief in the newsroom – they sit back in a chair and they’ll make a declaration about “Here are the important things we need to do,” or “Here’s a big story, I know exactly the things we need: feature story, column down the side.” When you think about how they come to those decisions, to be able to sit back in that chair and make that call – what is the correct package to cover a big story that’s A1? They’re taking a body of knowledge that they’ve accumulated through their career. What they’re telling you is, “I learned something through covering conflicts and covering fifteen budgets and five federal elections; that’s how I operate as a journalist who makes journalistic decisions.” I want to be clear with you, that continues to happen. I sit back in the chair and I say, “Here are the smart things we need to do, here are the stories we need to cover.” One of the tools I use day to day is also understanding what analytics tells me about the audience and how stories actually resonate with them. So yes, when I’m thinking about the big fall features that I want to plan for my audience, when I think about how to cover the Ontario election that’s a few weeks away, I’m thinking about A, what are the great things I need to do for the brand, as a great EIC would typically do. But I’m also thinking about, “What did I learn about the past fifteen elections that I covered?” The analytics have shown me everything from when it spikes, what are people looking for, keyword research to understand what people are looking for. Google is a platform just as Facebook, is, and [analytics] gives me more confidence than any EIC who has no idea about their audience, who has no idea what people are consuming. What they’re really relying on is, “I know it all.” And I think that’s a bit of an arrogant approach to our readership, especially an evolving one. You have to find a way to mix real data, and you can make your decision about what you want to do with it. I can look at data, you can look at data, and we can go in two different directions. I could decide I don’t want to cover sports, period, because my audience doesn’t want it and that’s a recommendation I can make, and focus on these [other] types of things. When it comes to the day-to-day conversations, when we’re encountering big breaking stories, when I’m supporting my staff, what I’m doing is looking farther down the field, as opposed to, oh, is that lead okay? My question is, “What are the five, six steps to move it? How I might disseminate the news?” And then, “Can I measure that traffic?” And then, “Can I give more intel to my staff so they can act on it?” … I learned through my career of actually covering the same thing that here are the things you need to do – and by the way, if you’re going to ask me why I made that call, I have the data to show you. And that’s permeating everything we do. I don’t think there’s an analytical side, then we turn it off and become journalists again. I think it actually informs how we act every day.
GA: Some of the points that Erika and Marissa made struck me as very interesting – about how you use information about the audience to make decisions about when you run a particular story or how you run particular story and how you present it. But what Kenny was just saying about sports made me wonder – are you using it to decide what to cover and what not to cover as well?
KY: It helps you. One of the more interesting things we’re seeing, in the digital space especially, is other sites that are quasi-news taking space away, and what they’re really taking is audience. They may not, in the end, take away audience from you, but you have to understand as a news organization where you’re going to win. We all understand scrums, the reason why everyone is sending the same reporter to basically the same interview. And we have to understand, where is the battle? Some organizations say our proposition is that we’re going to be everywhere: our mandate is broad. Some organizations, like ours, say we want to reach large audiences and we think we can win with these topics. So for us especially, from a start-up perspective, we can launch news sites targeting audiences that we think we can go after and we know there’s a type of strategy we can do. We can have a parenting site that we’ve launched, and we know specifically the audience demographic. It means we can be really focused. Trying to be everything for everyone is a difficult thing. I’ve worked in places where – the typical newspaper, you open up it up to all the sections and it gives you that, “Well, I guess we have to do every single one well.” But then you go to a place like Huff Post, where we actually have more than seventy or eighty sections. We have a wedding section and a divorce section, together, and why do we decide to do that when a newspaper would package it as lifestyle? We figured out there’s a lot of niches we can fill.
MN: You can use analytics in the micro day-to-day, minute-by-minute and second-by-second, and then you can also use the analytics for the 50,000-foot perspective, which tells you a bit about your brand – who people perceive you or your brand to be, and whether you like how they perceive your brand or not, and how you might want to shift it or enhance it. The CBC, for instance – CBC News doesn’t cover lifestyle. We don’t do recipes and that sort of thing because that’s not something people expect from our brand. In digital we did do it for a little while, and it didn’t do very well. So you can make story choices based on what you know resonates with your audience or where you want to move your audience, and then you can make more strategic ones about specific verticals where you see opportunity.
GA: Can I ask about individual curation? You’re mostly talking about aggregations of people, large numbers of people liking things at a certain time of day or liking a certain kind of article. I’m wondering about things like when I go to Amazon.ca they say, you got these last six books, so you’ll probably like this. To what extent are you able or do you want to be able to curate what you present to individual users in that way, and how useful is that?
MN: It will be interesting to see in the next five years how algorithms play into journalism. We’re typically not using algorithms. We generally curate our line-ups or our homepages, and it’ll be interesting to see if algorithms play into it. Some people may have seen the New York Times leak of their digital document that talked about the homepage being dead. Curation focuses on a very traditional audience pattern where people come to your homepage and consume what you’ve put there. And what search and social is doing is really disrupting that whole consumption pattern, so instead of consuming it in the way that you want, there are now independent little pieces of content that are out there. So there are two sides to that question. I think we will see news organizations play with algorithms and I think that will be interesting to watch. The other side, I think, is – you know, if we’re managing a homepage, we can make sure there’s broccoli and cheese. If people are landing on the story page [from somewhere else], your ability to control their news platter isn’t there. And that then drives design because we want to make sure that that story page is a homepage, that there’s a whole bunch of discoverability on that homepage.
GA: Algorithms are the means by which you can do that individual level …
MN: No, algorithms is the Amazon – Amazon figures out what you’ve bought and then serves up more books like that. I think there will be news organizations that observe your traffic pattern, so if you consume a politics story then on your platter when you come to our homepage, you’ll have more politics.
KY: That’s already in place. Think about your social network as your algorithm. You have told Facebook that here are the three hundred people that you trust, and a lot of sites can integrate that. So there are fifteen other articles that your friends read on the Toronto Star. And maybe because Bob liked it, you like Bob and you understand that there’s stuff Bob likes that you like too – and then you’re in there. So from that perspective – related content when you’re landing on our pages – algorithms are used throughout. On the business side, serving ads to an audience, that is already there. When you talk about, for example, wedding content, it’s not always that we’re creating a wedding page everyone visits. People are sharing wedding content, the advertising side is packaging it together, and I don’t know how [the audience is] going to get to it … A lot of us can’t wrap our heads around the fact that our product – we bake it, ship it, and then they have this product they’re going to read front to end, just like a newscast, right? You’ve got to sit down at six o’clock and you’re going to watch the whole thing. This is not the way people consume information, believe me, this is not the reality. There is a specific type of audience that still likes it, but if you want to reach Canadians in general, the twenty-seven million-plus who are online, most of them do not. We’re really fragmented in terms of audience. So we’re adjusting to how we can be part of their lives, not just by curating a homepage that they’re just going to visit once a morning.
MN: The curation aspect of traditional landing pages is becoming less and less important.
ET: That’s very similar to our experiences. The Star’s traffic is more desktop than mobile at the moment, but we’re looking strategically ahead to do a major shift. The curation aspect of that desktop page certainly is something we spend time thinking about, but more in terms of what we will do with it eventually, once the shift happens in terms of our audience to mobile, to social, to tablet, and then what does that desktop become? What does it look like? Who does it serve?
GA: What role do you think analytics will have in the evolution of journalism over the next five years, and are there any major new developments that you foresee?
MN: I think it’s going to change the attitude of traditional journalists towards their audience. Hopefully that’s already happened, but if it hasn’t, it will. I was on a panel a while ago with Zach Seward from Quartz – and if you haven’t gone to Quartz you should use qz.com, great site – he now has a great blog called Glass about the future of television [http://glass.qz.com/]. For their journalists, what they say is, “Okay, you’ve published your story, now what are you going to do to go and get the audience?” And that speaks to analytics. You have to understand how to go and get the audience. Publishing it is step one. I think you’ll see journalists – instead of just writing, thinking, publishing, gathering, that being their job – hunting down audience. [It] will become a key part of their job, and analytics will drive that.
KY: I have a story to illustrate that point. In my previous life I worked at the National Post, where in 2007 we created our own site called thenationalpost.com. And one of the things we did was retrain our newsroom [so that they can understand] what they can do to embrace digital storytelling. Twitter came along, and we spent a lot of time teaching journalists what Twitter was, but it’s safe to say that in those early days people like sports reporters did not get it. Fast-forward years and years – I’m seeing some of the people who initially were deeply sceptical about something like Twitter surpassing the number of tweets I have by five times. And when I speak to them, they relay things to me like, “Hey, do you know that story I did at the Olympics? So I tweeted it out and blah, blah, blah …” And then fast-forward: highest traffic ever. And he looked at me, saying, “It was so great to get audience …” The different aspects of it became engagement, him understanding who was his core audience, talking to them, understanding their concerns. He may not make a decision based on that about what stories he’s going to cover, but at least he has a deep understanding that his e-mail inbox is now Twitter. Now he knows he can influence the success or failure of a story by being on [Twitter]. What he’s doing is analytics. He’s looking at the success rate of – if he’s going to tweet live during a game, why is he doing that? Shouldn’t he be writing his story? But you have the engagement. You’re sharing a story, and your readers are saying, “I love this person, they’re talking all the time, they have great insight, and at the end of the day I’m going to read their story.” And you can measure that. You can go look at any sports reporters – I encourage you to look at them, I think they’re doing some of the coolest things on Twitter, specifically, because Twitter is a live-medium type of platform. It’s a lot of fun to watch.
MN: I have a great example. the fifth estate did a fantastic piece on sugar. And as part of the roll-out, they decided they were going to try to seed it on social, so they tweeted at Jamie Oliver and Jamie Oliver retweeted it and it was absolutely massive. The long tail for the traffic was huge, and it far surpassed all of our expectations. And that’s that fundamental shift from being passive – we published, that’s it – to actually going out and getting the audience.
ET: We just did a major training at the Toronto Star on social media. It was 180-odd people going through Facebook, Twitter, and different social platforms. It also opened up their analytics to them through Facebook Insights and Twitter Analytics. Culturally, we’ve tried to do this so many times in the past, but being given the data and seeing the immediate response in terms of the engagement through their stories – I’ve never experienced that in the newsroom. And the enthusiasm and excitement has continued. We talk a lot about reporters, columnists, being their own brand, being their own masters, and being responsible for the story well beyond publication time. This way they were really shown how effective it can be and it was tremendously, tremendously encouraging. It’s actually brought that discussion of, “Okay, I’ve got this really big story I’m publishing in November – oh, wait a minute: what’s my social strategy around this? Who am I going to direct it to? And how can I then think about this as I’m leading up to the story and well past the story?”
GA: Are there also changes in the nature of the technology? I’m wondering if the technology is going to change in ways that allow much greater depth for this information. Are you going to be able to learn more about more people in some substantially new ways? Are you going to be able to analyse and use this data in substantially new ways?
KY: The scary thing is that the data actually is out there. Omniture, for example, measures so much. I’ll give you one example. I spend on average about an hour, an hour and a half, delving into the metrics every day. And I’m solving business questions, perhaps from my staff, or I’m looking ahead, a year down the road, two years down the road. And the two business questions I ask myself are, A, “What is happening to my lifestyle audience?” That is, women over thirty-five, maybe first-time parents. And B, “What are the patterns I need to understand?” And so I did a day parting analysis: How is traffic on two channels happening literally by the hour, by the day? And I came to some interesting conclusions, which is that mobile is huge. I was sitting in a boardroom at a major newsroom organization in 2009, where the discussion and then the fear was that mobile would surpass desktop. That has been the rallying cry in digital space over the last five years. We all think this device here, the smartphone, is going to overtake desktop. And then with the tablet it’s been accelerated. Right now in some channels the mobile audience is actually catching up in a huge way. You have as many people reading stories on their phone as you have on desktop. That makes for me a lot of strategic questions: Are we monetizing, from a business perspective, that mobile audience? Obviously, I have to think about that.
MN: But less monetized than desktop.
KY: Yeah. So, we think about that, but also, what are the forms of content I’m providing that audience? And also when? And the second thing I want to talk about is video. Video is the future; it is monetizable in a big way, people are using it. And when I talk about video, let’s talk about audience. Let’s talk about the entire space, because I think a lot of the journalism field thinks that when we think about who we’re battling, it’s newspaper against newspaper, BuzzFeed versus Upworthy versus New York Times versus Huff Post. I am battling against Netflix. I am battling against people who are choosing to lose cable, I’m battling for Apple iTunes, I’m battling for people who are on YouTube. So when I think about audience, I think about where are they doing it? And where are they shifting towards? Because all they’re using is time, and they’re consuming information. When I look at video, I’m understanding patterns: What happens in the evening? That helps me go back to my video team and say, “We have a huge challenge. We have a really great infrastructure, we have great programming techniques, we know how to get video out there to our audience, but here are some things you did not know about our audience.” I’m not going to share much of that because I’m sure it’s proprietary, but it came to some interesting conclusions about how we operate in our day-to-day, our editorial. A lot of it made common sense. You know, are you watching video at your work at nine a.m.? Probably not. Are you watching video at other times of the day? Probably a lot. So what does that mean for us in terms of creating video and getting it out to you at the right time?
MN: And the right kind of video.
KY: And the right kind of video. The right type of video – how long is that video? It helps me solve immediate worries – “Can I own more of the video space?” But when I get challenges like, let’s take more of that YouTube pie, I have actionable things that I can do as a journalist as opposed to waiting for the business office, who doesn’t understand the audience. You know, I understand my audience better than anyone in my organization. So I would love to actually make the strategy to go and get that audience. Rather than be told, “Well, we heard that mobile’s great, so do more mobile stuff.” Great, that’s a great challenge [laughs]. Let me go and try to create some strategies to actually go about doing these things.
ET: That’s very interesting that you mention the business aspect of it. Being in your position and being in editorial and being involved in the analytics, it informs your decisions around editorial, but then it can also inform the business side in terms of being able to monetize what it is that you’re providing, or sell around it. Or have a sponsorship because we’re able to provide data on why this is important, who the audience is, what we expect the reach could be, and that it was conceived of at least partly with those aspects in mind. So it makes for much better cohesion between the business side and the editorial side. And in terms of business, I find at the Star that they’re very much involved in the analytics end, but it’s still surprising when we come to the table and say, “Hey, did you know X, Y or Z?” And it’s just like the light bulb goes on, and it informs a lot of the decisions that we make. … And we’re looking at new platforms every day. We’re entering into a new partnership in the next couple of months on a platform that can be a little more well rounded. Whereas we would go to Omniture for archival data, and certain other programs for real time, it has a little bit of both combined, which we feel will serve our editors and reporters in terms of making informed decisions. It also offers A/B testing. Things like that really drive display and presentation.
GA: What is A/B testing?
ET: It’s where you can segment your audience. You’ve got a story – going back to politics, it’s not a niche audience, but it takes a lot to get that information to the right people. There are important stories that we care about. They can be the dullest headlines in the entire world, but there’s important policy in there, important changes, important journalism. And in order to best present that to our audience, we can test which headlines resonate with the reader in order to drive more traffic, and also drive the audience to an important piece of journalism that deserves to be read. It informs a lot of our decision making. Because sometimes, we don’t know it all. We’re coming at it from the perspective of, “I’ve been an editor for ten years, I think I know this is how people want to read this.” But this allows us to get some immediate feedback on that.
MN: It’s an automatic thing, half the audience gets one headline, half gets the other, and then …
ET: … and it says winner!
GA: Kenny and Erika were just mentioning the use of analytics for business development, and I have two questions that critics of analytics bring up and I wanted to hear how you respond to them. First of all, people talk about problems of analytics in terms of protection of privacy. How do you handle that? And the maintenance of the distinction between news and what you might call promotion – as Robert Piccard was saying, perhaps the fanciful notion that as journalists we should be able not to be involved in the business side, that this somehow pollutes or corrupts what we do. And the notion that sometimes it just amounts to pandering: if your audience says we want Justin Bieber, that’s what they get, Justin Bieber. I’d be interested to know how you and your organizations respond to these kinds of concerns?
MN: I get this question all the time. I’m sure we all get this question a lot. Is it challenging traditional journalistic judgment? I heard this at newspapers. They’d bring in a consultant with a whole bunch of research and everybody would say, ”Well, it’s journalistic.” People would argue the journalistic judgment. I would say that “journalistic judgment” is a sample size of one. Analytics can give you a much larger sample size, and research with a larger sample size is going to be better than the biases, the experiences, and everything else that’s made up in your brain. I was thinking about this a lot this month when the news came out that there were 1,100 missing and murdered Aboriginal women. How many local newsrooms, how many local news editors, using their journalistic judgment, didn’t cover those murders and those stories that were so important? Robert Picard started his talk talking about how there are often these huge crises that we don’t notice until they’re a huge crisis. So I would challenge the concept of journalistic judgment being this sacrosanct thing that’s infallible, because in fact it really is fallible in a pretty fundamental way, and we’ve seen it here in Canada. I think that analytics are perhaps an affront to the traditional sense that we know best and the audience doesn’t, and I would just fundamentally disagree with that. I think the audience does want best. Kenny mentioned this idea of how we are the arbiters of brand. So yes, people want Justin Bieber or Rob Ford stories, but how do we approach those stories? And, you know, porn in the Internet is huge too but we don’t cover porn. So I think analytics is helping us understand our audience in a fundamental way that actually makes the journalism better, so that we’re not just using these biases and individual experiences to make important judgments around stories.
KY: I concur with all that, even what I said at the outset about the arrogant, smart editor-in-chief …
MN: Patronizing, it’s patronizing.
KY: It’s really, “I know what’s good for you and are you going to take it?” And you know what? That’s fine, I guess, for some places but it’s not how a lot of us operate. The brands we all work for, all the news organizations, may have hundreds of years of history – or in my case, my news organization is nine years old. The DNA of the Huffington Post is very different from the DNA of the Toronto Star. We share a lot of the same values, there are certain topics that we excel at, and so we take that as our proposition to the readership. People know that politics is part of what we cover, so we don’t waver from covering that. We also know that we’re not going to be the Report on Business for business. We cannot build our brand to be vanilla, we need to have that flavour. When people really like us like from a social perspective, they like something about what we do. When you come to our packages, our brands, there’s something that you expect, so we are very, very consistent in that. And using the analytics to help evolve. For example, I had to bring the Huffington Post to Canada, and it was a US site, and on Day One all I said was, “Look, this is a different country.” I have the DNA, I understand how the journalism is done, but I need to figure out what it is for my audience …
MN: I mentioned earlier the broccoli and cheese factor. Parents for years have put cheese on broccoli to get kids to eat broccoli. News consumers come, they have all sorts of – they consume recipes, they consume Justin Bieber, they consume very serious political news – and I think there’s lots of research to show that Canadians consume more international news and political news than a lot of other countries. So sometimes it’s a matter of making sure your platter has broccoli and cheese, not just broccoli.
ET: It goes back to what Marissa was speaking about earlier on, that gathering information about our readers is not something new. In particular with the Star, we’ve done reader surveys going back decades. That’s how we used to design our content in print, and we still do. Now, in the digital age, we have access to immediate analytics that tell us about our readership and we’re using those to serve that audience. It’s not quite the same audience, it’s actually quite different. There’s very little overlap, but in terms of what they want from the Toronto Star brand, they want quality journalism, policy pieces, they are very much coming to us for what we do best. Therefore, the idea of watering anything down just doesn’t apply.
GA: On the privacy question – I love news, I love news sites, and in a way I’d like them to know what I’m interested in. But I have concerns about disclosure of my personal information to corporations. I know attitudes about that kind of privacy are changing a lot, but can you talk about the privacy issue? What are your practices about that? Is it a problem? Is it not a problem? Or is it the world changing so fast?
MN: It’s a concern, and the [federal] Privacy Commissioner is really watching everything we do. The biggest privacy issue is not around things like knowing what browser or operating system you have, because that can’t be tied back to you. It tends to be around member centres where people have to log in to comment or do whatever. The technology that drives member centres, making sure that’s ironclad and that the servers are protected, is the most important aspect of it. That’s also why in our member centre, we require very little about you, and I think you’ll see probably as we go forward that people aren’t going to sign up for something where you have to give them absolutely all of your information. For us, it’s also thinking about which third-party platforms we’re going to use. Do we want to use Facebook as our commenting platform? At the moment, the answer is no because we think our audience will react to using it. You can actually use the Facebook platform to drive comments, and I’ve talked to a lot of publishers who are using that Facebook plugin, and they tend to get a very negative reaction from the audience because of these privacy concerns. So I think it’ll drive product choices in terms of what kinds of technology we’ll use. And then there’s the back-end concern around protecting the servers and that sort of thing.
KY: The biggest elephants, Facebook, Google, Apple, and large-scale publishers – not even publishers, just technology companies – you’ve already given your information to them. AOL as an organization does not gather as much data as [they do] – It’s pretty much, open the Internet and you’ve already given [personal information] … Navigating in different countries with AOL, every country has different specific needs … And yes, we – because we operate in all these platforms – are drawn into that conversation. For us, our customer is a consumer, so we’re always listening to their concerns. So when we do things like change our commenting system that will require Facebook moderation, we get a lot of feedback from readers, but it’s helpful because it gets a lot of understanding of where they are. Where are the troubling aspects for them? How can we respond to that? But what it actually means to us as a publisher, whether engagement drops for comments as a result of people not trusting it as much – those are decisions that we have to be aware of as we go forward.
ET: That would be very similar to the way we operate in the sense that we’ve got a membership sign-in, My Star, which requires very little information. The information that is gathered through that is more by your reader habits, not necessarily about you: I’m thirty-five, I’m a woman, I’m all of those things. But in terms of commenting, definitely, we have gone through three different platforms in the last little while and there are long, drawn-out discussions as to how this will affect the potential commenter: What will be the reaction? Getting into the crowdsourcing area, we try to do a lot of crowdsourcing, especially for really big projects that we’re committed to where we want a lot of reader engagement. And we ask a lot of pointed questions in our forums, and certainly, those are run through legal. The reason why we want to try to get more data on the individual is so that we can best understand where that information is coming from. So say, ideas on transportation or education, eventually looking down to [specific] clusters in the city as to where these concerns are coming from. So, we’re asking for more of your information, we’re invading that privacy slightly. It is your decision, but also we find that journalistically, that is something that would be of use to the public at large in the end.
GA: One term that comes up very often in your discussions is engagement. Is engagement going to be the thing that saves journalism?
KY: Engagement for me is the difference between ignoring the audience and [paying attention to it]. I think traditional journalism – when I exited that field, there was very little of that. The letters to the editor page, where you would publish fifteen smart letters and ignore the rest, was pretty much your community centre. I think the way people expect to interact with each other in news sources has changed. A lot of competitors areunderstanding that the conversation is back and forth … It means that you’re listening more. Doesn’t mean you’re always talking to everyone who’s talking back at you … and you also have opportunities to grab people. The new subscriber of the next era is not a person who’s going to want a newspaper that lands at her door. She made a transaction because they like your brand, they like what you’re doing, they had an interaction with you once that leaves a lasting impression. I can’t tell you how many times we’ve had an irate reader who would contact one of our reporters and by the end of the conversation – actual conversation – they’re converted and they’re apologizing for being a troll.So talk to them, and that’s on a micro level, but when you expand that to your overall strategy it pays huge dividends.
MN: Five years ago, or maybe it was more, when we talked about engagement we talked about page views per unique visitor. How can we get one more click? Engagement, absolutely, is a focus, but it isn’t about, well, can we get this one person to click one more time or stay one more minute? It’s really not about that. It’s about a relationship with our audience and a relationship with our brand, and I think for us, that relationship will stand us in good stead. So I would say for the big brands, that relationship with your audience, replying to people, saying thank you, always engaging with them, I think that has to be the focus.
GA: For people who teach journalism – what do our graduates need to know to operate in the world you’ve been describing?
KY: Don’t be afraid of numbers, first of all. The audience is the key portion of your job. It matters more than it ever did. All of us enter into this field because we’re storytellers, we want to tell stories in more different formats, we want to be behind the scenes, we want to be in front of the camera, we want to write a great long read. At the end of the day you understand the reason you’re doing it [is] for a public – you want people to read it, you want people to interact with it, you want to make a difference with that story …
MN: They need to understand the basic skills, but then they also need the critical thought for – okay, so if I see that people are using Facebook when they come home from the bar at night, what is the kind of content and how am I going to present it to actually engage with those people who are coming home from the bar late at night?
ET: Looking back five years, we wouldn’t have said, “Oh, you know, you can make a career as a social media editor.” Whereas now, you can make careers out of being an audience development editor. We’ve hired a content optimization manager. These are new and emerging jobs that are built on a journalism foundation, but also on a data analytics foundation, and it’s quite difficult to find the appropriate people to fill those positions.
Question (not picked up by the microphone, from the publisher of a website that deals with issues of interest to immigrants and recent residents of Canada): He wants to focus on positive aspects of the immigrant experience but finds that his audience responds more strongly to negative stories about difficulties immigrants face in Canada, such as racism. Should he change the focus of his coverage?
MN: It’s like that broccoli and cheese scenario. We want everybody to eat the broccoli, so what’s the right cheese to get them to eat the broccoli? In your scenario I would focus on the positive stories, because that’s your brand, that’s the core of your service, just as CBC news is not going to do recipes, for instance. So if doing positive stories is core to your brand, and it sounds like it is, I would use the analytics to help you package that so that you get the eyeballs. So when you post it on Facebook – my argument would be that Facebook tends to be a little bit more positive than Twitter, where it’s a little bit more snarky. So maybe you want to try and think about Facebook and then watch your referral rate, using your analytics from Facebook. If you’ve got one story and you’re going to post it three, four, or five times on Facebook in different ways throughout the day, then watch which one actually engages the audience and you can start to learn – a certain post at a certain time of day will get traction on Facebook, and if you post it exactly the same way at six a.m., it gets no traction. I wouldn’t let analytics change what your brand is. It’s a tool to help engage the audience.
KY: How old is the site?
Questioner: Two years.
KY: So you’re using that data to understand, what is your base traffic? And I think you need to understand, do you have any loyal readers, period? I think scale becomes a really big issue for a lot of start-ups because I think they simply don’t have a scale, which means that your homepage, just like my homepage – no one visits homepages any more, so you need to exist in other areas. So, you understand your traffic and understand that, potentially, search and social are your biggest drivers, and figure out what is working within those realms, and then capitalize on those wins. At the very end of the day it may be certain search terms that you’re seeing popping up, that’s where all the negative stuff is coming from. So you make a decision: Do I want to do more content that is using those key words? But I could flip it: make an emotional headline that asks, “Why are we being so X?” The A/B test is not to be dismissed because the stuff that you want to succeed, I think you have to help it succeed and measure it. If you have a piece that you think is just the greatest piece ever, and you put it out there and then you’re just hoping it does well, and you’re not supporting it, either supporting it to get it out to other influencers on social to see if they retweet it, get it on a certain platform, maybe – Pinterest is typically more inspirational. So is there something you could do in there? You have to throw a lot of stuff against the wall. We do it every day, even on a large scale, to see what sticks and realize, from those, things that work and help you shape a strategy going forward …
MN: We launched CBC Aboriginal this past year, and one of the interesting things was that the First Nations community of Canada is huge into social media. So we could use analytics to help us go and be part of that conversation. I would also use analytics to see where your community is, so you can go and play – with values that are still core to your brand, but be part of the conversation where they are.
Question: How much do you share your analytics with individual journalists as opposed to using the data on an organizational level? For example, the Globe and Mail had a board where you could see what the top ten stories were at any given moment, and it would drive some of us crazy as we watched our stories go up and down. So when and what aspects of the analytics do you share with your journalists?
KY: I think it depends on the type of organization. I agree, I think some education about why are you putting the numbers up, what reason is there for that other than being a scoreboard? In my organization there’s a big culture where we do understand what the numbers mean. We get a lot of information, we look at hourly reports, we look at what search traffic looks like; every channel editor gets a report every day about how they’re faring. But they get a lot of support, which is, “What does this mean for you? How does this help you in your day-to-day, other than looking at numbers?” I think a lot of us get a lot of dashboards in the morning and we’re reading through dozens and dozens of these panels and we’re – I don’t know what to look at, I guess yesterday was a big day, I guess that sort of went well. If you don’t have follow-up with guidance, well, a newsroom like yours needs to have conversations from the section handlers down to reporters about, “What does this actually mean and why we are doing stuff like that?”
Question: Kenny was saying that video is the future. Can you describe the characteristics of a really fantastic story that worked well on mobile, in terms of video?
KY: I think short, easily shareable – when I say short I mean short. What we’re really grappling with, with mobile, is that there’s a certain cut-off of time that people will no longer engage with it or even launch it. But from a platform perspective it has to be able to play. The responsive sites need to be built to support that. Sorry, I know I’m answering you with a product question, which is that we have to build a better platform for people to see it, but it’s true. And the second part is, how do you get it out to audience? Is it through social channels? Is it prevalent throughout your site, so you give users an easy way to launch videos? So those are a lot of things that we have to solve. Analytics help us understand – for example, a video analytic I look at is video views per visitor. And if you see it varies from channel to channel, then I ask questions like, “Are we using the right strategy for this channel to get video in front of people?”
Question: When you say short – thirty seconds?
MN: For sure, under three minutes. After three minutes, 50 per cent of the audience drops off. And I would also say that there’s a spectrum of video content, and we’re trying to figure out what is that right spectrum. So you have the raw, from the scene, live, happening now – for sure, a big appetite for that. We do get consumption of linear storytelling, and that’s The National. Every morning people watch The National because they missed it the night before. So, between those two goalposts there’s a whole spectrum of video content. The interesting thing about video content is that just as display advertising in digital has been commoditized and it’s a race to the bottom, we’re starting to see that happen with video. So I might argue with my friend here about it being the future because we’re seeing the commoditization of the money that comes from videos, much the same way as we saw with display advertising. But, having said that, the interest and consumption is growing exponentially because the data plans are sucking a little bit less, so the consumption and the desire for video content is certainly growing.
Schlesinger, Philip. 2014. “Multiplatform Enterprises and Journalistic Identities.” Paper presented to the annual meeting of the Association for Education in Journalism and Mass Communication, Montreal, August.
Tandoc, Edson C., Jr. 2014. “Journalism Is Twerking? How Web Analytics Is Changing the Process of Gatekeeping.” New Media and Society 16, no. 4: 559–75. doi:10.1177/1461444814530541
Vu, Hong Tien. 2013. “The Online Audience as Gatekeeper: The Influence of Reader Metrics on News Editorial Selection.” Journalism 15, no. 8: 1094-1110. doi:10.1177/1464884913504259
 Marissa Nelson was appointed senior managing director of CBC’s Ontario Region in October 2015.