How much is the media in the sway of algorithms?
I think they’re very, increasingly under the sway of the algorithms. I think every media company is trying to figure out ways to defend against the algorithms and also play along with the algorithms at the same time. I think you have to do both. You have to build other channels of traffic, and you also have to develop distributed content strategy, so that you’re not just dependent on what’s going on on your websites. But that there’s this sort of duality between playing along with what works in the algorithm and then also making sure that you’re doing other things to guard against the vicissitudes of the algorithm. So that you’re protected from month to month when something changes.
Defend against in the sense of try to have less of your traffic come from, say, Facebook?
Yes. So, I think a lot of places are going back to some of the quote-unquote older methods of getting people to your website, like partnerships with other websites, driving traffic back and forth, and building a strong email newsletter base so that it’s a channel that you can use to drive traffic. And then I would consider, it’s really the Facebook algorithm that has people much more in their sway than the Google algorithm changes. It’s ironic, because once upon a time people were so afraid of any change in the Google algorithm, and people still are, but those are fewer and farther between in terms of things that have an enormous impact, whereas Facebook seems to be changing their algorithm every few months or so. So search-engine optimization is another way of hedging against the changes in the Facebook algorithm.
But I’ve seen some figures that upwards of half of all traffic now, especially on mobile, is coming from Facebook.
Yes. And I think that for newer publishers, and for publishers who don’t have a lot of page traffic — so the publishers who are not the New York Magazines of the world or the New York Timeses of the world — you’re seeing significantly more than half often come from Facebook. Or from social media. So a lot of folks are trying to find ways to get around that. But at the same time, I don’t know any publisher who, any really growth-oriented publisher, who isn’t to some extent publishing, programming around the Facebook algorithm, and around what’s working and what’s not working. It doesn’t necessarily mean they change the topics all the time; that being said, there is a very clear, a very focused attention placed on knowing what’s trending in the algorithm and what works and what doesn’t. But it’s also using the formats. So right now the biggest thing is really video, and live video specifically, in the Facebook algorithm. So you’re seeing publishers who in prior years wouldn’t even think about doing video because they don’t have the money yet to do it, or to do it quote-unquote right, everyone feels like video is a price of entry right now on the internet, and if you don’t have some video in your stream, you’re really not getting the benefits of the algorithm. And I’ve worked with a lot of brands, just doing a few videos that have worked really well and gotten a lot of views have really helped grow their page dramatically.
But then, that could change? You’ve made this investment in video, then all of a sudden it doesn’t matter anymore? It seems like that has happened a lot in the past.
Absolutely. That’s the nerve-racking element of it. Again, you’re seeing some brands now build enormous video departments, and they’re making a bet that videos going to continue to be strong, which I do think that they will be. I think that video is an entire format and that Facebook is not going to pivot away from video, but when Facebook wants to emphasize something else, in terms of its own growth, it won’t be as big of a powerhouse to do video. So what I’m seeing is smaller publishers, or publishers who really are aware of that risk, are hiring a couple of social-focused video folks who are doing video on the cheap, who are doing live Facebook videos, who are sort of doing these slide-stream videos, which are like PowerPoints spliced together, having some gifs and imagery and all that. They’re doing video more on the cheap, not producing very in-depth, high-quality, highly produced videos. And to have one or two video people on your team does make a lot of sense. But most of the younger publishers, those that can’t afford to have huge video teams, are being smarter about it and saying, “Let’s hire one or two video people. Video is going to be something we’re going to be into, but let’s not blow it out just yet, until there’s really a revenue model there.”
And how do you know that the algorithm has changed — because you see the effect on your numbers, or because they announce that they’re doing it?
Once in a while, they’ll announce that they’re doing it. Sometimes they do not. Sometimes you just notice it with an effect on your numbers. And there’s a lot of Facebook groups of media people, and a lot of the social-media people talk to one another. And so when you see something big happen, everyone starts talking about it and hypothesizing that there must have been a change in the algorithm, if they’re seeing six or seven publishers know that something happened. And often Facebook will acknowledge it after the fact, that it’s made a tweak to the algorithm. So it can really surprise you.
Can you give me an example of a company that you’ve worked with that has gotten a surprise like this and figured out how to deal with it?
That’s a good question. Really, most of the companies I’ve worked with saw the video thing happen. But then Upworthy was a big example of that. That Facebook started pivoting away from the more clickbait-y headlines and Upworthy’s headlines then eventually really started changing, to be less clickbait-y. They’ve said it many times that they created this monster on the internet with the curiosity-gap headlines, and then Facebook started really putting weight on what happens when you get to the website, time spent on the website, once you click through. And so Upworthy pivoted and their headlines were a little less salacious.
Upworthy or Facebook said they created a monster?
I think Peter Koechley said it; he even apologized for the monster they unleashed on the internet … everybody started doing clickbait-y headlines. There’s a lot of places that continue to do that. So that’s probably the best example of a site really pivoting. But you’ll see that at times, just posting images and memes will do really well in the feed, and then people will start doing more images and memes, or shortform video will do really well, and people will start doing that. You really see people — I would say that most of the publishers that I work with, every two or three months you’re seeing some sort of pivot to publish the things that Facebook is giving preference to.
To get back to this monster quote, sure they created the monster, but they wouldn’t have created the monster if the environment wasn’t fertile for it. So in a way it was Facebook that created the monster by rewarding the monster once it was created, no?
Yes, I think that’s right. I think that’s absolutely right.
So I wonder if the people that you’re working with find themselves realizing, “Oh, okay, we can’t do stories about ducks anymore, we have to do stories about chickens, because that’s what’s going to get the numbers now,” and it may be that they wanted to do stories about ducks, but now they have to do stories about chickens, so it’s almost a willy-nilly reactive stance, if that makes any sense?
Yes. That definitely happens. What you’re seeing a lot of publishers do is, they’re heavying up on stories on topics that do well, and they’re lightening up on topics that do not do well. If it’s a more sophisticated publisher, they’re understanding that this is a topic that may not do well on Facebook but may do well elsewhere, and try to find an audience for it elsewhere, and may not just cut out those stories, but it definitely affects the stories being done, there’s no question about that. At least it affects the volume of stories being done. So, most publishers have editorial integrity and are still going to be doing the stories that they consider to be important, but they’ll just do more of the stories that are important and get a larger audience and fewer of the stories that will get a smaller audience. So there are some organizations, some sites, which I won’t call out, that will just do what works in the algorithm. Most of the organizations I’ve worked with are more highly journalistic or have some journalistic ethos. And so they’re also going to do stories that may not do really well with the algorithm … but they won’t be sharing those stories to Facebook. So if they know a story isn’t going to work on Facebook, they just won’t put it on Facebook. They will have a recirculation unit on their website, a next-story unit or something else like that, so that once someone gets to the page, that’s how they get them there. Or they’ll put that story in their email newsletter. Or they will optimize that story for search. Or they’ll try to find other ways to get people to that story. But any really smart, sophisticated publisher is only sharing stories to their Facebook pages that they think are going to work. So you see, that’s why you can see for some publisher’s Facebook pages, even if the publisher covers a very broad range of topics, the Facebook page might be much more focused.
There’s a sense that the algorithm changes lead to the creation of content that the creators themselves have no real interest in — for instance, I think the Times is doing these live videos that are just, like, a couple of journalists sitting chatting, and there’s no real reason to do it other than the algorithm.
That’s one way to look at it, for sure. They probably wouldn’t have done it if it weren’t for the algorithm. At the same time, it could become, they’re rolling out new formats, and certain brands are experimenting with them probably more vigorously and earlier than they might have otherwise, and then they will hopefully get to a place where they hit on something that really works, not just for the algorithm but also for the audience. So I think it’s a situation, at the Times especially, because I know they’re going very heavy on it, where they’re not going to produce poor-quality, really boring video for very long, if you think that they’re doing it now, but they’ll be forced to evolve into finding a format that works with the brand as well.
Have any of the companies that you’ve worked with been particularly whiplashed by, or taken by surprise by, Facebook algorithm change? Can you give specific examples of how the focus has shifted for one of the publishers?
There’s no question that that’s happened a lot. There are very few places I’ve been that haven’t made some sort of a shift. At Fusion, the shift was to both native and live video.
If you could magically figure out exactly how the media world works right now, it won’t be valid in six months. The game is going to be different.
Yes, I think that that’s right. But I think that, again, you have some publishers who are just chasing everything that happens. And that, in my opinion, is unwise, because, as you said, it changes every few months. So the smart publishers that I know are really looking at a diversity of sources of traffic, a diversity of ways to get people to their website, and so are a little bit less reactive to the vicissitudes of the Facebook algorithm. So the really smart places like, I’d say Bustle is one of them, sort of came out of nowhere a couple of years ago and now is at about 50 million uniques. I don’t know exactly their traffic breakdown, but I think at least 40 percent of it is from search. And you think of a site that came out of nowhere, like an Upworthy or Distractify, places like that, it’s usually Facebook. But I know that they’ve been really, really smart about making sure that they have a lot of search traffic. And the places I’ve worked with that are really smart — and I’ve actually really seen — when I was at Upworthy years ago, they were only focused on Facebook. Search-engine optimization was something, but it was not something that was on the radar at all in the first year or so.
In 2012, when you were at Upworthy, it was just SEO or just Facebook?
Just Facebook. It was like, “Facebook is what’s working, let’s go with Facebook. Let’s go 100 percent in on Facebook.” And SEO was a conversation that just wasn’t entertained. And you’re seeing now that the publishers that have come out since Upworthy went up to 100-something million uniques and then crashed back down to 40 million, or 30 million or whatever, really saw things like that happen and are much smarter about prioritizing search-engine optimization in their traffic mix from the very, very beginning. So Bustle was very focused on SEO, and making sure that everything is optimized properly for that, but then that’s its own beast, because it’s also really looking at the topics that work in search-engine optimization and doing more stories around those as well.
It’s like what goes around comes around, because the HuffPo was famous for SEO stories like “What Time Is the Super Bowl?” Where again, you don’t care about that content.
I think places have gotten more sophisticated than that at this point. So I think there’s really three elements of SEO. One is, how the site is programmed. And all the back-end architecture, and making sure that that’s optimized properly. Then there’s really how you write the headline, all the key words that you’re using on the site. And then the third element is really volume. I think a lot of people are not necessarily paying attention to volume.
Volume of content generated?
Yeah. Volume of content generated on specific topics. So I think a publisher like Bustle, which is really smart, will find topics that people aren’t necessarily covering as much, in their wheelhouse — because you can see competition on search terms in tools like Google Keyword Search. You can see what search terms are popular, you can also see the competition for those terms. Places like Bustle will find areas of content coverage that they would do anyway, or that are really in their wheelhouse, but that aren’t being covered by a lot of people. And they’ll decide to focus on doing a lot of volume around those topics.
You could make the argument that this is for the good, because it’s making sure that the public has access to lots of content on the things they care about, but it also from the perspective of old-school editors, it’s kind of taking things out of their hands. It’s not their judgment so much anymore.
I feel mixed about it. I don’t feel as nihilistic about it, in certain ways, that some people can get. At the same time, you have to be very audience-focused, and by that I mean not just what your consumer wants but what actually works in the distribution channel. Thinking about how is someone going to find this content. Again, for most publishers, I’m not seeing that that becomes the exclusive way that they create content.
There was an interesting piece in Recode recently by the founding CEO of Chartbeat, Tony Haile. He was talking about how in the past publishers would have the stuff they care about — schlocky stuff would fund the good stuff. Now everything’s sliced and diced, so if you want schlock, that’s all you’ll see on your mobile, you won’t see that they did this good stuff.
I think that’s true to some extent, I know Tony well, and I know that’s his theory about where things are going. It’s a possibility, it’s a definite possibility. But I think what publishers are doing to try to mitigate against it are, on mobile it’s much more difficult than on desktop, but you do see these next-story units coming in and merchandising other stories. “Merchandising” sounds like a really harsh word for! Marketing other stories within the body of a story. To get you to view another story. But it’s hard! It’s much harder. Publishers who have a large percentage of mobile traffic just automatically have less page views per visit than those with desktop traffic. So it’s definitely moving in the direction that Tony says it’s moving in. I don’t know if it’s going completely in that direction. But it’s harder to get people to just go somewhere and discover. I think what people are doing is in Facebook feeds, so you’re publishing most of the things to Facebook that will work really well, and once that page is engaged, then a story that’s quote-unquote important but would not necessarily work as well on Facebook as some others, they’ll still post that because it will get more reach than it would have otherwise.
Does this somehow explain the Trump phenomenon?
No question! Just from a personal perspective, I was thinking at the time yesterday that every time I launch the Times app, I log into the Times app every day, and the top stories every single time are some Donald Trump–related story. They must be seeing that engagement is higher with the app because they’re publishing Donald Trump stories. When I was at Fusion recently, we needed to be really conscious of not having too many Donald Trump stories in the Facebook feed. Because they would perform really well, and we wanted to be more balanced.
Okay, you wanted your coverage to be different from what the algorithm was pushing you to be.
Yes. It could become all Trump, all the time. We didn’t want, at Fusion, the brand to be known for the Donald Trump coverage, that’s not what the brand was about. But those stories would perform really well. So knowing that those stories would work really well, we had to say, there still is editorial judgment that comes into managing the Facebook feed. The New York Times is not posting all Donald Trump to their Facebook feed all the time, because that’s on the New York Times’ mission. But you could ask Cynthia Collins, who manages their social media there — I’m sure that there is some leaning toward those stories that perform better.
You don’t want to give your feed over to that, so you do less, but the question becomes, how much less?
Yes. And I’m not on the editorial side, so you probably need to talk to someone in editorial for a better sense of the production. But there’s no question in my mind that the algorithms have driven a lot of the coverage of Trump. I think about it all the time. Anything he does, it’s salacious, and it’s like watching a train wreck, and everybody’s sharing it, and everybody’s producing those stories, and so you get that bubble. There was great story, I forget where it was, but it was doing side-by-side a liberal person’s feed and a conservative person’s feed, and the way the stories were covered, and the types of stories that they see in their Facebook feed, the Hillary stories in the liberal person’s feed vs. the Hillary stories in the conservative person’s feed. It’s just fascinating. It’s partially what Eli Pariser of Upworthy, the phrase that he created, the “filter bubble” … seeing more and more of things you agree with and less and less of things you don’t agree with.
If you’re conservative, you’re going to get a lot of Benghazi, if you’re liberal you’re not going to get any Benghazi, but either way you’re going to get a ton of Trump. The conservative person sees Trump saves a baby, and the liberal person sees Trump runs over a baby — everyone’s fascinating by it.
There’s no question the media is really looking at itself now and appreciating what their role has been in creating his rise to power, or power that he has right now. The amount of attention the media has given him because it’s really — I don’t think it would be a real hyperbole to say that these algorithms have had a strong effect on the election, if you believe that the media influences elections. Because the media has started to heavy up on those stories.
Can you recommend a good subject for a story about a site using Chartbeat to manage a story?
You would probably want a news publisher. I know that the Times does that a lot in terms of their website. They’re using Chartbeat to look at their homepages. I know ABC News Digital uses Chartbeat a lot in, because in Chartbeat you can see how a headline or how a story is performing in relation to how other headlines had performed in that slot previously, so you can see if it’s overperforming or underperforming. Ironically, I had just been doing some consulting for Chartbeat, so I know some of those who actually used it for that. So it’s really the publishers that have larger home pages and larger direct traffic who are using Chartbeat to optimize there. If you wanted to look socially, I’d probably look at a different subset of publishers who are perhaps testing how something is performing on their site, or testing how something might perform in a dark post on social media before they post it to the full audience.
A “dark post” means it just goes out to a small subsection?
Yes. It just depends on what you want to look at.