Google’s News Experiments & The Quest To Solve The “Read State” Issue
Hey news publishers. Stop acting as if your content only appears printed on dead trees and tap into the dynamics that the web offers. That’s a blunt summary of advice from Josh Cohen of Google News, from a wide-ranging interview with him on Google’s experiments with new ways of delivering news. Cohen, business product manager […]
Hey news publishers. Stop acting as if your content only appears printed on dead trees and tap into the dynamics that the web offers. That’s a blunt summary of advice from Josh Cohen of Google News, from a wide-ranging interview with him on Google’s experiments with new ways of delivering news.
Cohen, business product manager of Google News, says Google has no ultimate solution for the future of news online. It does have a vision of a super personalized news product that tracks someone’s “read state” and keeps them constantly informed with updates. But to turn that vision into reality, it’s conducting a variety of experiments. Some will succeed; some won’t.
The experiments aren’t meant to compete with publishers. Cohen stressed that Google’s not a content play and has no “Hulu for journalism” pretensions. Google’s a technology company, he says, one with tech that it hopes news publishers can tap into.
For publishers to be successful in a personalized news product, they may have to consider the “Living URL” model of stories, Cohen said. Think Wikipedia, written by journalists. And think about how newspapers might learn from a classic Christmas film, Miracle On 34th Street, where sending your customers away might actually make them more loyal.
Below, you’ll find Cohen’s comments on these and other issues. It’s the second part of a three part series, from an interview I conducted last month. Also be sure to read the other two installments, Josh Cohen Of Google News On Paywalls, Partnerships & Working With Publishers and Under The Hood: Google News & Ranking Stories.
What’s Up With Hyper Personalized News?
Both Google CEO Eric Schmidt and Google vice president of search product and user experience Marissa Mayer have talked about the concept of a hyper-personalized news system that remembers what you know, finds what you want to learn about and would be as easy to use as flipping through a newspaper or magazine. It’s been raised as an idea by Google as far back as 2007. Is this real? Is it coming? Is Google working with publishers on it? Cohen wouldn’t say much:
The short really unhelpful answer is, sort of. I think if you look at a few of the things we’ve launched in the last month or so, and launched being a broad term for publicly available, you get a sense of how we’re approaching some of this.
Journalism Should Be Written For More Than Print
Continuing, Cohen spoke more broadly of the idea that journalism still largely acts as if it is meant for publication in a hard-copy newspaper instead of also appearing in the instant access, hyperlinked world of the web:
News online by and large hasn’t adapted to the medium and still is largely brochureware, where people are taking not only just the physical article that was in the paper and just putting it online but also just the way you tell stories.
It [storytelling] just hasn’t really adapted and taken full advantage of the medium. Beyond just the challenges that go with it, it’s a completely different way of telling a narrative. And so there hasn’t been that real transformation. As a result, you get a lot lower engagement online.
Fast Flip: To Combine Best Of Online & Offline
Cohen said that Google Fast Flip, launched in September, is one example of how Google is experimenting to combine the best of both worlds, offline ease of reading and browsing with online’s “smarts” that allow for personalization:
The premise behind Fast Flip is really a step in that direction of trying to figure out how do I create a good online reading experience that is engaging, that captures some of that browse experience?
Fast Flip is focused right now on a specific set of content, more of the longer form content that’s more suited to a magazine, which is why you can sort of get that idea of a custom magazine that you can sit back and browse, flip from one page to the next, and without having to take 5 to 10 seconds for each of those pages to load.
You literally have that offline browsing experience with all the advantages of being online, customization, aggregation, and all those types of features.
Fast Flip: A Test, Not Google’s Ultimate Solution
However, those who interpret Fast Flip as “Google’s plan to save newspapers,” Cohen said, are making a mistake. It’s not perfect. It’s a test, and a test that might not work for everyone or at all.
What it is, is Google’s attempt to try and experiment in one of those categories. I think Fast Flip has a lot of advantages for, and this is my personal view, a certain set of content. How does that work for all forms of news, for all forms of content, I think remains to be seen.
We want to put that [Fast Flip] out there in Labs [Google Labs, where Google releases experimental products]. We’re excited about it. I think the response has been really positive both from users and from publishers. But we want to test it.
I think we all see it as one step in that direction of what does a news experience look like? Could that be a Google-hosted experience? Can that take place on a publisher side of it? Sure, I mean yes, the answer is ‘yes’ to both those. But that’s, more than anything, it’s an idea of ‘Look, we’re experimenting with this.’
Continuing, Cohen explained that Google sees certain issues with how news is delivered online, has thoughts on how things might be improved, and that Fast Flip is meant to test some of those theories:
That’s kind of how we approach things. We go out there and test it. And we have certain assumptions that are behind Fast Flip, which the data will prove us right or wrong, and then we’ll go from there and either continue to expand it and see what’s working and what’s not and iterate on it, potentially blow it out, deeper integration into [Google] news, separate products. Who knows? We try and be as open as possible to letting the product and the users’ reaction to it drive that direction more anything else.
In fact, Google has just begun a small test of Fast Flip being integrated directly into Google News.
Google To Publishers: Can Our Technology Help You?
How about that laundry list of ideas that Google provided to the Newspaper Association of America, which approached companies it selected for ideas on how to monetize news content? Cohen said this wasn’t a specific plan but rather a list of off-the-shelf technologies and systems that Google already had which might be adapted:
The way that we intended it was a response to say, ‘Look, this is our thinking on the space.’
Continuing, Cohen explained:
“We really tried to map out how we saw that ecosystem working [such as news publishers trying to process micropayments for content], how some of the existing Google technologies could potentially plug in to that, and we could power a site in the same way that Google Maps can power your mapping solution. We don’t create content but we create technology, so how can you use that?”
After the document leaked out, Cohen said it kicked off a number of discussions with publishers wanting to know more. Focusing on the micropayment aspect, he commented that subscriptions can and do work, but an easy way to charge is vital
For a certain set of content, and for a certain type of publisher, subscriptions, not only can they work, they do work, for a subset today. Is it a panacea for all forms of content, simply to put in a paywall? Personally, I don’t believe so. But again, if there’s a set of content that it works for, you’ve got to have a technology solution that works.
If you’ve gotten over the hurdle of someone saying ‘Yeah, I’d be willing to pay for this, I’m going to take out my wallet,’ don’t have them go through 50 leaps to get there. Then it’s really doomed to fail.
Google News & Personalization
Cohen then circled back to my original question, about how all these news experiments relate back to a more personalized news experience that Google espouses.
“We always talk about delivering the right results,” he said. “So yes, Google’s working on this. What forms that might come out in, could it just be a better personalization of Google News or certain types of feeds you can subscribe to or custom sections that are there today? I think it’s kind of any and all of the above.
By the way, since I conducted the interview, Google rolled out Custom News Sections, a way to personalize Google News so that you can browse stories that match particular keywords you’re interested in. A form of this has existed since 2007, but the update rolled out a directory, so that Google News readers could share custom sections with each other, along with the ability to have more complex keyword matching.
Challenge Of Explicit & Implicit Personalization
Of course, Google also has personalization that learns through watching behavior. But both explicit and implicit personalization features have much room for improvement, Cohen said.
I don’t think it’s where we want it to be. Personalization is hard to do.
The easiest of those are explicit personalizations. Even that, I think, is hard to get right, because a user will come and make certain decisions or choices about the type of content that they want to see, and yet if they miss something because of the personalizations [that they did], it’s still your fault.
You know, ‘These are the sources I want to see, these are the topics I want to see,’ and then they miss something, and, ‘Why didn’t you tell me about that?’. And again, that’s with a lot of direct signals and instructions from the user.
Then you continue down that spectrum on the implicit side of it, of understanding my reading pattern, not only what stories and sources and topics that I’m interested in but also my experience reading those specific stories. And so I think Eric’s talked a little bit about this, the ‘read state.’
Read State: What Online Can Detect That Offline Cannot
Read state, Cohen explained as he continued, is a key challenge that Google feels online news faces, the need to figure out where someone has left off in following a particular news story.
Think about it this way. The traditional newspaper reader would get their morning paper, read some stories and be done. The newspaper had no idea what they read. So when writing updates to those stories, the newspaper was forced to assume you knew nothing. It had to get the most important breaking aspects up at the top of a story, writing in “inverted pyramid” style so that if a reader drops off, the less important facts are safely buried further down in the story.
Online papers could be smarter. They could understand what you’ve read, where you left off and keep you informed with only the new material you need, because they’d understand your read state. But online news isn’t written this way. It continues to be produced as if people are reading offline. As Cohen explained:
Every single day I have to put something out in the paper. So there’s an on-going story. Every single day I file another article. A deadline comes in, 6, 7 o’clock or whatever, I file it, and it goes out there because I have to put something out there in the paper.
As a result, often times you have to have a certain set of facts, even if they’re just one little update to that story. It generates a larger story either to fill space or because I can’t just put a quick headline update to it and link back to my other sources to it.
Part of the reason that you see that is, one, there hasn’t been too much innovation in the space. But also because people don’t take into account your read state. So if I know that you’ve come there, I can give you the full story, or I can give you a quick update, a bulletpoint summary. That’s another level of personalization that I think is not there.
Living URLs & Bringing “The Miracle On 34th Street” To news
In theory, the idea sounds great — this “Living URL” idea that Google’s Marissa Mayer has especially suggested, the story that lives in a single place, constantly being updated. But aside from current technical issues Google News has, where it can’t even handle stories like that (see Of Living URLs, Newspaper Rankings & California Fires), there are current business issues that prevent it.
For example, what happens when one paper reports on a story, then a different publication reports a unique and specific fact. You do you merge what the two competitors say together, especially when they often don’t want to acknowledge each other?
That gets to another issue, people not linking out. Publishers trying to be all things to all people, instead of a focus area, of whether it’s a regional area where it’s the LA Times and I covers LA or it’s a topic where I’m the Washington Post and I can cover politics or I’m the Wall Street Journal, and I can cover business…
My value is my editorial filtering. I recognize that if I send you off [my site], and I just put a link to an update, if I’m the New York Times, and the Wall Street Journal has a good update, I’m going to link off to it, so that I know that this is a good source that tells me what I should be reading, even if it’s not on their own site.
That’s what’s going to set me coming back. Not because I’m trapped into their web site, and I have to know that all the information is coming from there. There’s a comfort level, there needs to be a comfort level, to send people out.
That’s certainly a Google bias towards this. Where our focus is largely on getting people off of our site, because we recognize if we provide value in serving them the most relevant search results, whatever it is, news results, that this is going to be a jumping off point for them.
Call it the Miracle On 34th Street approach to news. For those unfamiliar with the classic film, the Macy’s department store Santa Claus (who is actually the real Santa) sends shoppers who can’t find what they want to Gimbles, Macy’s archrival. While at first Macy’s management is horrified, they’re won over as a customer declares her loyalty to Macy’s for putting the customer first.
In my own experience, the mainstream media traditionally has not linked out. It’s more common where they have blogs, such as Bits from the New York Times or Digits from the Wall Street Journal. And among news blog, it’s very common to crosslink (my Blogs & Mainstream Media: We Can & Do Get Along post gets into this more).
Should Newspapers Become Like Wikipedia?
The Living URL / Read State concept also sounds similar to something that already exists: Wikipedia. Should papers all simply become Wikipedia-like, where stories about a particular topic reside on a single page that’s constantly updated? Cohen’s not certain himself and figures there will be lots of experimenting.
I don’t think we have that answer. That’s something else that I think we’ll certainly experiment with on our side and experiment with publishers directly. I think the concept makes a lot of sense. How do you put that theory into practice?
I know you played around with certain parts of it, and you felt certain parts didn’t work felt or not. That was like one take at it. There will probably be 50 different takes at it from a number of different sources….
I think everyone recognizes the value in an online world, of having that persistent URL and a single source to get all the updates on a given story. That doesn’t mean there aren’t changes that need to take place on the editorial workflow, on the product design and also on the search side of it about how you pick those things up, too, which I think you pointed out.
Is A Living URL System In Testing?
But does Google have something like this in testing, that hasn’t been put out there? What about all the rumors that Google has something in the works with the New York Times or the Washington Post? Cohen wouldn’t say but rather pulled back again to stress there’s no perfect grand plan system Google has up-its-sleeves:
I will say this. We are directly working with a number of different publishers. We talk about the monetization, but product and content creation as well, whether it’s a set of tools or a structure or do so.
There are definitely discussions we’ve been having. There was that story that was reported about these collaborations with the New York Times and the Washington Post. It took bits and pieces of a lot of these different experiments and turned into a brand new product that would be the “future of news.”
While it is accurate that we’re working with the Times and the Post – among others – about improving the online news experience and how Google might contribute, as with many things, it often gets turned into something much grander.
Google Unlikely To Build A Hulu For Journalism?
How about the thought of a “Hulu” for newspapers and other journalism outlets, where Google might compile all the stories together for publishers, with single monetization or a subscription basis?
Cohen didn’t see Google as putting out a specific collective solution but rather staying focused on publishing tools that publishers could use individually.
Broadly speaking, the area of creating platforms for content is something we do today, such as with Blogger or YouTube or Knol. These are all examples of Google trying to make it easy to put content online. That’s consistent with what we want to do,” he said.
But Fast Flip seems like the start of a Hulu. Or is it something that Google may license as a tool for anyone to use? Indeed, that might be its future.
“Fast Flip is a longer term vision. We’re not saying it must be hosted at Google,” Cohen said. It might evolve into something publishers can host on their own sites or customize, he said.
Spotlight Added; Newsmaker Comments Go – Learning From Everything
How about the new Spotlight section of Google News. What’s the purpose behind that, and what’s selected for it?
It’s been really successful with us so far. There’s interest from publishers that it has potential to give another platform for content that often doesn’t do that well with Google News, which changes hourly. It’s for long shelf-life content, the enterprise stuff, the investigative pieces.
And thoughts on things that have gone away, such as dropping the attempt to get quotes from newsmakers on stories. Disappointing?
I don’t mean this to sound too Pollyannaish. Google is trying to encourage itself as a place where you can have failures and think more radically about the approach to a product. There’s no sense that I have get this perfect or it will never launch.
The comments feature is something we killed, but we also learned something from it …. People are encouraged to ask ‘What did I learn from this?’ That’s what makes it interesting to work on the products.
Are there things Cohen particularly likes? Cohen said he was happy with many “under the hood” changes that he thinks help people better trust Google as a news aggregator.
I was curious if more people browse stories at Google or do keyword searches. Cohen wouldn’t give specific figures, but he did say that at a typical news site, he knows that searches tend to be in the single digit percentages. In contrast, browse versus searching at Google “is much more balanced,” something he didn’t think was surprising, given that ultimately Google’s a search engine.
Google News: Trying To Serve A Balanced Diet Of News
Finally, I’ll end this part of the interview with what Cohen said was the overall mission for Google News, to “educate and inform.” In particular, Google News is aiming to expose things you want to see alongside things you should see. Or as Cohen put it, “serving the vegetables as wells as the dessert.”