Most folks prefer to consume news tailored to a community (Techmeme, Hacker News, Drudge Report, etc.) over news that is personally tailored to them (News.Me, Prismatic, etc.)

Why is that?

My hunch is it's because we're interested in news as a way to relate to other people –- "Did you see that game last night?!" "I can't believe Instagram did that!" -- not because we have an inherent interest in "staying informed."

There are use cases for both, I think.

Communal news = Common denominator.

I started smoking cigarettes occasionally because I'm pretty darn socially awkward. Cigarettes (in New York) are a common denominator. Knowing what happened in sports is a common denominator.

Maybe "connecting" or feeling a connection with someone else has more net importance/reward for most people than feeling self-satisfied in your personal knowledge.

Another take: community news also has virality built in. Virality or word of month increases demand. Increase in demand = increase in value or in some cases, just perceived value.

The question is more about human curated vs. machine curated. Even if you had an algorithm that was as efficient as a human editor (there isn't one out there), I'd bet that people would still gravitate toward news that is editorialized. Only reason machine curated news is worth exploring is that it's scalable from a business perspective.

I agree with you, Kelsey.

Ross, it has nothing to do with the quality of the content. I've been really impressed with the relevance of the content that Prismatic recommends to me, for example. It's spot on. I always find stories I'm interested in. Yet I find myself going back to Techmeme and Hacker News instead, again and again, and I think it's because I want to know what other people are talking about/reading.

Josh, you're totally right. I wasn't impeaching the quality of the content per se. Was more pointing out that the strength of sites like Drudge, Techmeme and Hacker News doesn't come from them being targeted toward a specific community but rather that there is a much stronger human hand in deciding what content is important.

I like to put this in terms of personality mediated content. We like to think we like the things we like because of personal taste and interests, but we also highly value the validation that comes from what Kelsey calls that common denominator. There is a great article on Pandora in the NY Times about an upset user that didn't like the "inaccurate" recommendations that were being fed to him. Turns out they were algorithmically accurate, but not socially. I think curation hits a sweet spot because it gives the impressions that we consume based on taste but the comfort of content being fed to us by a trustworthy personality. Algorithms fail to emotionally connect us to the content.

"Algorithms fail to emotionally connect us to the content."

+++ great point

It's always struck me as sort of a straw man to pit algorithms against humans re: curation.

At NewsFoo, I ended up in a great conversation about an as-of-yet undiscovered "fourth paradigm" for curation: editorial, algorithmic, social and ________. Seems a bit obvious, but effectively combining all three (a Middle Way of sorts) ended up being the most compelling argument.

The frame that @Josh sets up here is actually a really good one for showing why that's the case. Humans (even a whole crowd of them) can't tell you what kind of device you're using, or whether you happen to be standing on a subway platform or sitting on your couch. Algorithms can't give you @Kelsey's shared experience or @Keenan's emotional connection.

Agree with your second point Josh, that we read the news as a means of participating with groups — our friends, family, village, city, industry, country, and world.

The size of the group dictates the extent to which it looks like "social/community" or "editorial." The larger the group, the more intense the effects of the power law (shirky.com), and the more that group starts to look like "editorial."

What we're arguing about are different types of conversations. Humans don't consume information any more than they publish it — machines publish and machines consume. Humans talk about things with other humans. Everything else is an historical accident.

I think a big part of why people read news is to "be in the know" or make sure you aren't missing out on important information. In cases where you are seeing what is popular or editorially selected, you feel more comfortable that you are seeing the most important information. When it is too tailored to your interests and personally applied, you always wonder if you are missing the important news. "Fear of missing out" is one of the most powerful human emotions to drive behavior around.

I wouldn't be surprised if in a few years, a large proportion of the news stories people read will in fact be auto-curated by Facebook based on their and their friends' activity. They'll just see it when they visit Facebook, so whether they'd prefer something else will be beside the point. (Same for Twitter.) But yeah, I think in the foreseeable future people *deeply* interested in a news topic (like tech, or politics, or finance) won't be relying primarily on algorithmic-curated news, because the choices made by those algorithms just aren't as good as those where humans decide at the last stage. I can't really justify this claim without writing far more than I'm prepared to now, but to summarize: the algorithms are just too dumb.

Machine-generated recommendations will get to a point where they are indistinguishable from social recommendations.

It's a matter of presentation. If content is packaged as a machine-generated result, it lacks additional rich layers of social information — what does interest in this information say about me? What group does this put me in and do I care to identify with that group? Does that group represent ideologies I agree/disagree with? Will this knowledge be socially valuable to me? Etc. That stuff is hidden in every bit of information we consume.

There are utilitarian cases when we want "search result" type information, but typically socially relevant information caries emotional power, and emotion trumps information every time.

Social error—

"Westergren likes to tell a story about a Pandora user who wrote in to complain that he started a station based on the music of Sarah McLachlan, and the service served up a Celine Dion song. “I wrote back and said, ‘Was the music just wrong?’ Because we sometimes have data errors,” he recounts. “He said, ‘Well, no, it was the right sort of thing — but it was Celine Dion.’ I said, ‘Well, was it the set, did it not flow in the set?’ He said, ‘No, it kind of worked — but it’s Celine Dion.’ We had a couple more back-and-forths, and finally his last e-mail to me was: ‘Oh, my God, I like Celine Dion.’ ”

This anecdote almost always gets a laugh. “Pandora,” he pointed out, “doesn’t understand why that’s funny.”

nytimes.com

Gabe, it's interesting that you mention Facebook as being at the center of news in the future. My high school and college friends (i.e. my current graph) don't represent my interests at all. And though I know you can "Subscribe" (Follow?) reporters and publications, it's not what I'm looking for when I head there. Seems like the wrong context to consume news, to me at least.

Here's my thinking: Facebook will be forever optimizing the stuff it presents to users to increase engagement. Presently this is the newsfeed, though it could be that plus other kinds of summaries in the future as well. So, at some point, for the typical person, the typical news story consumed might be via Facebook algorithm. (What % of news stories consumed NOW are via Facebook? 10%? Just an uninformed guess, but whatever it is will grow.) And Josh, a smart enough algorithm will figure out which data to weight, i.e. will determine which friends share stuff of interest to you, and which don't. And again, I'm not asserting all of this is what news readers will actively decide is best for them, it's just what might end up happening.

As usual, Gabe is right. A big chunk of actual news consumption happens on highly personalized platforms (Facebook and Twitter) and this will continue to grow. Twitter is the front page for Politics and Tech. Facebook for general news and big, moving stories. The direct attempts at "personalized news" seem to fail but adding news to a larger social experience works great.

That means publishers don't need to personalize because the social web will do it for them. News sites publish to a core audience who shares to social platforms where the content is personalized via follow graph and/or machine learning. News may start on non-personalized sites read by junkies but that isn't where it is consumed by a growing number of people!

hopefully personalized is not always the opposite of communal and social. Our approach at Nuzzel is both personalized and social, but it's definitely not algorithmic....

Seems like the big theme here is that these "ways" cannot be separated. The "Middle Way," as Libby writes above, is already here, and has been for a few years. When we think of personalization we think of big machines scraping content, categorizing by topic, and redistributing by interest. Those products never quite cut it, not because their recommendations are bad, but because the recommendations are made in absence of social context.

@Jonah I'm curious to know why you think Twitter and Facebook lend themselves to different types of content (politics/tech vs. general). Does it have to do with features, or is it just about who is using each product?

Look at us, ahead of the news cycle! Alexis just wrote something up on Google News re: all these same issues, too.

TL;DR: "The people who are producing the inputs to the system and dependent on the outputs will begin to adjust their behavior. They'll change to make themselves more legible to the machine, and those that do so best will prosper."

Strikes me that neither news consumption system we posited here (communal or personalized) lets you avoid the construct of "people gaming a system" entirely. Hacker News' Achilles heel is groupthink. Google News' is its algorithm.

/* Please keep this branch open :) More to add -- just not enough time right this moment. Good stuff in here....*/

I think an algorithm that bases story weighting on follow graph MAY contribute to determining the best content. But, the best story consumption engine will be much smarter.

Jonah mentions machine learning. If a crawler could index the most commonly used words in an article (with synonyms), map them to broader interest categories, it's possible to map interests to consumers.

Gabe said: a smart enough algorithm will figure out which data to weight, i.e. will determine which friends share stuff of interest to you, and which don't.

I think the "which data to weight" will be based on interest matching not social clicks. When a "sports" link is shared with x word in the caption, display it to y consumer group.

This is an awesome Branch, first and foremost. There's a couple interesting points being raised. On a whole, I think there's a delicate balance that's being missed in the communal vs personal news debate, and that's that editorial, algorithmic, and social means of delivering news need to ultimately serve the consumer. It's 2013. Consumers deserve a better filtered result. The abundance of information and the scarcity of attention requires that we deliver both unexpectedly relevant, and surprisingly familiar content to every single user. The publisher's role is to prioritize information, stretch a consumer's experience or world view, and provide a delightful experience.

I think about this in two ways:

1) The social ordering of content. If you think of three concentric circles, the outer rim would be the most broad: what the web likes ("hits"). Then comes what your friends like. The last, most inner circle would be what you like. Mass appeal to personal appeal.

2) The publisher's brand. Using the above example of circles, a publisher's brand is the most generic, outer ring. Their homepage is their declaration of purpose. The next ring is the like-minded consumers that create the community that embodies the ethos of the publisher. And, finally, the inner circle is about the individual consumer who only cares about specific topics within this portfolio of content.