Creating web content is incredibly easy — but filtering content is really hard. In late 2014, I realized I was reading too much bad content. I felt enraged by some of the articles I clicked on, because they were such a thoughtless waste of my time.
I got so frustrated that I decided to invest serious effort in fixing the problem on my end, instead of fruitlessly swearing at my laptop. I hoped to determine what what was non-optimal about my media habits, and how I could improve them. So I audited my habits (with a spreadsheet and everything!) — and what I learned might surprise you.
Current clickbait solutions
I’m not alone in my anger about clickbait and my desire for a better media diet. There’s plenty of mocking commentary about this, like The Onion’s satirical site ClickHole, or the amazing Twitter feed Saved You A Click by Jake Beckman. Example:
It’s becoming an arms race: One side makes ever-sneakier clickbait … while the other side makes ever-better filters. Much of this stuff gets spread via Facebook, yet Facebook knows that it’s a threat to user satisfaction, so the company is working to make the news feed algorithms less vulnerable to bad content in clever packaging.
On the publisher side, we have examples like The Daily Beast, which just got its “newly quantified and gamified news reader” profiled by Nieman Lab. The Beast’s reader organizes Beast content for each user, and it allows readers to signal that they didn’t like a certain piece and thereby see less of that in the future. Of course, The Beast isn’t a clickbait publisher, but its app allows users to slowly and intentionally improve their media diet. I’m looking forward to more of these.
Another quantificational example is Nuzzel, a service that trawls your Twitter and Facebook feeds and pulls together the most popular links that have been shared by people you follow. I use Nuzzel, and it’s really handy. Yet, as you’ll see from my findings below, a fair amount of irritating content gets into my feed via Nuzzel.
On the other hand, qualitative solutions typically involve human curation. Like most media professionals, I’m on Twitter all the time, and I use Twitter lists to track a small core of people who provide consistently good content. (My primary list is secret, so I can’t give you a link. Sorry!) I’m hardly the first to point out that Twitter is a tough product to master, but if you invest many hours in curating your sources, it’s unbeatable.
And I actually pay (gasp) for one human-curated media source — the LAUNCH ticker, which was founded by investor Jason Calacanis. His team collects news and commentary about tech and digital media. The Ticker’s sensibility is close enough to mine that at least one item per day is genuinely interesting, and it keeps me up-to-date enough that I can skip trawling big tech sites’ daily newsletters. For each item, the Ticker team provides one-paragraph summaries, which give a lot of information so readers can easily skip clicking through. (This contrasts with most publisher newsletters, of course, because most publishers have incentives to convince readers to click.)
Although there are a lot of items in the Ticker that I don’t read, it has a better signal-to-noise ratio than most newsletters. It’s also worth noting that Calacanis has a technical product, Inside.com, that’s like Twitter except that it offers one-paragraph human-written news summaries. Inside.com has potential, but it hasn’t yet earned a permanent slot in my media perambulations.
How I audited my media habits
I acquire links in three basic ways. One is that I find them while specifically researching topics. Another is that they are sent to me personally by friends and family. The third is that I receive them throughout the day via mailing lists, digital communities and social media feeds. While auditing, I ignored the first two categories and I focused on the third — the links that I passively receive in an intentional way.
For one month, I rated every link that I passively received on a 1–5 scale:
1: “I’m actually angry that I clicked on this link.” (Example: A newsletter wielded my fashion anxiety against me, so that I clicked on a slideshow of “5 Style Traps To Avoid.” I then unsubscribed in a fit of rage.)
2: “Ugh. This might be useful, someday, sort of, but it’s unlikely.”
3: “Okay, I can see this being useful under the right circumstances.”
4: “Wow, this is really cool or useful. I’m glad I saw this.”
5: “Maybe I’ll write an article about it!” The piece was so useful, interesting or funny that I find myself bringing it up during “real conversations” in the “real world.” (Example: This article in The Economist about major players in “dark net,” i.e. the underground and un-tracked internet economy, came from my curated Twitter list and it’s fascinating.)
I also tracked various characteristics of each link:
- Where I saw it
- Whether I shared it on social media
- Whether I saved it in one of my bookmarking services
There’s a blank version of my spreadsheet here, if you want to copy it and try this yourself. After you’ve tracked all your information, you can make an assortment of tables using the Google Sheets “Pivot Table” feature. (For instance, to sort your link awesomeness by source, make a pivot table where Rows is “Group by: Source,” Columns is “Group by: How interesting or useful or fun this link is 1-5,” and Values is “Source” summarized by COUNT.)
Finding #1: Things I regretted clicking on
It wasn’t all traditional clickbait-like top-10 lists. (I have no idea how they got me to click them, I swear! I was tired and distracted!) I feel uncertain about how to evade listicles, but I think (hope) that I’m clicking on them less and less?
Anyway, the other things I rated #1 fell into two loose categories:
Cynical, thoughtless zeitgeist capitalizers. These are the people who write 500-word fluff posts about something that’s currently fascinating (say, Uber). Then the writer slaps a title on it that makes it sound like they’re doing legit analysis. This is hard to dodge — my main strategy is to start ignoring sources that link to things like this, but I can’t always do that, because surprisingly good sources occasionally contract this illness.
Personal insight posts from names I didn’t recognize. Typically, I thought these would be interesting because they were about topics I find interesting (e.g. “Why I’m Quitting The Newsroom”). But it turns out that these posts are almost always useless to me unless they’re from someone whose opinion I already care about — a friend, or perhaps a thought leader who’s written good analysis elsewhere. Hopefully, this is something that I can immunize myself against by recognizing the trend.
Aside from these types of clickbait media, there was one specific stream that I will not name, which yielded such consistently awful content that I started leaving giant bolded notes to myself in the middle of my spreadsheet: “Don’t click on [offending stream]!”
This stream is part of a product that I use a lot. So at first, I couldn’t figure out a way to avoid it — and the parent company has a lot of data about me, so it constantly matched me with headlines that I couldn’t resist. (No, it’s not Facebook.) Finally a friend showed me how to use the AdBlock Chrome extension to block specific areas on a webpage, and I eliminated the offending stream from the product. When I’m in a charitable mood, I assume that this company’s content team is saddled with terrible incentives and bad work culture, but I’m still annoyed at them. It really harmed the brand in my eyes. Don’t be this brand.
Finding #2: Content observation anxiety
I’ve always considered myself an open book, but there’s a lot of stuff that I save for later without sharing on social media. This is partly because I save some stuff that’s not top-quality, yet I want my social presence to distribute only the highest-quality information. But it’s partly because I feel self-conscious.
Several recent surveys have discussed the fact that what people want to be seen reading is different from what they want to read. Indeed, I save a lot of links privately, in a format that I believe others can’t see. And I was hesitant to show friends my spreadsheet, even when they asked, because I wasn’t always sure how they would interpret my interests. I found myself most self-conscious about reading content that was critical of the industries where I work.
I suspect that my feelings on this are typical, which means that the designers of filters and curation services must be careful about what information is public — including the information that’s openly recommended to their users. Most curators are already aware that sexual content should be filtered, but there’s lots of other content that might not seem controversial, especially if it’s being gauged by an algorithm that can’t measure sarcasm.
For instance, let’s say the user is a full-time employee of Corporation #1 and he’s been reading articles about his workplace. What if he’s at work showing something to his boss on his screen, and a recommendation service pops up with a new article critical of Corporation #1?
Finding #3: What are media brands good for?
Of the best content that I came across during the month (i.e., rated 4-5), there was only one media brand that came up more than once. (It was Wired, in case you’re curious.) This doesn’t motivate me to visit big-name media sites for the purpose of discovery.
This backs up other trends that media people have been discussing: For instance, as streams have begun to dominate, visits to media brands’ home pages have gone way down. John Herrman at The Awl predicts that large publishers will soon perceive home pages as a liability — “uncomfortably revealing glimpses into the social sausage factory” — if they don’t already. Blogger Chris Blattman has noted that, in his experience, audiences now read more blogs but fewer posts by each blogger.
However, there were publications that were clearly consistent at a certain rating level. This motivates me to read those brands’ stuff more frequently, and more carefully, when I’m directed to it.
Also, when I’m actively doing research, writing an article or drawing important conclusions, then I consider the source. Sometimes I’ll specifically search for what a respected source has to say about a topic: For example, if I’m researching the business fate of mobile games, I’d look for “Harvard Business Review mobile games.” So media brands still count! (I think.)
Finding #4: My best sources
Over the course of the month, more than a third of the links I clicked came from Twitter. A quarter came from the LAUNCH Ticker, and a sixth came from Nuzzel. The rest came from a scattering of other sources. I was surprised to see how much I get from Twitter, Nuzzel and LAUNCH combined, but it made sense when I looked at how well those sources perform.
Between them, LAUNCH and Nuzzel ensured that I didn’t miss anything big in tech or digital media. LAUNCH also had the distinction of never sending me anything terrible (that I clicked on). This might be because LAUNCH includes a one-paragraph, human-written summary of each news item, so I already know a lot by the time I click through. I plan to look around and see if I can find more lists like this.
My Twitter lists performed very well, especially my main private list. This is unsurprising, since it’s a list of sources that I’ve carefully curated over years on Twitter. Sometimes I look at my main Twitter stream — all the people I follow — but that’s much less productive than my lists. So Twitter has been a good return on investment.
Overall, the quality of stuff I got through Nuzzel was least consistent of my top three sources. Since Nuzzel tracks everything that’s “most shared” among people I follow on Twitter, it makes sure that I see the top stuff, but it surfaces a lot of terrible share-bait too. Also, an unanticipated side effect of Nuzzel is that I feel nervous about following many new people in new disciplines on Twitter because that might throw off my Nuzzel ecosystem. I fantasize about the day I can apply Nuzzel to my Twitter lists so that these problems are mitigated (a feature that’s rumored to be currently available to Nuzzel employees).
However, much of the good stuff I see is only shared by one or two people I follow — people who are out exploring the edges and making unexpected connections, rather than reading the same stuff everyone else is reading. Ultimately, the smart people who explore outside their filter bubble are useful sources for that reason alone — and I plan to ensure that this is part of my ongoing value.
Finding #5: The infamous Facebook
When I started this project, I instinctively felt as if Facebook wasn’t a good source of media for my purposes. Sure enough, it wasn’t. I clicked only four links from my Facebook feed during the month, and all of those links ranked 3 or lower on my coolness scale.
Sometimes it’s more productive to look at my Facebook wall in an unsorted state, sans algorithms, but not always. (You can access your own unsorted Facebook wall by clicking here.)
However, I suspect that once I finally get around to setting up a Facebook list, the media I acquire there will be more useful. I haven’t done this yet because with over 1,500 Facebook friends, I know it’ll take time to go through them all and set up my first list — and I’m already getting such good results from my other sources that I haven’t felt motivated enough. New Year’s resolution.
This exercise highlighted the amount of time and work I put into curating what I read. I understand why the average consumer throws up their hands and immerses themselves in their Facebook feed. But this has also shown me the width of the gap that Facebook has to cover.
This exercise was time-consuming, but I’ll probably do it again sometime this year. (I’ll do it after I make my Facebook list and can include it among the sources I measure!) I have a better understanding of what clickbait looks like, the difference between keeping and sharing content, and how I can continue to hone my content acquisition, and after reading this story, I hope that you do, too.