Relying on indirect metrics will sink you
Recent upheaval in the publishing sea deserves a deep dive into the metrics we rely on – and a peek at what we could be doing better.
The content triangle
The business of online publishing looks like a triangle.
At one vertex are the content producers (traditionally also the owners), with readers at a second. There's a long standing and implied relationship between the two: I write, you pay me to read it.
That model has been under pressure for at least 15 years. Declining readership and income since the internet became global meant that revenue had to be found elsewhere. That's advertising: the last vertex in our triangle.
Advertisers is a catch-all term which really includes demand side platforms, agencies, and some analytics companies. Anyone who makes a living by understanding online audience engagement – who tries to find the knobs and disks that increase that engagement, slips neatly into this bucket.
That bucket thrives on ‘eyeballs’ – readers, to you and me – and it’s a straight numbers game. Finding one million eyeballs is better than one thousand, because a million clicks translates to more sales opportunities.
Casually seen as less important was the question of which eyeballs those were. When the target is to beat last quarter's numbers, volume is the fix-all. So the advertising world is intrinsically disinterested in what the content is, as long as those readers are reached. And, for the longest time, reach meant revenue.
But because of the distinctions between the three players in our triangle, there is no direct relationship between reader, writer and funder. This set up relies on indirect metrics. The number of shares or clicks on an article has come to define the article in terms of the value can be ‘extracted’ from it. But that value is not measured in terms the reader might align with.
BuzzFeed and Feeding the Beast
That glaring inequality was smoothed over for a while. When clickbait was seen as fun rather than unhealthy, who really paid attention? Buzzfeed did. They played the numbers game better than most, and basked in online glory for a decade. But their business relied on indirect relationships – that triangle – and so they were, fundamentally, always at the mercy of tech firms and their streams of data. Offloading your distribution to platforms also often comes at the expense of you developing a real relationship with a committed audience and their “user needs,” as we discussed in our last newsletter.
We did some work to back up that theory. We know that BuzzFeed generates traffic and revenue for some of the large platforms. Looking at a recent batch of about a thousand articles published by BuzzFeed and BuzzFeed News, we looked at the various types labelled by Overtone’s AI – from Quick Hits through In-depth Opinion – and tracked how well they performed on social networks and in search results (and their estimated traffic from those results).
A few findings:
Buzzfeed News (now closed) and normal Buzzfeed (still operating) produce very different content mixes, with News understandably producing more in-depth, newsy pieces and Buzzfeed more snackable content.
These different types of articles also perform radically differently on social vs search results. For instance, higher depth pieces like Investigations are better highlighted by Google’s search algorithm and lower depth pieces like Quick Hits and Update Me stories do better on Facebook.
The other factor of course is that it takes more effort (and resources) to produce in-depth journalism than it does to produce low-depth snackable content. Buzzfeed seems to have chosen to keep producing the later while cutting the former. Here’s what that looks like.
Essentially, what we see is that (outside of news) BuzzFeed is making content that does well on other people’s platforms – but isn’t so great for themselves. Compare that with the non-news content, below. The purple area for regular content shows a disconnect with metrics-driven, implied performance. Sound familiar?
More results coming soon: let us know if you’d like to review the worksheet.
Relying on indirect metrics is like sailing a boat blindfolded.
You get some feedback from the waves you navigate, but you can’t feel the storm coming. And, as we’ve discussed before, huge changes are sweeping through our world. Is the problem specific to BuzzFeed? Not at all. This week, we learned that Vice is filing for bankruptcy protection. Whispers on the grapevine mention ongoing ripple effects. This is a systemic problem.
Our agility in dealing with such shifts depends on the foundation we build our businesses on. In an ocean of content, relying on indirect metrics will sink you.
To give all this some context, take a look at a selection of news and non-news articles from Buzzfeed, with Overtone assessment for each. What are your thoughts – which would have done well on social and search, and which should have done well?
BuzzFeed News
This article investigates the age of Tom Sandoval, a reality TV star from the show Vanderpump Rules. It gives fans’ perspectives and provides background context. (Depth 5, conf 0.98)
The article delves into the experiences of tech professionals who unexpectedly lost their jobs during a wave of industry-wide layoffs. It follows along with people’s experiences showing, for some at least, an unexpectedly better future. (depth 5, conf 0.79).
This article (which is longer than its headline), examines the postponement of a social media influencer’s trial. It provides insights into the reasons behind the trial delay and offers an overview of the current legal proceedings associated with her case. (Depth 5, conf 0.96).
BuzzFeed Regular
This article aims to be a comprehensive guide to essential air travel rules. (Depth 5, conf 0.73).
The author shares highlights from an interview conducted on Rye Lane, a vibrant and diverse street in London. The piece provides insights into the experiences and perspectives of local business owners. (Depth 5, conf 0.68).
This article explores near-death experiences and shares personal accounts from those who have gone through them. It delves into the profound effects of these events. (Depth 5, conf 0.98).
Updates
Last month was busy for Overtone. Saman Nia, our machine learning engineer, represented Overtone at WAN-IFRA’s data science day in Vienna.
Meanwhile,
tuned in to the News Product Alliance summit.Early June, Overtone will be attending the Media Party, where we will be bringing our capabilities to life with some of our partners. If you are in Chicago, please let us know. We’d love to see you!
Someday, eventually, someone will figure out how to, um, square the triangle. Maybe.