Discover more from People vs Algorithms
Media, Dismiss AI at Your Peril
For better and often worse, digital media is built on hacks. Artificial intelligence will offer the next set and media companies would be wise to embrace it.
Welcome to People vs Algorithms #59.
I look for patterns in media, business and culture. My POV is informed by 30 years of leadership in media and advertising businesses.
Last week, when the Wall Street Journal reported on Vice Media's list of unpaid debtors, the $900K owed to Ranker.com stood out to me as a reminder of the sorry state of digital media and the litany of "hacks" on which it was built. Ranker itself is a fine idea — a website where users submit and rank cultural “lists” like "The Best Marvel Movie Actors Ever" or "Every Major Movie Sequel That Came Out In 2022." Its virtue, unlike most media properties, is users do the work, making the economics much more like its social media cousins. This low-cost, page view rich model is well suited to hostile economic times in digital media. In short, Ranker is a clever "user-generated mouse-trap hack”. But why did Vice need them?
The exchange between the companies amounts to the familiar "attention arbitrage hack" that works like this: Vice sells premium advertising against a brand that promises the attention of culturally relevant young people and the well-produced video that seemingly captures their attention. The convoluted systems that bring people to watch videos on Vice.com, or the social platforms on which it runs the content, does not produce enough viewership to fulfill a good-sized media buy needed to sustain Vice’s content ambitions or business broadly. So one of the many things you do is take your video and hand it to other people, like Ranker, who offer a surplus of low-cost page views inside their "non-premium" brand. You share a portion of the revenue with them to get advertising you sold in front of an audience and fulfill your commitments. Now you owe Ranker money. But you need that money to make your premium media business function. Argh.
Unfortunately, web pages do a lousy job of producing views. Ever wonder why so many seemingly irrelevant silent video players float all over web pages? This, of course, is the "video player page hack", surely one of the most maligned of all media hacks.
Social platforms are where most publisher video is consumed, but without a pre-roll ad product or publisher-friendly revenue share, the trade does not work very well. YouTube is the best on this account, but they take half the revenue, enforce strict ad frequency policies (smartly), and in the end are a notoriously difficult place for publishers to make a buck. Snap was nice for a while, but that opportunity is fading. But advertisers want to advertise on video. You do what you have to do. Hacks!
The aforementioned "arbitrage hack" joins a long line of hacks that sustained digital media. Sadly, many have petered out. Artificial Intelligence (AI) will present the next frontier of hackerdom for those that choose to innovate with and around it. For those that ignore it, the economics will continue to get worse. Before we get to that, let’s look at a few more.
Historically, media has invented great products to make money with video. Cable television is the best example, but few digital publishers ever got to taste its sweet nectar. Cable benefited from the ultimate "bundling hack" whereby most of the country pays monthly for channels they rarely watch, with the added benefit of a large, frictionless ad market providing a generous second revenue stream. Admirably, Vice benefitted from this hack and still makes good money here. But this is slowly evaporating too.
There "roll-up hack" was also quite popular for a time, wherein a company like Vice would roll up the unique visitor numbers of completely unrelated sites like OMGFacts, Dose and Distractify to bolster a comScore position in order to show advertisers dominance in an audience vertical and say hacky things like "we reach more influential millennials that other media brand in the world," a proud gesture that didn't seem to move the needle much.
I would also put "search engine optimization hack" on this list, but creating content to serve hungry Google searchers is really less a hack and more an industry neurosis. At least it is useful to a consumer. For now.
I could go on. Some hacks have disappeared like the advertising "pop-up hack" or worse, the "pop-under hack". The "email list building sweepstakes hack" sustains and is favored by traditional magazine publishers. The ad impression spewing "slideshow hack" has also stuck around because consumers seem to like clicking buttons to see celebrity photos.
In truth, making money in digital media has always been a hackathon — a constant dance to find the next content or ad format, support the next distribution point, embrace the next ad tech optimization, and drive the next impression — like a shaky hand probing a needle-marked arm and tired vascular system in search of the next fleeting high.
Perhaps we can characterize them. Hacks usually fall into one of three categories:
1) The short term, "do what it takes to make money hack" which is good for a while;
2) The much more rare hack that starts like the above but turns into a "great sustaining business hack";
3) Those that push legal boundaries and "don't end well hacks".
Facing a difficult media reality and wanting to raise money on an illusion, media bullshitter and Ozy Media CEO Carlos Watson succumbed to 3) and was arrested in Brooklyn on Wednesday on security fraud charges. Better to stick to 1) and 2).
When advertisers want what you don’t got
Now the industry is looking for the next frontier because the current set of hacks are not working very well. That certainly is the sense I am getting. I have taken three calls this morning from wonderful people currently or recently employed in the digital media space who might describe the grass as "brown, everywhere". It bums me out.
Because, despite the litany of hacks, historically digital media was full of excitement, innovation and promise. We danced. With the unstoppable share-gobbling platforms that offered better proximity to the consumer and with it, data and scale. With the commoditizing inevitability of programmatic. With content marketing, and how making stuff for every advertiser polluted our margin structure. We scoffed at the “influencers”.... nah, we were important media brands, the OG and only important influencers. There was the insanity that is GDPR, the new art, and new expense line of "consent management". Cookie depreciation came to slowly rob content publishers of targeting signal transfer. Subscription... everybody move to subscription! Affiliate, a temporary godsend…
We danced awkwardly with video cause video was cool. We wanted to do more but without a place to play it, it too became the ultimate margin sucker. Few crossed the video chasm.
In the end, advertisers want what they want, and, like it or not, they want more video and quantifiable performance. The video side proved difficult without a reliable distribution mechanism. Performance demands data, purchase intent and scale which few publishers can offer. Here, retail media is stealing share like hungry raccoons. Retailers have two things publishers do not; users looking to buy something and zero content costs. Today at $37B in revenue, Amazon's retail media revenue alone amounts to 80% of the entire publishing ecosystem.
And, as Brian Morrissey. would say, the enterprising are left as "general contractors" for brands, leveraging hollow brands and nimble service competencies to make low-cost content and fancy event "activations." Contractors are good hackers.
The next hack
Alas, there's a potential new hack on the horizon, but as much of a threat as it is an opportunity. This depends on how you see things and how enterprising you can be. It is, of course, AI.
Among other things, AI threatens to upend the only reliable distribution mechanism left for publishers, search. When the AI pulls the guts from everything that publishers work hard to create only to repackage that content in a tidy, custom-made response to a question, there's reason for concern. Do not for a second believe Satya Nadella and the nice people at Microsoft that the inclusion of "citations" and web page references in the AI response will in any way replace the spigot of traffic that traditional search produced. The web page was just a formatted response to a search query. More of that response is being absorbed into the chat interface as the web is reordered around chat-powered AI. Attention and value will shift commensurately.
There’s no turning back. Capitalism likes nothing better than technology that disrupts existing business models with new ways to serve a consumer and better economics. AI offers it in spades. Unlike crypto, the early use cases are obvious and as enticing to the user as they are to business innovators. It promises to reinvent every category with unprecedented human-like personalization at a fraction of the cost and in many cases, a vastly superior product. Wharton Professor Ethan Mollick makes the point well. AI is a source of competitive advantage in so many places. It's gonna spread fast:
There is no doubt it will have a large effect on anyone doing information-based work. Early AI assistants, like Copilot, already cut the time for complex tasks like coding in half. This will do the same, or more, across many industries. I think every organization that has a substantial analysis or writing component to their work will need to figure out how to incorporate these new tools fast, because the competitive advantage gain is potentially enormous. And there is no instruction manual. You can only learn through trial-and-error.
Don't be fooled by naysayers who seek to "unplug" the smart robots before they become fully sentient. Ignore the spate of AI gotcha stories, who seek to unmask the demons underneath of AI as proof of anything other than QA'ing a chatbot is more like raising a child than catching software bugs. Here I agree with Mike Solana at Pirate Wires who pointed out the insincerity of journalist AI fear-mongering, like the NYT's Kevin Roose baiting the Bing bot by leading it to the nasty content that Roose was hoping it would spew back at him.
While large language models (LLM’s) have effectively modeled the relationships between words in ways that are highly effective and semantically sound, it is still a long way from human cognition. LLM's models skim the surface of human knowledge. There’s lots of good stuff to read to understand this better. This from NYU professor Yann LeCun and postdoc Jacob Browning is good. Want more depth, take some time with this lengthy explanation on how ChatGPT works from Stephen Wolfram.
It doesn't really matter to me if this is "intelligence”, fancy automation, or just a better toolkit. AI changes media. Jon Stokes, co-founder of Ars Technica summarized his take on the state of things, which I happen to believe. A point that's been playing out in media for several years. Real people, lots of people can make incredible things, shitty things, informed things and entertaining things. They do this for lots of reasons, not the least of which is "just cause." This, of course, creates a surplus of supply against the diminishing returns of attention. The value of most media drops.
Almost all of the fears around AI that are circulating right now — fears of cheating, of “disinformation,” of scamming and spamming, etc. — all actually boil down to one thing: fear about what happens when a whole bunch of people, some of whom are stupid and/or irresponsible and/or malicious, instantly level up and get really good at making cultural objects....
Again, all of this amounts to a fear that randos will suddenly get creative superpowers that put them on par with the vetted, the certified, the bona fide, the trained, and the institutionalized.
It’s kind of wild to me that so many people are deathly afraid that the masses will suddenly get really good at making high-quality arguments, short stories, paintings, photographs, movies, etc. You’d think we’d all be excited about the impending flood of high-quality everything. Yet here we are flipping out over the prospect of unprecedented cultural abundance.
Embracing a strange future
It seems to me there is only one path forward and that is to fully embrace the change that AI is presenting us, the next great hack. I would break these down into six new frontiers:
More efficient creation: the ability to make more and better with less. It doesn't have to be worse, to be very clear. A person with better tools for creative inspiration, for research, for editing will make better content for less. A smart person will use tools to differentiate, not overload the system with cheaper crap. Anybody can do that.
New formats: Chat plus search is a new way of accessing information. How can our brand play a role in making that experience more effective and trustworthy? How might AI search against our repository of content make the experience better for the user? How can we use tools to create completely new categories of entertainment or news, like we are starting to see with weird experiments like the goofy "perpetual Seinfeld machine" on Twitch. How might chat like service augment content experiences in new ways?
More effective personalization: How might AI with some knowledge of my interests and preferences create wholly new and individualized media experiences just for me?
Better curation and discovery: This is where much of the AI innovation has been to date with algorithmic-driven feeds. Early reactions from the Supreme Court justices suggest section 230 may evolve, but the notion of algorithmic driven content will persist... largely because the old way will never worked in the modern sea of content. But, the idea of personalized, roll your own, algorithms are bound to gain popularity as more people look to have content tailored to a specific point-of-view.
Better advertising: Advertising is essentially telling stories and matching buyers and sellers. It's the later part that's gonna get more interesting, particularly as AI gets unleashed on mountains and SKU and pricing information from every corner of the commercial world.
Cheaper software development: Enabling the above has been extremely difficult for content companies whose core capability was not software. AI will make new application development accessible to more companies. A raft of API’s from Open.ai and others will open up new frontiers.
I am sure there are others.
But imagine yourself as an educator facing a new reality where much of how you teach is undermined by the tools suddenly available to every student. You either ban them or embrace the pedological shift and figure out how we make better minds and outcomes with them. Ethan Mollick at Wharton has elected to embrace the change:
However, based on my experience, I think focusing on how people use AI in class rather than whether they use it will result in better learning outcomes, happier students, and graduates who are better prepared for a world while AI is likely to be ubiquitous
I would suggest the media world does the same.
Have a great weekend…/ Troy
On the podcast: Publishing as hackathon
Software has eaten the world, and for publishers that means a continuous battle to fend off second-order impacts of decisions made by tech companies battling each other or governments attempting to rein them in.
This week, we discuss the second-order impacts cascading through the industry from attempts to regulate digital privacy and from, of course, the development of AI.
Terrific listening, says everyone that listens. Listen here.
I was reminded.
Somehow I stumbled upon this… a reminder of my youthful love for the Violent Femmes and how they hacked the vibe. And below, as bonus content, a charming interview with the local news host in Milwaukee.
Violent Femmes are an American folk punk band from Milwaukee, Wisconsin. The band consists of founding members Gordon Gano (guitar, lead vocals) and Brian Ritchie (bass, backing vocals), joined by multi-instrumentalist Blaise Garza (joined 2004), and drummer John Sparrow (joined 2005). Former members of the band include drummers Victor DeLorenzo (1980–1993, 2002–2013), Guy Hoffman (1993–2002), and Brian Viglione (2013–2016).
Violent Femmes have released ten studio albums and 15 singles during the course of their career. The band found critical acclaim with the release of their self-titled debut album in early 1983. Featuring many of their best-known songs, including "Blister in the Sun", "Kiss Off", "Add It Up" and "Gone Daddy Gone", Violent Femmes became the band's biggest-selling album and was eventually certified platinum by the RIAA. After the release of their third album The Blind Leading the Naked (1986), the band's future was uncertain and they split up in 1987 when Gano and Ritchie went solo. However, they regrouped a year later, releasing their fourth album 3 (1989). The follow-up album, Why Do Birds Sing? (1991), contains the fan favorite and concert staple "American Music".