Reminder: The Rebooting’s next Online Forum is a deep dive into how Forbes has unified its audience data. Forbes chief innovation officer Nina Gould will join me and BlueConic’s Matt Jarman to walk through how Forbes has wrestled its disparate sources of audience data into a unified view. Nina will provide examples of how Forbes is using this unified view to create personalized experiences.
Join us on July 30 at 1pmET for this live interactive discussion. If you cannot make it, registrants will get a link to the replay and a guide to the takeaways.
Today, a TRB Pro members piece on idiosyncrasy as a strategy and the emerging and messy contours of our post-text, post-webpage world, as seen in the shifts to livestreaming, Substack's big new funding round, Google Discover as the next shoe to drop, Anonymous Banker on the folly of publishers building their own AI products and why AI browsers are yet another Very Big Deal.
At one of The Rebooting’s private dinners last week, I asked a publishing executive how he expects his company’s output to change as media slips fully into an AI age. The publisher focuses on newsletters and a “faces” strategy of high-leverage individuals complementing the brand. This executive expects more idiosyncratic approaches.
While much of the focus of the AI-and-media discussion has focused on the important issues of getting paid for training data and driving efficiencies in a more-with-less era, comparatively less has been placed on the essential question of how publishers can win loyalty in a world mediated by AI.
That's led many to lean on idiosyncrasy, both in terms of the brands they create and business models they pursue. In the Traffic Apocalpyse™, nobody knows what their businesses will look like in three years, so you might as well just do things. That will lead to messy brands and messy strategies.
The best media brands traditionally have institutional voices. Say what you want about The Economist, it is certainly consistent. Humans are by their nature messier than the artificial construct of a brand identity. Humans tend to be contradictory. Media brands themselves will need to become more like big-tent political parties with enough room for dissenting voices and side quests.
A brand like Emily Sundberg’s Feed Me is a case in point. It’s very idiosyncratic because, as Emily likes to say, it’s a product of her experience in the world. That makes it a difficult brand to pigeonhole. The organizing construct, to me, is around what it’s like to be a striver in New York. The flattening of global culture means that audience is worldwide, not just literally in New York City. Brands like Feed Me cut across categories in ways typical media brands do not.
Fresh Hell is a good example of the type of idiosyncratic publication that will do just fine in a world of AI slop. Tina Brown is a great writer. Her writing has pace. You can feel the momentum, and she knows when to drop in little one-liners that can even elicit an audible chuckle. This is the kind of writing that isn’t optimized to get the attention of algorithms or AI agents.
These brands are smaller. Feed Me gets outsized media attention as some kind of Gen Z Whisperer because it is today’s version of Gawker, only optimized to a vastly different NYC where al fresco dining sometimes happens in St Mark’s, wearing around yoga outfits is as acceptable in the West Village as in Houston, and TikTok trends like Labubu and Dubai chocolate have as much resonance. For all the New York or Nowhere merch, NYC is as much like anywhere else as it has ever been. The best restaurants end up having outposts elsewhere. You can go to Lucali in Carroll Gardens or in Miami Beach.
Feed Me and Fresh Hell are both Substacks as well. Substack’s $100 million funding round is testament to how it's managed to become a consumer brand – I tense when mislabeled as a Substacker although I've come around to B2B influencer, if for a lack of more palatable options – and also a sign of how text has its limits. Substack’s bet is it can create an everything app that’s beyond newsletters. For instance, Substack-hosted publisher The Free Press has seen success in its livestreams as a conversion tool.
Text content is an area many are fleeing, or at least rapidly diversifying away from. Felix Salmon was ahead of his time a decade ago when he declared himself “post-text.” The Information’s oddly named TITIV is a sendup to TBPN, the VC-friendly counterweight to the independent media tech companies tend to dislike. TBPN showed the pathway to using livestreaming and insider access to build a brand, even if most of the engagement on TBPN is not actually live. It’s all about the clips that go far and wide. Cycling in many guests means more clips, and more clips means more distribution. People love seeing themselves on video.
The Information's TBPN homage was glitchy on the first day when I tuned in. You gotta take your lumps It felt too produced. The best livestreams tend to be from people’s kitchens. And a lot of TBPN’s success comes from the host dynamic. That’s hard to cast. Idiosyncratic media needs to present as organic.
I applaud The Information for making the bet. I’m working on a something along these lines for TRB, in large part because I want to save the reputation of the webinar. What is a webinar if not interactive digital video programming? The live angle, to me, is only of use if you make the program interactive, otherwise it’s just a schtick. This is a good time for met to remind you to attend our next online forum. It’s on July 30, featuring a deep dive into how Forbes has unified its audience data. Sign up.
Publishers are at risk of making too many small bets at a time of uncertainty. Those reliant on text content can see the trend lines and they’re not great. Publishers are fighting rearguard battles to get paid for their text content, but it’s going to be difficult to make this sustainable. Text is just too easily reconfigured and summarized. A news article is quickly regurgitated on Reddit and elsewhere.
One question that swirled is a kind of naive media question: Do the tech overlords care if they drive so many out of business? The sensible response was absolutely not, they run businesses and generally billionaires become billionaires by being fairly ruthless. There’s PR, sure, but Mark Zuckerberg will come around to never being particularly popular no matter what kind of haircut he gets.
This isn’t new. The aggregation economy provided incentives to aggregation publishers to draft off the original work of others. Why go through the hassle and expense of an investigation when you can aggregate it in a slideshow and bank more traffic? The modern version of this is podcasting and newsletters.
Substack is making traction convincing major publishers to dip a toe into its ecosystem as one of dozens of hedges against the Traffic Apocalypse. And it is finally getting over its aversion to advertising and following the typical Silicon Valley path of hating advertising until they run the numbers.
The traffic woes publishers are facing are going to get worse as Google moves onto Discover. I can’t over-emphasize how important Google Discover has become to many publishers as a reliable gusher of traffic. The reality of all these models is people are operating the old ones while building the new ones. AI Overviews coming to Discover is likely just some product manager trying to hit an OKR. And it could have massive downstream impact on publishers. Tough business.
I co-hosted a salon gathering last night at Joe Marchese’s Casa Komos, which has an essential design feature most offices need: the front desk of the Union Square office space converts into a bar. The theme of the evening was, of course, media in a post-traffic world. One exec joked to me that publishers should state their individual p(doom) number. Seeing off existential threats has become a habit.
The optimistic take on Google Zero is that it doesn’t literally mean zero and decay is typically a long process. I’m reminded of the shift to the internet, when a generation of middle-aged executives tried to run out the clock since your personal risk profile changes after 50.
The conversation veered into the need for publishers to make better AI experiences with their content than the AI companies do with their content. This is sensible. Too often, publishers have waited for relief from regulators or the courts.
Anonymous Banker was unimpressed. We huddled in a wellness room for an impromptu podcast segment at the party. A staffer apologized after entering the elicit podcasting scene. AB is a good reality check. He pointed to the obvious: publishers are so far outgunned in a world of $100 million a year AI talent. That’s why you’ll hear more about throwing parties for brands.
The interesting dance publishers are playing now is they’re trying to negotiate deals with these massive AI companies while also not closing themselves off from them. It might feel good to push the red button and say “enough is enough,” but it is also insufficient. Publishers need to change their products, and they’ll need to work with these AI platforms as well as fight for their rights.
Hearst and Condé Nast signed with Amazon to license their content for use in Amazon’s Rufus AI bot. This is a good compromise for what to do with all of this affiliate content publishers have been creating as part of their SEO strategies. The problem for all of these deals is the money in the new deal rarely replaces the money in the old deal. The math doesn’t math, at least not nearly as well.
NotebookLM doesn’t get as much attention as Google’s other AI products under Gemini. But it’s seemingly onto something as an AI consumer product. The Economist and The Atlantic are “featured notebooks” that give a glimpse into what feels inevitable: a choose-your-own-adventure approach to content that is multimedia and audience-directed.
Publishers have taken baby steps in this direction, letting people listen to a story, for example, or see transcripts of interviews or annotations from experts quoted in The Washington Post’s case. These are scratching at the surface. I continue to ask at our dinners and gatherings if people expect to be shipping article pages in five years. It’s well past time to rethink that atomic unit, particularly as its economic utility fades.
The NotebookLM experience is still clunky and a good reminder that for all the breathlessness these AI tools are still flawed. The new focus is moving g to the browser level. The battle for search will move there to make these agentic visions a reality. We might not get flying cars, but we will get robots booming our vacations.
It moves up the p(doom) number on the webpage back button. It’s hard to see how an open web of cul-de-sacs fits with AI browsers that do the work. And those will be built without much publisher input. An exec testing Perplexity’s said he got in off the public waitlist despite being a Perplexity partner. Some things don’t change. Publishers are far downstream of tech.
Which is why many don’t know what to build. So much is in flux that publishers are derisking with a grab bag of incremental changes that are likely to fall short of changing habits. As one exec said at the Media Product Forum, “We do just enough to stay behind.”
On last week’s PvA, we discussed a recent study into how Gen-Z consumes news content. It turns out that they scan headlines and heavily rely on the comments to understand whether to trust it.
This is a notable change. Institutional publishers are struggling to keep up with many distribution and monetization challenges, but the biggest might be the eradication of this default trust that comes with an institutional brand. I’m not very optimistic about rebuilding that trust with little product tweaks such as adding a bias meter. I suspect the product will need to be rethought from the studs to maximize transparency, participation and personalization.
Thanks for reading and being a TRB Pro member. I know I've promised this, but I have a lot of new initiatives for TRB Pro that will go deeper with expert content about how to build sustainable media business models.