Paradox in the Information Economy: How AI is Increasing the Value of Publishers
Emma O’Brian:All right. Well, I think we’ll kick it off and get started. Thank you all for being here for a conversation about the AI paradox, how AI made publishers more valuable, not less. I’m Emma O’Brian,. . .
Emma O’Brian:All right. Well, I think we’ll kick it off and get started. Thank you all for being here for a conversation about the AI paradox, how AI made publishers more valuable, not less. I’m Emma O’Brian, General Manager of Dow Jones Factiva. If you’re not familiar with Factiva, we’re a B2B content marketplace with arguably the largest number of Generative AI licensed sources for enterprise use. Emma O’Brian:A few years ago, the prevailing wisdom was that content would become increasingly commoditized. Then generative AI came out, and it made content creation dramatically easier, faster, and cheaper. By that logic, content should be worth less than ever. But over the last two years, we’ve seen publishers sign major licensing agreements, enterprises are becoming more concerned about trust, and advertisers continue to pay premiums for certain environments. Today, we’re bringing together perspectives from publishing, marketing, and advertising to explore what I think is one of the most crucial questions in media right now: What actually creates value in an AI-powered information economy?Emma O’Brian:With me today, I have Stacy Kemp, executive lead for the CMO program and principal at Deloitte; Will Dougherty, SVP for Inventory Partnerships at The Trade Desk; and Scott Havens, chief growth officer and head of consumer at Dow Jones. Thank you all for coming. We’ll start with a kind of open-ended question: Is the premise even true? Maybe Scott, we’ll start with you. Two years ago, what was the most valuable thing about being a publisher, and what is the most valuable thing about publishers today?Scott Havens:Thanks for having me, Emma. I would like to say that I, for one, am very excited for the era of AI because I have worked with premium content brands for the last 20 years and never chased traffic and audience scale for the sake of it, but tried to build content utility for consumers. Two years ago, when search changed and traffic started to plummet, optimization was probably the most valuable currency of the time for digital publishers in particular. The old tricks of optimizing for these platforms fell out of favor. Then the rise of AI created exponential content. Scott Havens:Publishers like Dow Jones andThe Wall Street Journalin particular have seen the value of what we do rise. This feels contrarian or paradoxical because there’s so much out there, but actually, you don’t have any reason to trust what you’re seeing now, whether it’s generated by a machine or even a creator. We’re seeing this play out in a few ways. First, we’re able to strike amazing deals with platforms. We’ve talked about a couple at the News Corp level. This is a real model because these LLMs aren’t very sharp if they don’t have great original reporting and data. Second, we are 75% direct now with our consumer base. It’s partially a numerator and denominator issue, but we’ve got so many people coming directly to us because they are seeking refuge from slop or untrusted content. I’m actually very excited for this because I did not push my teams into search and social optimization in a way that other brands built their businesses on—brands that, frankly, are gone or will be gone. Emma O’Brian:Stacey, you’re at an interesting intersection where you’re advising CMOs, but you’re also managing the CMO representation for Deloitte itself. How are you seeing that interconnection from the customers that you have today, their needs, the AI ecosystem, and trust? How are you applying that to your own business?Stacy Kemp:It’s interesting because I think AI obviously has a big role to play, but this shift would be happening even without AI. Other factors drive opportunities to connect with audiences that are accelerated by the use of AI, but also just offer opportunities beyond what would have existed five years ago. What is becoming most important is that the connection between a brand and the return on advertising across any channel has to be absolutely measurable. Whether it’s the rise of AI or the rise of retail media networks and other ways advertisers reach target audiences at more pointed points of relevance, advertisers can get measurements back and adapt strategies. That is as much a piece of it as AI, and of course, AI accelerates all of that. The more we can be really crisp on the value proposition of what each publisher brings—both in terms of access to audiences and return on that access—is where the value really has to be. Emma O’Brian:Will, you are also at an interesting intersection of ad spend and delivery mechanisms. How do you at The Trade Desk specifically think about that interplay and how that’s changed in the age of AI?Will Dougherty:None of us have a contrarian take—we all agree. We’ve seen traffic change quite a bit over the last couple of years. A lot of what we associate with AI is actually much more of a smoke screen or red herring for changes Google has made. Long-term, to Scott’s point, that is probably going to be the best thing that has happened to durable, strong publisher brands. Search created an incentive for a lot of publishers to chase traffic at all costs. They wrote outside of their brand and their relevance to the consumer, optimizing for a transient experience where they were just trying to get users from Google search onto their page. Content and brand started to splinter from each other. Once that happens, you’ve degraded your brand, and you live at the behest of a search algorithm which you have little to no influence over. Will Dougherty:What AI has done is make information retrieval extraordinarily cheap. It’s very easy to get simple information. Personally, it’s wonderful; I don’t have to click through a bunch of links of dubious provenance to answer simple questions. But when I need an actual editorial piece, a point of view, or something with expertise and perspective, I seek out brands I trust and have a relationship with. From a publisher monetization standpoint, we’ve known for a long time that the most valuable consumers on a publisher’s website, streaming CTV content, or listening to podcasts are the ones who return and keep a durable relationship with that brand. Will Dougherty:In some ways, I think this is a short-term disruption but a long-term benefit, especially for publishers to re-establish their connection, brand, and what that stands for with their consumer audience. It’s been really encouraging how quickly we’re seeing publishers lean into that. I had a meeting earlier with another large national publisher, and last week was their single best day of organic traffic in years, none of it coming from search. It’s all coming in because of investments made directly into their consumer base. As advertisers, we’ve never had more supply, but we have a finite amount of quality supply. We’re getting more signal and data, and where there is high brand equity in a premium ad environment, advertising works more successfully. Like everybody else up here, I’m very optimistic that this will pull us out of a bad direction led by search into something where consumers, publisher brands, and advertising have a much more egalitarian relationship. Emma O’Brian:Thinking about the different inventory on a website or news outlet, there’s a long tail of very trustworthy, reliable content that maybe doesn’t have as high of a reach as mainstream or premium publishers. How do publishers and advertisers truly understand how that information ecosystem changes and capture that value for their content or ad spend?Will Dougherty:No publisher brand will be single-channel anymore. They will exist in a few different permutations, building their audience and brand across many channels—whether through website stories, writers on social media, or premium content in social environments. They need to be consistent in who they are as a brand and cultivate that. The authenticity of that brand becomes more important in an AI-derived world. It has never been cheaper to be mediocre. AI is very good at being mediocre, but it is not great at being great. My advice to publishers is to conduct an honest assessment: Do we know what our brand is? Do we know what we stand for? Are we investing in ways that reinforce that with our audience? Do we know our audience? The same data marketers use to make effective advertising, publishers need to be using to grow an effective audience. Emma O’Brian:That’s a really interesting change of conversation from traffic, looking for a value number rather than just a reach number. How are you seeing that change?Stacy Kemp:That was sort of the point you made about the value proposition of being able to clearly articulate the unique value provided to an advertiser by accessing specific audiences at a specific time. What is exciting about AI is that it gives us the technology to do that at a scale that was previously impossible. If a publisher can stay true to their brand and build trust—which is the ability to build integrity between what you say you are and what you do—they can articulate the value of that trust to advertisers in a language that makes sense against their audiences. That’s the type of conversation we need to have versus broad-brush traffic or search conversations. Emma O’Brian:Scott,The Wall Street Journalobviously has both reach and trust. How do you keep winning?Scott Havens:We are winning, and as I mentioned, this is in many ways a really good phenomenon for us. But we need to stay multimodal and multiplatform. An entire generation raised on phones and social media gets information through the creator economy. Forget about AI for a second; Substack and Beehiiv creators are commanding huge audiences. We have to think about how we show up on these platforms in an authentic way, not just copying and pastingJournalorBarron’sarticles onto Substack, because that won’t work. Scott Havens:We’re also thinking about how we bring our personalities out more. The Wall Street Journalhas built a Talent Lab under Emma Tucker to find, coach, and elevate some of our individual personalities. It’s not just the institution; it’s institution plus personality, which is a far stronger combination. You bring that trust, history, and credibility, but you also have somebody that people feel like they can connect to. We’re doing all of those things in spades right now. Emma O’Brian:How do you think about optimization for LLMs, visibility across distribution models, and reach expectations for publishers and advertisers?Stacy Kemp:It’s an interesting question. I spend a lot of time with retail companies, so it’s hard for me to think about this strictly in terms of content because there is a whole piece of agentic commerce sitting within that. For B2C companies, marketing to an AI agent comes down to price, availability, and speed. An agent doesn’t care about brand values. It’s about thinking through a bifurcated strategy where you look to maximize the power of an LLM in vastly different contexts. That creates a complete rewiring of how operations work, along with relationships with agencies and media partners. Will Dougherty:It’s such an interesting question because when LLMs launched, the novelty was mind-blowing and felt like a magic trick. Very quickly, our expectations rose regarding what we wanted them to do. Even within the LLM community, brand equity and preference developed—using Claude for specific tasks, ChatGPT for others, or Gemini to go through email. We started to build preferences within those ecosystems. Will Dougherty:Next, we’ll see how expertise is understood within those ecosystems. You’re going to see brand integrations specifically inside LLM models. If I want a perspective on the market, I may ask the LLM to show me whatThe Wall Street Journalthinks or who their leading authority is. It becomes less about getting an answer to a question and more about why the question is important and what we learn from the answer. Integrating that expertise into LLMs creates a massive opportunity for the really premium parts of our information systems to stand out. Scott Havens:As Emma knows, we built a connector with Factiva into OpenAI, and I think this will be a really interesting area for publishers to explore. How people interact with LLMs will evolve beyond the current chat interface. Regardless, if Will and Stacey are subscribers toThe Wall Street Journaland want to use an interface powered by an LLM, we want to make that available while recognizing them as subscribers so they get deeper, better information. Scott Havens:We ran a report recently on account abuse to see who was scraping our site through user accounts. We sent a stern note to one user about violating terms of service, and he wrote back saying he built an agent on Salesforce to read theJournaland paid for his account. We realized this could be a new subscription model for us. We unlocked it for him, of course. As AI agents get more sophisticated, we’ll have a huge audience that wants specific pieces or the full breadth of what Dow Jones offers, and that’s exciting. Emma O’Brian:When you think about value capture, what is more valuable—a human eyeball or an agentic eyeball, and why?Will Dougherty:There are different degrees of value. If you’re a commerce brand, having accurate information about your product structured and propagated correctly into LLMs is critical. That’s a skill developing in real time. However, why a human tells an agent to make a purchase—why they choose a brand in the first place—is connected to moments of emotional resonance. Premium content at its best helps people feel something and engages them. If you only harvest demand and never plant seeds, you eventually run out of things to harvest. You need to fill the top of the funnel with your brand story, authenticity, and partnerships with publishers who create those emotional connections. It’s not binary; you let the agent handle the task of finishing a purchase, but why we choose a brand isn’t just about finding the lowest price. The existence of Zara doesn’t make Gucci less valuable. Stacy Kemp:They will both coexist, just as they do today. We’re all learning to operate with agents, whether personal assistants, shopping agents, or task-based tools. Advertisers need to understand their data and leverage both agentic and human strategies. Deep insight into the end customer tells you when to show up with a brand message versus an availability or price message. Orchestrating that messaging at scale was impossible two years ago; now the technology exists. Understanding when a customer is using an agent allows you to provide that agent with the data it needs to push your offer to the top. Scott Havens:As a publisher, we have to be careful that we don’t lose a direct connection with our customers. One of the biggest mistakes of the last 20 years was building on top of third-party platforms that controlled the data, scale, and monetization, leaving publishers unsure of who their audience even was. Whatever we do with AI, we must maintain that relationship and understand our consumers directly rather than outsourcing it to a third party. Will Dougherty:To Scott’s point on planting seeds, for publishers with a clear reason for being and true utility, platforms like Google, Meta, and YouTube have actually been good partners for building future brand awareness. Publishers don’t typically spend a ton on brand campaigns, so showing up on massive platforms reaches billions of people. The mistake was building an entire business model on those platforms without converting users into direct customers. I agree with social platforms for distribution, but my strong position is that they broke the scoreboard regarding who gets credit for driving value. The value should be distributed much more to the edges rather than soaked up in the middle. Emma O’Brian:If we continue that thread, what data is most important to pay attention to in this changing ecosystem if not traffic?Will Dougherty:If I were to stack-rank in order, having a direct consumer relationship as a brand or publisher is number one. If you have identity and a direct relationship at scale, you’re going to be fine. Second is data around your ad experience—how many ads someone gets, the user experience, and managing churn to create returning traffic. Third is managing relevancy around those ads. If you get those three things right in order, you’re going to be fine. Stacy Kemp:I agree. With the proliferation of media choices, clearly articulating value against those dimensions is key. Those who rely on black-box vanity metrics will find themselves left without investment. Scott Havens:On the publisher side, the premium on proprietary data for our user base has only risen. When we strike LLM deals, we don’t provide proprietary user data; we provide access to the journalism archive. In our B2B verticals like Energy, Risk, or Factiva, that proprietary data cannot be scraped by an LLM. At Dow Jones, journalism is our bedrock, and we build data, analytics, experiences, and events around it. That proprietary data piece is skyrocketing in value. Emma O’Brian:All right, lightning round. Will, five years from now, advertisers will pay the biggest premium for…?Will Dougherty:Premium content. Emma O’Brian:Stacey, five years from now, brands that win with AI will be the ones that…?Stacy Kemp:Transform their marketing operations. Emma O’Brian:Scott, the most underappreciated asset in the AI economy? Last word. Scott Havens:Original reporting. Emma O’Brian:Stacy, Will, Scott—thank you for joining me, and thank you all!