The Tech Oligopoly Is Over. We have a generational chance to not let AI become the next one in media.

We’ve danced this dance before, haven’t we?

Over-reliance on Google and Facebook cost content and news publishers dearly. Since Google's ad server embedded themselves in to most publishers and Facebook’s Instant articles became the hero of the technology stack, the whole industry hitched it’s fortunes to the big tech platforms and chased traffic and ad dollars on terms set by Silicon Valley. We learned the hard way that building a strategy around someone else’s ecosystem is fragile as hell.

By the time we’d realized that the long game had ceded ground to a quick Google-infused crack injection to the arm, it was too late. And in 2025 it’s already too late to sanction Google; history is about to repeat itself and so we should probably think ahead instead of spending all our time thinking behind.

Whether it's the various lock screen advertising businesses that went bust literally overnight due to an API change by a tech giant or a single algorithm tweak being capable of changing a publishing business’ audience and income overnight, I see an uneasy similarity in how AI is entering our industry - and I wonder if it will mean the same happens again?

If it does it'll be far worse.

But we've already learnt our lesson right? Surrendering control of content distribution and monetization to an oligopoly is a dead end for sustainable growth.

The new shiny lights

AI giants offer new scale but threaten to become new gatekeepers. With gen AI on the rise making the segmentation of site visitors a lot easier , it’s easy to see parallels to the old platform era. A new wave of businesses promising fast growth during a stagnant economic moment (but secretly mining your data)? Couldn't imagine it.

Companies like OpenAI (who need to claw back some money to pay the ridiculous debt they are riddled with through investment) and Perplexity (who will probably be in the same boat shortly) have already made their intentions pretty clear in the media space.

They’re not even here yet, but I can hear the ogre's footsteps thundering. They'll come harder for the open web than anything we've seen to date, by more than 10x.

AI businesses will - in the next two years - assemble and aggregate the open web at a rate that makes Google’s efforts look tame.

The offer on the table to a content publisher without email/paywalls is tempting; immediate ease and traffic is definitely the safe instant option for an executive to take, so I’m not trying to suggest nobody use AI businesses. I’m just saying that history repeats and pendulums swing back and forth, and it’s happening again. The reality so far actually is worse. Early data shows that AI-powered search engines, which scrape and summarize content, send 96% less referral traffic to news sites than traditional Google search.​ Puts Google's villain status in to perspective?

Don’t outsource innovation (again)

Outsourcing innovation in monetization to external platforms is a risky shortcut, and one that I don’t think large content publishers will survive if it happens again. 

If we allow AI platforms built on media metadata to be the sole source of new ad targeting or content recommendation innovations, we’ll be ceding the future of business to them. Consider how much of today’s AI is fueled by publisher data: our articles, keywords, and user engagement patterns are the training fodder for these models. Now consider that in 2025 ads are everywhere, not just news like previous generations.

If an AI company develops an algorithm from that data which develops dominance and we simply buy it as an Agent As A Service (AaaS haha), where is anyone else's competitive edge in this equation? (I speak as a PUBLISHER because I have been one and am passionate that the area thrives). This is the threat that publishing content comes with these days.

Every competitor can purchase the same tool – meaning no one stands out, except the AI provider who now sits between us and our revenue. Instead, content businesses should invest in owning a permanently defensible part of the innovation cycle - first party data. This is the opportunity that publishing content comes with these days.

Defensibility comes from what you do that others cannot. For defensibility of online publishing businesses, first party data comes in just behind the no.1 priority: having a clear publishing voice and brand that your community loves. AI can absolutely help amplify these strengths – but it can’t be the strength on its own, especially not if it’s the same AI everyone else has access to.

Don’t be seduced by vendor pitches that claim their AI solution will instantly solve your revenue challenges. And honestly, we’re nowhere near being able to claim that AI can produce better outputs than data does. I’m hearing this being said right now in real time, and honestly? I believe these “AI performs 30% better than data segments” claims are either misinformation or misreporting. After all, 87% of stats are made up right?

The reason first party data hasn’t thrived thus far is not because first party data is dead - as some AI startups are beginning to suggest. I don’t believe it’s because CDPs and DMPs overpromised anything. In my view, it’s because the idea of accounting for your business’ customers with digital data tokens that have you know them better (the nirvana) has generally always been cobbled together in bit parts, arbitraged by middlemen before it could be used properly by the owner, and then bastardized by adtech.

Programmatic adtech has been data’s worst enemy instead of it’s best friend, campaign measurement has been done by the parts of the ecosystem that either cannot or should not be doing measurement, and the tech hasn’t been there till now to have durable first party data (without having your customer’s email addresses to hand). None of these things are true/will be true in the future.

So I get it - we all have PTSD from being underwhelmed so far by data driven products in our industry. But just because it’s been hard thus far, does not mean we give up on ‘business owners being the only ones to own their first party data’ being the right/ethical/compliant thing to do.

Nothing good comes easy, but in 2025 even deterministic 1P data tokens are now able to be created > deployed > used for known and unknown visitors - ethically and compliantly. The world beyond 2025 is different, and will allow first party data to be a huge success for the owners of it.

First-party data: real, accurate foundations in an AI era

AI doesn’t make first-party data obsolete – it makes it more critical than ever. A dangerous misconception floating around is that in the age of ‘machine learning’, the machines just… learn? Right? Wrong. It’s untrue that you no longer need to cultivate your own audience data. The easy to tell (and sell) narrative follows the shiny lights: “If AI can predict user behavior at such scale, why invest in gathering data directly from users?” Without getting philosophical and thinking about a world where AI puts humans in to cohorts and decided what content each cohort sees, let me assure you: the opposite is true.

To all content owners: the deterministic insights you have about your audience – gleaned from subscriptions, site registrations, content preferences, purchase history, and other first-party interactions – are gold in the AI era. They are precisely the assets that no one else has in the exact combination you do, which means models trained or tuned on your data can deliver experiences and results that generic models cannot. First-party data is what will allow your AI-driven initiatives to actually reflect your audience and your community - not the same vanilla predictive probabilistic outputs that will be impossible to recognize in a lineup.

Moreover, brands are increasingly seeking out publishers with robust first-party data to participate with as third-party data connections fall away. Advertisers won’t be impressed by your AI tool in a year, they’ll be impressed by your accuracy and uniqueness as a business partner through data connectivity, which comes from the foundations laid by first party data creation - and maybe the diligence & accuracy your fancy new AI agents have as a result of it.

Your hard-won data (your customers) isn’t just a defensive asset, it’s a revenue driver when used smartly (and ethically). Conversely, if you neglect first-party data, you’ll find yourself at the mercy of those who do aggregate open internet data (your customers) – likely the tech or AI giants – effectively recreating the dependency we’re trying to escape.

The next few years of AI will thrive on good quality human (non synthetic) data, and if it’s not your data, then it will be the data generated by Open AI or Perplexity by mining your content. Yes you'll need AI and tech partners but please for the love of our industry please let's spread bet this time 🙏 

In short, treat first-party data as the bedrock of your future monetization and your best defence against generic AI. It’s the element that makes your use of AI proprietary and your business resilient to what's about to hit this industry.

The era of being captive to an oligopoly is over.