The future of data in media publishing: fewer processed data-sausages, and more home grown produce.

_{Spoiler: Scroll to the bottom for my 2025 Publisher Monetization Primer, subscribe and I'll send you the whole thing!}

Early in my career, my boss once asked me when I kept asking how our data product was made: “Do you really want to know”? It took me years to realize that he didn’t really know the answer himself, nor did he feel compelled to. Over the recent years it’s become clear just how hard it is to answer that question when it comes to 3rd party data as used in adtech.

We’ve proudly proclaimed for many years that data is oil, but when you look at what data we were mostly referring to in adtech, it wasn’t oil - it was more like processed sausages.

After 17 years in the industry, I can’t honestly tell you that data targeting is what drives actual business results in the open web adtech game, but again - maybe it's because we’ve largely been feeding each other nasty ingredients. The culprit behind it all - “third-party cookies” - are still lying around in the sun sadly, to keep Google’s legacy adtech business alive until they decide to kill it.

You hate them as a consumer too. Getting hammered with flight ads when you just bought a flight? Got targeted by a fertility clinic insensitively? Being politically targeted? Getting over charged when you return to a site to purchase? Or just having a shit website loading experience? Bingo.

To give another analogy, 10 years ago I explained to my then-team that if I tried to get past border security in any country with a photocopy of my passport, I’d get deported faster than you could blink. Photocopies of site visitation data are the 3P data; a mirage of what you thought you were getting. Not the real copy.

“3rd-party data is mostly really 4th party data anyway” 😆

3rd party data by definition means it’s rented, photocopied (legally or illegally but usually without user consent), or found lying around on the floor. The airtag stuck on your back when you’re not looking so someone can track you without you approving of it. All of those things fly in the face of what we should do next in adtech or in data. Or publishing. Or just moral code.

We’re all sitting here spending way too many brain cells villifying Google for being the Big Bad while our industry continues to feed on these sausages, and excitedly flock towards the new sausage factory - AI. Google will likely quietly exit adtech and keep inventory monetization on YouTube running. Long term I wouldn't want to have a bar of open web adtech if I was them; the complexity, bad press and margins don't add up. The thing Google are guilty of isn’t keeping 3P cookies going - it’s not having killed them when Safari did. So they’ll now probably move in to enterprise AI, productivity and cloud - whilst Big AI businesses like Open AI will swoop in to alter the open web forever in their image.

It’s happening again, but we have a moment to take control. Let’s use it.

It’s time to let go of old ingredients and take a look in the mirror and build a better industry - otherwise it'll happen again. This starts with suppliers (digital asset owners) setting the precedent.

My goal is to advise businesses with customers visiting their digital properties that it’s time to build data assets, not rent them. I’ve worked for publishers, adtech, martech and BigTech - and if I ran a publisher today I’d start by generating first party assets I actually own. No lease back agreements of easy promises like ID graphs, no bulk buying mystery meat segments cooked on the floor. I’d control the source, control the output, and guess what? I imagine I’d realize that user privacy mysteriously gets respected along the way, and that it does work for what I want to use it for.

I’ve been turning data in to monetizable enterprise products for many years and assumed I was always helping to improve things… so my catharsis (and occasionally guilt) is real, and I see the next paradigm glowing on the horizon. We’re all ready to rebuild, and we’ve all learnt what not to do. We can do it better this time, right? I’m optimistic.

I chat with enterprise customers daily, and here is a (hopefully helpful) round up of the main shifting sands from the last few days:

👎 Google are keeping 3rd party cookies alive, their AI Overviews are gobbling up 34.5% of publisher clicks and the only ad-stack unifier publishers currently trust with their entire authenticated + unauthenticated audience (Google Ad Manager) may still get carved up by the DOJ.

🛍️ OpenAI is cueing up one-click commerce that leapfrogs websites entirely while lobbying for zero IP restrictions. An admission of intent to plagiarise - sounds familar?

🏠 Tariffs are on pause, and publishing data teams around the world are using the moment to rebuild their house.

👮🏻‍♂️ ID graphs are in the legal crosshairs (for being data sausage factories), and America’s flirting with a national privacy law.

🤝 Media agencies are morphing into data-collab engines – which only works if suppliers turn up with michelin-grade ingredients for the cook-off.

🤖 Agents are everywhere, but they’re like untrained graduates ^{(click the link)} right now, they need training (which is good data + your instruction + guardrails). Deploying AI agents to connect to your customers without any proprietary/unique/ethical private customer signals is a risk - like hiring a savant who has never met your customers: impressive vocabulary, but might just say or do weird things. Agents risk becoming your most annoying employee with a megaphone if mis-implemented. In particular, for a publisher implementing retention and personalisation agents, 1P data is the non-negotiable core.

All up, I feel like now is the time for all businesses to stand on their own feet when it comes to data. Build, not rent.

I am sure you can see the pattern emerging above; every dependency brings with it a potential knife to the throat down the line, and we don’t know enough about what the world will look like next year to go all in on one monetization strategy alone - and certainly not one that relies on data being scraped by another business. We did that already and it didn’t go well.

So, the only thing I genuinely feel strongly about right now is that as a business owner, you should build your own house as much as possible.

Third-party anything now carries a whiff of PTSD, and that includes the main ingredient in the meal - data. The headlines may be about agents and AI, but if businesses want to build unique data driven monetization tools I would not recommend you let AI gobble up your content metadata - and as a result lose most of your long term USP and defensibility. This is particularly important in an AI world where your content and metadata is already AI food - and AI LLMs are already using it to go straight to sell products to YOUR customers.

What about ID free targeting, I hear you say? ID free is cool.

It is - but it’s not the same thing at all. For me it’s like saying “I prefer mince to a steak”. In reality, I wouldn’t sub out one for the other unless what I wanted wasn’t on the menu. At it’s simplest, ID free ad targeting is predictive (and usually contextual but sometimes cohort based) and barely remembers things.

ID free, cohort and contextual targeting offerings are built for ad buyers (and sold to publishers when they don’t have 1st party data). These products are useful when they're the right solution, but if we hadn’t clung on to stale sausages and then panicked when the rug was pulled, those products probably wouldn’t be as trendy right now.

Such solutions usually exist to help ad buyers chase the next cohort of customers to sell to - but don’t really cater as much to the user’s relationship with the publisher/business that the customer is loyal to. They might in time do that with the speed of AI, but they might not - like I said we’ve seen a similar story play out before.

So in summary predictive, cohort based or contextual ID free monetization capabilites are a solution - just as AI agents are - but not instead of knowing your customers (with consent and privacy and without needing their email addresses).

This idea goes beyond ad targeting now and becomes about fostering a customer relationship between supplier and buyer. Middlemen sometimes forget that they exist to serve the supplier and buyer and get wrapped up in their own agendas.

AI is likely to sell people more stuff, but 1st party data is the trustworthy critical ingredient keeping your customer close to your business. AI will help you interact with your customer, but 1st party data is what enables you to talk to the person you want to with respect.

I don’t think it’s possible to overstate how important that is when a) loads more individuals will be publishers in 5 years and b) everyone will be able to start a brand indistinguishable from the rest without a loyal fan base. The one safe key you can trust in a world of homogeneous brands is your relationship with YOUR customer. That’s literally first party data.

Let’s go.

I think this is a huge moment. And a hugely exciting one. I’m personally keen to help digital asset owners to build a private, respectful, deterministic data layer that recognises every visitor without begging for their email address or outsourcing the task to the Yellow Pages of data products.

It feels like we are playing things by the day right now (as you can see by the amount of seismic changes that occur every week right now) - but one idea is reliable, and that’s owning 1st party data to anchor the tech stack rebuilds being conducted right now by all of the established media publishing houses.

The goal is to respectfully and ethically tokenize authenticated and unauthenticated customers, own the keys, secure the data, and collaborate with emerging data technologies (clean rooms).

That’s when data becomes oil, and it’s why Big Tech does so well. It’s the strategic plan we all want for a future world of publisher/business independence.

Building 1P data is not even just a better way to build ad products. In a world where most businesses have an online footprint, it’s a key part of the puzzle to build equitable economic growth for businesses of all shapes and sizes in the years to come.

Two years from now, when OpenAI pulls the same stunt Google just perfected, I’d much rather be the business who took as much control as possible, not the one who banked the easy money or assumed AI would make me squillions on its own. Thankfully I can currently see the biggest publishers rebuilding their own technology stacks (the rise of PubTech). In this world, a 1st party data asset is a non negotiable to anchor it, reduce all the errors and admin that usually ensues down the line with rented/fragmented data, and make it a worthwhile business investment.

Let’s label it the bulletproof diet: 100% first-party protein, zero fast food data-fillers, served with a side of Big AI and agent-driven activation that actually respects the farmer: the end user.

_______________________________________

TL;DR: I advise digital businesses on agent, data and tech-led monetization strategy - mainly those who want anonymous, deterministic, privacy-forward data that returns revenue and yield via addressability and defensibility in to the future. I love writing, expressing ideas and I’m getting in to it again. I hope the above was a helpful perspective!

I’m happy to share you my research and frameworks (you can see below for one of the pages), subscribe if you want access and I’ll send it through. And bring your thoughts or experience – I want to learn too, and help cook up something better and more ethically sourced than the data sausages we’ve been feeding each other for the last 20 years.