What Comes After Digital Marketing?

Digital marketing is ending because digital marketing won.

No, I don’t mean companies are going to stop advertising on Google. Websites are going to continue to exist. Emails will get sent. SEO is going to keep going. People will keep making YouTube videos and podcasts. Companies will ask people to sign up for email and creators will ask people to like, comment, and subscribe.

In 1996, the entire U.S. internet advertising industry brought in $267 million. In 2025, according to the IAB and PwC, it brought in $294.6 billion. That's more than a thousandfold increase in thirty years.

Globally, digital now accounts for about 69% of advertising spend. So no, digital marketing is not dying in the normal sense of the word. It’s just that digital marketing is so dominant that it’s just marketing at this point. Putting the word “digital” at the front is old-fashioned.

Of course, there’s also a deeper change under way too, and I think it matters much more.

For about 20 years, digital marketing worked like this: reach the right person → get a click → bring them somewhere you control → convert them → attribute the result.

That model is not gone. I use it every day. But more and more of marketing no longer fits inside it cleanly. It seems to me we’re in this strange and awkward period where the old model is clearly breaking down, but we don’t know what’s coming next.

Part of what made digital marketing so exciting in the first place was that it made marketing seem observable.

You ran an ad. Someone clicked it. They landed on your website. They bought something or filled out a form. Google Analytics dutifully put a number in a box. That is something straight out of the dreams of Claude C. Hopkins, who authored Scientific Advertising in 1923.

Marketers used to have to worry about figuring out how many people were mailing in inserts from magazines. They used to pay companies to survey people to find out which TV shows they were watching so they could back into figures around how many times their ads were being seen. Then came the Internet and all of a sudden you could target people with amazing precision.

You could go onto Google, and you still can, and have your ads show up only on very specific five-word searches. You could, and still can, target people on Instagram and Facebook by broad interest group for a $20 test. You could, and still can, run a really good A/B test for less than $1,000. Then you could measure conversion rates to the decimal point and calculate return on ad spend before lunch.

Well, at least it seemed that way. As I've written before, marketing measurement was never quite as clean as the dashboards made it look. Even in ideal conditions, attribution tells you what happened more readily than it tells you what caused something to happen. And in the real world, customers have the inconsiderate habit of using multiple devices, talking to friends, seeing things offline, forgetting where they heard about you, and generally behaving like human beings.

Still, the click gave digital marketing a center of gravity. You could organize channels around it like search, social media, email, and display ads. Your website got the clicks and the clicks showed up in your analytics.

That was the paradigm. And now the anomalies are piling up.

Thomas Kuhn used the word paradigm to describe the shared framework that tells a scientific community what problems matter and how those problems ought to be solved.

Kuhn's point was not that one weird result destroys a paradigm. Paradigms absorb weird results all the time. But it’s when the anomalies stack up one on top of the other, then you have enough to create a crisis. This happens when enough important anomalies accumulate that people begin losing confidence in the old framework's ability to explain what's happening. Only then does a rival framework have a chance to take its place. The Stanford Encyclopedia of Philosophy gives a good overview if you'd like the serious version.

Sounds a lot like marketing in 2026, right?

Take search, for example. Google increasingly answers questions on Google. Pew Research Center analyzed nearly 69,000 searches and found that people clicked a traditional search result on 8% of visits when an AI summary appeared, versus 15% when one did not. Only 1% clicked a source inside the AI summary. We’ve been headed for a zero-click world for at least 10 years now.

Now if you look into this, you’ll find that Google disputes the broader claim that AI search is gutting web traffic. It says aggregate organic click volume has remained relatively stable and that clicks from AI results are higher quality.

I largely agree with that, but I don’t think you have to resolve this entire argument to see the writing on the wall. Google itself says AI Mode can answer complex, multi-part questions by issuing multiple searches across subtopics and synthesizing the results. By 2026, AI Mode had passed one billion monthly users.

A lot of people still think of Google as “the site that gives you the links.” That’s not how it works these days. A lot of the time, it will just answer a question outright. That's why I argued recently that SEO isn't dead, but it is getting much weirder. The core job is no longer just "rank a webpage." It is making your expertise legible to whatever system your customer is using to find answers.

Advertising is changing in the same direction. Google's own documentation says Performance Max uses AI for bidding, budget optimization, audiences, creative and attribution. You supply the objective, assets, audience signals, and data. The system increasingly decides what to show, where to show it, and to whom. Dentsu estimates that 75% of ad spend will be algorithm-driven by 2028.

Oh, and that tracking that would have made Hopkins absolutely green with envy? That’s been on the way out for years too. Apple requires apps to get permission before tracking users across apps and websites through App Tracking Transparency. Firefox now isolates cookies by site through Total Cookie Protection. Chrome hasn’t fully killed off cookies yet, and even if it doesn’t do that, it might not even matter. Customer journeys routinely take place over multiple devices and platforms that have mutually incompatible stats on their dashboards.

I've written an entire post on the absolute state of conversion attribution, so I will spare you another thousand words of bellyaching on it. Let’s just say that we’re seeing less of the customer journey with each passing day.

Then generative AI showed up and made the situation even stranger.

AI is changing a lot of things right now, but it’s not the sole reason why digital marketing is getting weird. It’s just one of the more visible changes to an industry in the middle of a ton of changes.

Privacy restrictions predate ChatGPT. Algorithmic social feeds predate ChatGPT. Automated bidding predates ChatGPT. Zero-click search predates ChatGPT. Attribution was already becoming unreliable.

AI just poured gasoline on all of it. Right when the excessive tracking regime was falling apart, along came these mighty machines where people ask questions in complete privacy and receive equally private responses. There is no master list of questions that people ask ChatGPT and every response is different for every person. It is damn near untrackable.

And, of course, as you know very well by now, AI can make text, images, and video very easily. One large peer-reviewed study looked at hundreds of thousands of consumer complaints and corporate press releases, more than 300 million job postings, and thousands of United Nations press releases. By late 2024, the researchers estimated that LLM-assisted writing appeared in roughly 24% of corporate press releases, 18% of financial consumer complaints, nearly 10% of small-company job postings, and 14% of UN press releases.

Talk is cheap, in other words. Incredibly so. You can make competent-looking text, in great volume, in moments, with no training, on a free AI model.

But just because you can make landing pages and email flows way faster, it doesn’t mean marketing is worthless. Rather, it changes what’s scarce.

Now I need to put a giant warning label on the next section.

This is my best guess.

The old king is gone, the new king hasn’t been crowned. And we don’t know when the new king will be crowned. Welcome to the interregnum.

The old digital marketing framework is plainly struggling to describe reality. But it’s anyone’s guess what will replace it. Anyone who says they have The Definitive 2027 Marketing Playbook is probably selling it for $997 with a countdown timer. Yuck.

But if I had to name the direction of travel, I would call it marketing under algorithmic mediation.

For most of the digital era, marketers used machines as tools to reach people. These days, machines stand between the marketer and the person. Search engines are summarizing what we write. Social media feeds are determining which of our messages reach people. Ad platforms are choosing audiences and placements for us. Recommendation systems are ranking products, businesses, and creators and we have to try to figure out how and why they do what they do. Inscrutable.

People are making serious B2B decisions after asking ChatGPT and Claude to synthesize data from dozens of different sources. A customer can read reams of pages about your company and show up with a complete understanding (accurate or inaccurate) of who you are. And all you see on their intake form is source of Google, keyword of [insert brand name here].

That changes what marketers need to be good at. And, weirdly enough, I think it brings old ideas back into relevance.

When something becomes abundant, abundance stops being impressive.

Twenty years ago, having a polished website was evidence that a company had its act together. But that’s not impressive now. You can do that on Squarespace in maybe two hours.

Ten years ago, publishing 100 useful articles represented a substantial investment of time and expertise. I could probably have ChatGPT do that for me before I finished writing this one with my dumb little meat hands on this keyboard with the space bar so well loved it’s polished to a shine.

So I increasingly think the important divide is not digital versus offline. It's high-trust versus low-trust.

If a colleague I like recommends a company, I’ll probably listen to their recommendation because they’re high-trust. If I see a customer review, I’ll trust it more than your home page. I’ll trust an industry publication more than you calling yourself an industry leader.

To impress people in the years to come, you need things that don’t easily scale like conference talks and podcast appearances. You need unique contributions like original data sets and hard numbers proving your capabilities like detailed case studies with real numbers. You need to show your knowledge by doing things like creating a knowledgeable YouTube channel.

This is one reason I have become much more interested in old-school marketing techniques. Not because I think we should all abandon Google Ads and start mailing catalogs. No, I pay attention to the past of marketing because a lot of older marketing operated in a world where attribution was lousy and reputation mattered enormously.

We're getting some of that world back. And let me back that up by giving you some more data from buyer research. In 6sense's 2025 study of nearly 4,000 B2B buyers, the winning vendor was already on the buyer's Day One shortlist 95% of the time. Buyers initiated 79% of first contacts. Eighty-five percent had prior experience with the winning vendor, and 75% personally knew sellers at vendors they were evaluating.

Meanwhile, Edelman's 2026 global brand study found that 88% of consumers considered trust an important or critical purchase criterion. Trust basically tied with quality and value.

I’m not saying we’re about to go into a world where back-slapping at trade shows is the future of marketing. But I am saying that marketing tactics that depend on low-trust interactions are going to be a lot harder to pull off when you can’t guide them with halfway reliable data.

Here are three bets I’m making based on my understanding of the bizarro macro environment we’re in.

Anyone can say they’re good. Anyone can ask AI to write 2,500 words about how good they are. So you need evidence from other sources proving that you are as good as you say you are.

That comes in many forms, including reviews, mentions, links, interviews, partnerships, customer stories, building public expertise, and publishing original research. None of this is new, by the way.

People need to know your work and they need to know you’re worth reaching out to. It helps with the trust problem on the human side. And it makes you more legible to the machines that determine whether people will see you.

If your company says one thing about itself, that's a claim. If customers, publications, directories, partners, and other credible sources repeatedly say the same thing, that starts looking like a fact.

As third-party visibility gets worse, your own data gets more valuable. You need to know who your leads are and which ones are good. You need to know what they buy and how much they’re worth. And it’s worth asking them outright how they found you.

You need to connect marketing activity to actual sales outcomes whenever you can. And then accept that there will be gaps.

I've argued before that attribution is useful without being omniscient. Use attribution. Measure lift where you can. Talk to customers. Compare regions and time periods. Look for multiple lines of evidence.

Your dashboard is a map. But don’t look at the map so close that you forget to look at the road.

I think this is going to be one of the defining marketing skills of the next decade. You have to make it easy for machines to answer basic questions about you correctly.

I’m talking very basic things:

This is not about discovering one weird schema trick that forces ChatGPT to recommend your plumbing company. I frankly think the machines are way too smart for that these days anyway. You might be able to fool them for like a week.

I’m talking about building a coherent information environment so the machines know what to say about you. So they tell the most flattering version of the truth. As I wrote in What SEO Can't Do, search optimization has always been partly a translation problem: taking real expertise and making it legible to the systems people use to find it. There are just more systems.

This isn’t Les Mis and we’re not getting a clear-cut revolution. SEO still works. Google Ads still work. And the same is true of emails and websites and content and analytics.

Einstein figured out relativity, and people are still using Newton’s physics. The existence of a new paradigm does not require every tool from the previous one to burst into flames.

Likewise, a good search campaign does not stop producing leads because I've decided to spend an afternoon talking about epistemology on a marketing blog.

The change is at a higher level. The click-and-attribute model is becoming less capable of explaining the whole marketing system. And that’s because we increasingly operate in a world where machines choose what gets seen, summarize what gets known, obscure parts of the journey, and make competent execution cheap.

So I don't think the answer is to become less digital. I think the answer is to become less attached to digital marketing as a self-contained category.

Brand, search, PR, events, partnerships, customer experience, content, paid media, sales, and measurement are not staying in their lanes anymore. Any borders that existed between these types of marketing are becoming fuzzier and fuzzier.

Which brings me back to the original question. What comes after digital marketing?

I don't know yet.

But I have a reasonably high degree of confidence about which bets survive most plausible futures.

Be known. Be corroborated. Know your customers. Measure real business outcomes. And make the truth about your business unusually easy for both humans and machines to find.

Digital marketing isn't disappearing. It's becoming too ubiquitous, too automated, and too entangled with everything else to remain a coherent world unto itself.

What comes after digital marketing may simply be marketing again.

My company helps B2B service businesses generate qualified leads through data-driven SEO. We do the work and we build the tracking to show you what's producing results.

If you're interested, book 30 minutes of my time and we can talk about whether it makes sense for your business.