Digital advertising is set for another massive disruption wave

I experienced the mid 2010s digital advertising upheaval first hand. As The New Yorker's Ken Auletta described it, "The Math Men Overthrew the Mad Men." As one of those Math Men, personally spending $10s of millions without talking to a single human at an ad agency, I hardly realized just how big of a tectonic shift I was amidst.

And, 10 years later, the next wave of upheaval is upon us. Here are 4 megatrends underway:

#1: the duopoly welcomes other advertising channels to the party as “everything becomes an ad network”.

Anyone who had a large audience (usually because they sold products or services that a lot of people wanted) realized that their customers were also ad inventory. With AdTech tools (targeting, auction mechanics, etc) largely commoditized, anyone with an audience could now become an ad platform.

Retail Media Networks (‘RMNs’) are on the rise:

What started as a few sponsored ads on Amazon shocked the industry in February 2022, when the company disclosed $31 billion in 2021 ad revenue for the first time ever. Walmart, Target, Kroger, and just about every other major brand quickly followed suit.

Fast forward a couple years; Activate predicts that Retail Media will top $100B in ad revenue by 2028. While this is still smaller than Google’s Search business ($175B), alone, it is larger than all of TV advertising ($87 billion).

The current duopoly has proven resilient – and signs do not point towards a fragmentation of the ecosystem. For all of the competition – there are over 200 RMNs in the US – Amazon still controls over 75% of that retail media ad spend.

And it’s not just RMNs:

AppLovin, whose stock price has increased 759% YTD at the time of this writing, is another big winner in this trend. Few outside of the industry have ever heard of the company, yet its market cap has grown to over $100 billion.

What does AppLovin do? At the most basic level, AppLovin has a network of mobile games. Those mobile games have many users. AppLovin sells advertising space against those users, primarily to other mobile games who are looking to acquire users. This feedback loop works well: generally, if you are a user of one mobile game, you are open to downloading another mobile game.

So what’s the big deal? Recently, AppLovin launched a beta program opening its ad network to non-gaming brands (think: your favorite DTC detail or consumer products brand). And it seems to be working. While it’s early days, and there are plenty of questions of whether it will scale, early results are promising. Eric Seufert does an incredible job laying out the bull & bear case here.

Note: there is some skepticism in the industry about AppLovin’s incremental ROAS — but I’ll leave that for another post!

#2: last touch attribution gives way to more holistic forms of measurement — and people wake up to some inconvenient truths.

I wrote about this subject in more detail a few months ago in “The CMO's next big battle: over-attribution”

“Last touch attribution” gives credit to the the advertising platform where are user most recently saw an ad immediately preceding their purchase of that given product. This has been the dominant mode of measurement for digital brands for the past decade.

Why? In contrast to the methods that preceded it (more or less “finger in the air”), it was a revolutionary increase in accuracy for poor marketers trying to measure ROAS accurately. Make no mistake: it really was revolutionary. You could actually ~deterministically see that a user saw an ad and then, shortly thereafter, made a purchase.

However, recently, last touch attribution has come under fire. The weakness of this approach is that there are often many contributing factors that lead to a sale. Crediting the most recent ad impression is often limiting.

It also has another side effect: advertising channels that are more likely to be the last touch (e.g. Google: searching right before you purchase) will be over-credited in this model. And less direct advertising (think: billboards) will be under-credited. The net effect: for a decade, a wave of marketers (myself included) under-appreciated the magic of “brand” advertising that was harder to measure though still effective. 

As brands have matured (the DTC revolution is well into its second decade at this point); and innovative measurement firms, such as Measured and Haus, have democratized access to more holistic ROAS measurement; more marketers have adopted a broader point of view.

The consequence: while still dominant, the dominant advertising channels of the past decade are under a higher level of scrutiny from sophisticated marketing analysts than ever before. To their credit, their marketing analytics teams are furiously working to build out a new wave of measurement solutions in this light.

For further reading on the subject of marketing channel effectiveness and how attribution can ‘trick’ marketers into prioritizing one source or another, I recommend two incredible reads:

• How Brands Grow

• The Long and the Short of it: Balancing Short and Long-Term Marketing Strategies

#3: privacy regulations lead the industry to scramble for creative workarounds.

In the face of both external (government: CCPA, GDPR, ATT) and internal (employee pressure, whistleblowers) pressures, many of these large ad platforms have also dialed back their deterministic targeting capabilities.

As one illustrative example, Facebook no longer offers the ability to target third-party audiences from the likes of Acxiom, Epsilon, and Experian within their ad platform. Without arguing the ethics for or against this type of ad targeting (arguments do exist on both sides!) I will simply make the claim that it is impossible to put the toothpaste back in the tube. Once marketers were given the ability to reach specific audiences, they would never tolerate those capabilities being stripped from them.

The result:

• The rise of zero-party data: brands prioritizing collecting voluntarily shared data from consumers, at the expense of other KPIs, in order to personalize marketing & products for them.

• Data clean rooms (e.g. Snowflake, LiveRamp, Infosum): which allow brands to comingle their zero-party data with other brands, thus forming a data co-op to mimic big data.

• Data marketplaces (e.g. Snowflake Marketplace) bringing all sorts of third-party data sources, available for connection, in a privacy-centric way. These data sources answer questions ranging from what else do my customers buy (Affinity, Yodlee) to what is their job title (ZoomInfo, Clearbit) and did they just visit a certain retail store (Safegraph, Foursquare).

These trends, together, represent marketers’ collective attempt to rebuild the AdTech infrastructure of the 2010s in a privacy-centric way.

#4: and, of course, AI.

Disruption wave #1 (2010-2025): The ability to microtarget dozens of audiences to find the best ROAS dramatically dropped the barrier to entry for new marketing-led brands to thrive. The catch: given highly produced creatives were still expensive, this disruption was limited to social platforms that “accepted” low-brow creative.

Disruption wave #2 (2025-???): The ability to use AI to generate highly produced creatives for “free” will supercharge the prior disruption wave for all advertising channels, including the ones that require high-brow creative such as TV ads, film trailers, and so on. Herein lies the next wave of great consumer brands.

All this sounds a bit exhausting — but the game never stops 🥵. Personally, I’m beyond excited to observe – and participate in – this next wave.

*Many figures came from ChatGPT and Perplexity. You’ll have to forgive them for any hallucinations ;)