Will AI Shake Up Advertising as the Internet Did?

Nobody yet knows whether artificial intelligence will mostly reinforce or disrupt advertising markets, much as nobody could have predicted huge shake ups in advance of Google or social media emerging. But it seems entirely possible that AI will assist some channels more than others in gaining or keeping their market share.

That already seems to be happening.

As use of third-party cookies dwindles, another shift in ad placement will follow, further boosting commitments to big retailers who possess actual first-party buyer data. And while internet app and content sites will continue to drive most of the spending, the biggest growth will be advertising on major retailer sites. 

By 2030, forecasters expect major retailers to be scooping up as much as 28 percent of all U.S. advertising spend. 

Channel/Venue	Estimated Spend (USD Billion)	Source
Internet Apps & Content Sites	520	Statista, eMarketer, Interactive Advertising Bureau (IAB)
Major Retailer Sites	290	Statista, eMarketer, Forrester
TV Broadcast Networks	35	Statista, Nielsen
Cable TV Networks	25	Statista, Nielsen
Out-of-Home Advertising	70	Statista, Out of Home Advertising Association of America (OAAA)
Radio & Podcast Advertising	30	Statista, IAB
Print Media	10	Statista, Pew Research Center
Other Channels	45	Industry reports, expert estimates

If one assumes ad revenue since 1995 was earned by different segments including print media, TV broadcasters, cable TV networks, cable TV operators, internet apps and content sites as well as major retailers, the general conclusions would be that print media commitments have shrunk, while revenues earned by the other (mostly electronic) segments have grown. 

Ad volume is dominated now by internet apps and content sites, as well as major retailer channels, which is a big change for advertisers accustomed to using formal media channels. In fact, for some of us, the growth of retailer channels is the biggest surprise. 

Media Channel	Annual Spending (USD Billion)	Percentage Share
Internet Apps & Content Sites	300	52.19%
Major Retailer Sites	120	21.04%
TV Broadcast Networks	49.2	8.55%
Cable TV Networks	32.5	5.65%
Out-of-Home Advertising	40	7.01%
Radio & Podcast Advertising	20	3.48%
Print Media	19.25	3.35%

Advertising at one time underpinned media business models. Today it underpins technology and retailer business models. By 2030, as much as 78 percent of advertising will be going to technology and major retailer firms, while media claims only about 16 percent, including outdoor media such as billboards and displays. 

All that assumes that artificial intelligence will only reinforce existing trends, and not disrupt them. In other words, if AI aids hyper-personalization, predictive targeting, dynamic content optimization or programmatic platforms, it might only reinforce existing trends. 

Likewise, voice interfaces, immersive venues or context awareness should only reinforce existing trends. 

Most observers likely see AI as driving more effectiveness for current methods (targeting, personalization, behavior-based inferences) than creating entirely new venues for placements. 

There is one major caveat. If AI somehow creates new venues, channels or platforms, we could be looking at vastly-different spending patterns in a couple of decades, much as the internet created alternative platforms and venues.

Linear Video Might be Dying, But it Still Throws Off Huge Profits, Compared to Steaming

Changing markets are very hard to track, as share in a t;raditionally-defined market often shifts to suppliers in “other” markets. Looking only at entertainment video, for example, one might see a “share pattern” that crosses industry boundaries. 

In this illustration, showing estimated time spent with video content, linear video is shown in blue, streaming video in orange, social media in green, gaming in orange. 

“Time spent” correlates with, but is not the same thing as “market share” measured by revenue. 

In the United States in 2023, for example, “video entertainment” still was led by traditional linear TV, but streaming services are displacing linear usage. Also, social media and gaming represent alternative uses of free time entertainment that arguably are substitutes and replacements for entertainment video consumption. 

Segment	Market Share (2023)
Traditional Linear Networks:	57%
New Suppliers (Streaming Platforms, On-Demand Content):	23%
Gaming:	10%
Social Media:	10%

Linear video arguably still “over indexes” as a revenue source, even if its consumption is down dramatically since 2000. 

Segment	Revenue (USD Billion)	Percentage
Traditional Linear TV:	68.8	28.4%
- Advertising:	49.2	20.3%
- Subscriptions:	19.6	8.1%
Streaming Services:	58.6	24.1%
- Subscriptions:	48.4	19.9%
- Advertising:	10.2	4.2%
Gaming:	56.2	23.2%
- Digital Games:	48.4	20.0%
- Hardware & Accessories:	7.8	3.2%
Social Media:	38.2	15.7%
- Advertising:	37.2	15.3%
- Other Revenue:	1.0	0.4%
Other Video Sources:	7.2	2.9%
Total:	241.0	100.0%

And, revenue is not the same as profit. 

Average profit margins for media companies typically range from 10- to 20 percent. Assuming a 15 percent margin on linear TV revenue of $68.8 billion, the profit would be approximately $10.32 billion.

Profit margins for streaming services vary significantly. Netflix, for example, had a profit margin of around 10 percent in 2022 while most services arguably were not yet profitable. Applying a conservative five-percent margin to streaming revenue of $58.6 billion would result in a profit of approximately $2.93 billion, all of it earned by Netflix, which in 2022 had around $5 billion in profits. In other words, except for Netflix, few, if any, streaming services were yet profitable in 2022. 

The average profit margin for video game companies is around 20 percent. If gaming revenue is $56.2 billion translates to a profit of approximately $11.24 billion.

Meta reported profit around 15 percent in 2022. Assuming a similar margin on social media video revenue of $38.2 billion, the profit would be approximately $5.73 billion.

So gaming and linear video generate 71 percent of all profits of the actual profit, with social media and video streaming contributing the balance.           

If Internet was About DTC, AI will be about Autonomy and Automation

Summed up in a single word, some might say the internet’s primary impact has been “communication.” Some might say artificial intelligence will bring “automation.”

Others might say the main economic impact of the internet has been disintermediation, the removal of steps or stages in distribution chains. “Direct to consumer” is another way of illustrating the impact. 

Outcomes	Internet	AI
Primary Impact on Value Chains	Disintermediation, personalization	Automation, autonomy
Benefits	Lower costs, increased competition, improved customer experience, agility, customer engagement, different business models	Increased efficiency, reduced costs, improved accuracy, 24/7 operation, potential shift of human activities to higher-value tasks, predictive maintenance.
Challenges	Brand awareness, data privacy, ethical considerations	Job displacement, ethical considerations, safety and security
Overall Impact	Disruption of traditional value chains,	Revolutionizing value chains
Direct to Consumer	Enabled direct communication and interaction between businesses and consumers, leading to increased transparency.	Amplifies D2C trend by providing deeper customer insights and enabling personalized marketing and sales strategies.
Personalization	Enabled personalized experiences by collecting and analyzing customer data to personalize product recommendations, pricing, content, and advertising.	Automates and optimizes the personalization process through real-time learning and adaptation, resulting in highly customized and dynamic experiences.

AI is likely to be different. As value chains already are disintermediated and built on “direct to consumer.” Instead, “automation” or “autonomy” are likely to be the distinguishing characteristics of AI impact. 

At the moment, it is hard to predict whether automation or autonomy will be the bigger trend, though both will happen. 

AI can automate repetitive tasks, leading to more-efficient, accurate operations. But AI also can operate autonomously in some use cases, learning or adapting without human intervention.

So AI might produce “efficiency” as an outcome of its automation efforts. But it might also create “effectiveness” in its autonomous operation role. Automation will be based on pre-programmed rules, while autonomy will be based on independent learning. 

So AI automation will tend to produce “efficiency” outcomes such as “do it faster” or “do it more comprehensively” or “more accurately.” 

AI autonomy might also produce “effectiveness” outcomes such as “do it a different way” based on trial and error.

T-Mobile Preparing to Use Millimeter Wave for High-Traffic Urban Areas

T-Mobile, in a test, aggregated eight channels of millimeter-wave spectrum to reach download speeds topping 4.3 Gbps without relying on low-band or mid-band spectrum to anchor the connection, the company says.  T-Mobile also aggregated four channels of mmWave spectrum on the uplink, reaching speeds above 420 Mbps.

Though T-Mobile has not relied on mmWave spectrum to support 5G, it is testing 5G mmWave for use in crowded areas such as stadiums. T-Mobile also suggests mmWave might--and likely will--support its fixed wireless home broadband services.

Verizon’s use of millimeter wave has been to support usage in dense urban areas and high-traffic locations such as stadiums, airports, business districts. In large part, Verizon has been more aggressive about using millimeter wave assets because it has had the smallest allotment of crucial mid-band spectrum. 

AT&T has the same strategy--supporting usage in dense urban areas and select business locations--but AT&T has been more cautious in deploying its mmWave assets, compared to Verizon.

Competitors criticize the sustainability of fixed wireless as a platform that will eventually be unable to keep pace with capacity demands of its users. 

At the moment, Verizon and T-Mobile are careful to offer fixed wireless home broadband in areas where they have lots of capacity on the 5G network, allowing them to devote spectrum to fixed wireless without impairing mobile experience. 

Average monthly data consumption for 5G fixed wireless ranges from about 300 gigabytes on Verizon’s network to 450 GB on T-Mobile’s fixed wireless network. 

Compared to that, average monthly data consumption in North America is about 8.6 GB. Basically, fixed wireless, when used as a home broadband platform, consumes two orders of magnitude more bandwidth than does a typical mobile phone customers.

Google Says Gemini AI is Faster than ChatGPT-4

Google argues its new Gemini large language model outperforms ChatGPT-4, using the MMMU method.