Good afternoon,
The Medium identifies a few key trends each fiscal quarter that reveal the most important tensions and seismic shifts in the media marketplace. The key trends help you answer a simple question: "What's next for media, and where's it all going? How are the pieces lining up for business models to evolve, succeed, or fail?"
Read the three key trends The Medium will be focused on in Q3 2023. This essay focuses on "Artificial intelligence (AI) and cloud computing applications and services are increasingly dictating content consumption."
Amazon’s Q2 2023 earnings report last week generated a lot of buzz in large part because its advertising earnings are up 22% year-over-year ($10.7 billion in Q2 2023 up from $8.8 billion in Q2 2022). Meanwhile, Google’s advertising revenues were up 3.2% ($58.1 billion up from $56.3 billion) and Meta’s were up 12% ($31.5 billion up from $28.2 billion).
On the earnings call, Amazon executives pointed to machine learning (ML) and artificial intelligence (AI) as the difference makers: “Our teams worked to increase the relevancy of the ads we show to our customers by leveraging machine learning and improve our ability to measure the return on advertising spend for brands. Third-party unit mix increased to 60% during the quarter, the highest level we've ever seen, and we're continuing to see good growth in the number of sellers and the unit sold per seller.”
And that’s the other reason it generated buzz: AI has been the big buzzword on earnings calls: there were 25 mentions on the Amazon call, 65 mentions on the Alphabet earnings call and 54 mentions on the Meta earnings call. It is not clear if Wall Street simply has identified total mentions as a key metric for estimating future business success, or whether they actually believe it is a game-changer.
I have had a hard time finding examples of the key trend “Artificial intelligence (AI) and cloud computing applications and services are increasingly dictating content consumption.” Not because they don’t exist. But because they tend to focus on the wonkier elements of walled gardens (like this good essay — “An internet of silos” — from The Martech Weekly’s Juan Mendoza) or go deeper into how “the possibilities of activating data have increased manifold” because better data tools and the adoption of the cloud data warehouse.
Amazon offered an unusually helpful example. And it’s worth briefly diving into the logic of what we learned.
Key Takeaway
Because Amazon's earnings call treated AI as more than a buzzword, we now have a minimal window into Amazon's role in shaping the future of advertising and media.
Total words: 900
Total time reading: 4 minutes
Wins with ML
In its letter to shareholders, Amazon offered multiple examples of how it is “inventing on the behalf of customers”, and it included this one related to Amazon Advertising:
“Introduced new, more advanced machine learning models to help advertisers reach previously unaddressable audiences with Amazon Ads. The new machine learning models analyze a range of signals to help advertisers predict and reach highly relevant audiences with optimal cost-efficiency, which is critical to reaching desired audiences as the advertising industry moves away from third-party cookies. Performance improvements include an increase in click- through rate, an increase in return on ad spend, and a decrease in cost per impression.”
It’s a compelling sales pitch. Amazon has nearly 200 million monthly active users, and until recently, it had been underdelivering for both customers and ad buyers. Customers were not seeing relevant ads, and ad buyers were neither reaching desired audiences nor delivering cost-efficient campaigns.
Enter machine learning, and Amazon’s first party data advantage is not just scale, but also rapid evolution towards client needs.
Amazon is The Medium
Amazon CEO Andy Jassy added on the call:
“...while we will build a number of these applications ourselves, most will be built by other companies, and we're optimistic that the largest number of these will be built on AWS. Remember, the core of AI is data. People want to bring generative AI models to the data, not the other way around.
AWS not only has the broadest array of storage, database, analytics and data management services for customers, it also has more customers and data store than anybody else.”
In one way, this is a narrow statement about Amazon’s dominance. Amazon will account for $23.95 billion of the total $29.69 billion US retail media search ad spend market in 2023, per eMarketer.
Moreover, in a world shifting away from third-party cookies, retail media “may be the best/one of the only ways to target people at all”, as Mike Shields wrote after visiting this year’s Cannes. So the narrow interpretation of Jassy’s comments and the quarterly earnings is that Amazon is finding growth through retail media search.
Through the broader lens of The Medium, this might be the most significant statement of Q2 if not all of 2023. Why?
As I wrote in May’s “Consumer Data May Be Too Complex for Media's DTC Models”, AI is becoming increasingly important to dictating which marketing content we will see and/or consume depending on the data a media company has on us. The view of any individual or cohort of consumers by a singular AI program is “now truly holistic, and the decisions that are made about serving us content from data within a warehouse may be made without human insight or guidance.”
The biggest challenge of this has been where to look and why, and Amazon's transparency — albeit minimal — is a welcome solution to that challenge.
AWS has more than 1 million customers. More than 150 advertisers and marketers rely on AWS for cloud storage. Their demand for Amazon's services is growing faster than it is for the likes of Google, even with YouTube's nearly 3 billion monthly active. And ML is rapidly improving the performance of Amazon's services for all of them.
Jassy's point was that Amazon is not some monolith shaping the future of AI and ML in the cloud. Rather, its services in the cloud are increasingly meeting what its customers need. Its supply of AI and data services (especially the latter) is meeting growing demand, and therefore it is quickly beginning to shape marketing content we will see and/or consume.
Because Amazon's earnings call treated AI as more than a buzzword, we now have a minimal window into Amazon's growing role in shaping the future of advertising and media.
Must-Read Monday AM Articles
The demand for “premium content” is being redefined by creators, tech companies and 10 million emerging advertisers.
* In a lawsuit filed in the Southern District of New York, Donaldson’s team claims that Virtual Dining Concepts harmed the integrity of the MrBeast brand by prioritizing growth over customer satisfaction.
* With the new ad tier gaining traction slowly, Netflix has had preliminary discussions to sell ads through other partners, in addition to Microsoft, people familiar with the matter said. Netflix is reworking its pact with Microsoft to reduce the revenue guarantee.
* As networks are largely without new scripted episodes heading into the fall amid Hollywood’s dual strikes, the state of unscripted programming is suddenly in a brighter spotlight than ever before. But this moment of scrutiny comes as those who work in the genre continue to deal with an uncommonly depressed job market, not to mention challenging expectations for the coveted openings.
* Top executives at several advertising agencies told The Information they’ll commit at least 10% to 20% more with YouTube this fall than they did last year. Ad spending commitments to traditional TV networks, meanwhile, are down 15%, multiple senior ad-buying executives said. ($ - paywalled)
* Netflix’s data is used to inform decisions about what shows and movies to produce, whether to renew them, and whether to share them with any given viewer through the company’s famous recommendation algorithms. ($ - paywalled)
* Since the streaming era, movies and television feel less special, labor conditions have plummeted, and turbulent mergers and layoffs call into question which legendary institutions will still stand in another ten or twenty years.
* YouTube released six new ways to create on YouTube Shorts, including livestreaming
AI & cloud computing applications and services are increasingly dictating content consumption
* For growth-stage organizations, using a data warehouse in conjunction with a Consumer Data Platform – either packaged or composable – is key.
* If advertisers are still clinging to their data management platforms, they are being left behind.
* While the rise of on-demand streaming has made DVR less relevant, in a way it’s more important than ever. ATSC 3.0 DRM threatens to take away user control it ensures that users will be funneled into apps in which the content providers call the shots—even for free, over-the-air TV.
* Limits on data would challenge how easily AI companies can build future versions of their language models. But the sheer size of these models also is a challenge for those seeking copyright protection, lawyers say. ($ - paywalled)
* Adobe Chief Strategy Officer Scott Belsky is betting that other generative AI models will eventually have to reckon with severe copyright restrictions, at which point many will go the way of Napster.
Legacy media companies are throwing in the towel on their bets to own the consumer relationship in streaming and beyond.
* Vox Media, the parent company to websites such as New York Magazine, Eater and SB Nation, will no longer use Chorus — its proprietary content management system — to power its own websites, sources told Axios.
* Director Stephen Soderbergh argues the need for data transparency in Hollywood
* Some national brands have expressed frustration for the lack of data on how well their campaigns are performing with Apple TV's MLS broadcasts.
Other
* Failing to address the concerns of WGA and SAG-AFTRA reinforces the perceived disposability of creative professionals, further alienating them from joining and staying in this industry and inevitably resulting in a decline in quality content and diminishing returns for advertisers.
* While it’s certainly open to interpretation, both GDPR and the California privacy law can be read as requiring that users provide consent for the use of IP addresses. That’s a big challenge, since many consumers don’t have a direct relationship with their TV manufacturers, let alone the comScores and Nielsens and vast ad tech ecosystem that may use or touch IP addresses.