One of the best books about Hollywood is “Adventures in the Screen Trade”, in which the screenwriter William Goldman writes: “Not one person in the entire motion picture field knows for a certainty what’s going to work. Every time out it’s a guess and, if you’re lucky, an educated one.”
He was referring to the fact that despite their market research, gut instinct and experience, no one in Hollywood has any idea how well a film will do before its release. He gives the example of one of the highest-grossing films of all time – “Raiders of the Lost Ark”. It was turned down by every studio except Paramount.
Film studios are not alone in not knowing everything. In 1994, a writer for Wired magazine noticed that the URL“www.mcdonalds.com” was still unclaimed, so he registered it and then tried to give it to McDonald’s. Their lack of understanding about the Internet was so amusing that it became an actual Wired magazine story. It proved difficult for Wired to give it to McDonald’s for free because they struggled to find someone who knew what the Internet was and whether they should accept a random journalist’s offer.
Kevin Kelly is the founder of Wired magazine. He is also the author of a wonderful book called “The Inevitable”, in which he writes: “Artificial Intelligence will define how things work and who we are in the future”.
Given the AI hype machine has been in overdrive, it’s hard to work out what is useful and how it could define how things work. There appears to be no formula or even guideline for working out where to place your bets.
In fact, the biggest problem with AI hype is that most of what is discussed under the term “AI” is not AI at all, but machine learning. Secondly, most writers about AI use the two letters as a shorthand for “what is ChatGPT usable for?”
AI use cases: from single point to the total customer funnel
The current use cases for AI in retail media revolve around automating personalisation, content creation, media planning and buying, audience creation and customer insights.
In the wider world of advertising outside retail media, WPP—one of the big advertising holding companies—worked on around five Super Bowl spots where AI played a role in elements like production, ideation, and media.
The clearest use case is video. One new example is from ByteDance, the Chinese company behind TikTok, which recently launched Omnihuman-1—a new AI tool that creates realistic videos from a single photo, of any length and style, combined with audio, with adjustable body proportions and aspect ratios such as portrait, half-body, and full-body formats.
You can see how such tools could be useful for Sponsored Brand Videos from Amazon Advertising and Walmart Connect. Sponsored Brand Videos allow you to share video content directly on the product detail page based on audience keyword searches. Given that the average ROAS for a Sponsored Brand Video targeting branded terms was 72% higher than other retail media ad types on Amazon, AI-powered video creation tools—by lowering content production costs for a high-performing ad unit—could increase conversions and revenue.
Amazon has now taken its AI-driven advertising propositions to the next level, where advanced AI works with humans across the entire marketing funnel. Amazon Ads now offers Brand+, which they describe as AI-optimised TV ad campaigns. The AI-driven optimisation platform is intended to simplify brand-awareness campaigns by delivering video ads across Amazon’s properties, such as Prime Video and Twitch, as well as premium video publishers like BuzzFeed and Fox.
Amazon claims to combine trillions of shopping, browsing, and streaming signals from across its ecosystem to model streaming television and online video audiences, predicting the consumers most likely to convert over time. Advertisers can apply AI-optimised automation to purchase TV advertising that reaches unique audiences of customers most likely to buy their products or services.
During Brand+ beta testing, some advertisers experienced more than a 10% increase in sales and over a 70% increase in website traffic.
Brand+ builds on Amazon Ads’ recent AI-driven programmatic media-buying innovations, including the launch of Performance+, which automatically optimises campaign performance for lower-funnel outcomes, and Ad Relevance, which uses advanced machine learning to match ads.
Amazon provides the example of a travel company looking to drive awareness among its highest-potential customers. The company could upload its first-party signals to the Amazon DSP, and Brand+ would analyse these signals along with Amazon Ads’ shopping and streaming signals to power machine learning models, identify patterns, and recognise customers who have searched for travel gear, bought travel guides, and streamed travel shows. Brand+ then delivers TV ads for the travel services to this audience of potential customers. The machine learning models continuously optimise the campaign by analysing new signals and engagement insights, helping reach high-potential customers early in their journey.
If “no one knows anything”, what should we do?
If, as Kevin Kelly says, “Artificial Intelligence will define how things work and who we are in the future,” what should people working in retail media actually do?
The first step is to carve out time and space to learn and experiment. Right now, as retail media continues its growth trajectory, short-term pressures dominate. Nevertheless, becoming more technical and proactive must be a strategic priority rather than an afterthought.
Here are some ways to think through applying AI in your retail media role:
- Allocate 10% of your time to exploring AI-related material and assessing how it could apply to your role in retail media.
- Allocate 10% of your time to experimenting with generative AI applications, testing text, audio, image, or video productivity tools that can streamline content creation for multiple marketing channels.
- Approach everything as an experiment rather than expecting immediate, perfect results. Or, as Simon Andrews puts it: “We just have to keep playing and experimenting.”
- Avoid thinking solely within your department. Instead, consider how strategy, audiences, and media can be driven by a single AI-powered system.
You are in: Home » Retail Media » OPINION Nobody knows anything: AI will define how things work and who we are in the future
OPINION Nobody knows anything: AI will define how things work and who we are in the future
Colin Lewis
One of the best books about Hollywood is “Adventures in the Screen Trade”, in which the screenwriter William Goldman writes: “Not one person in the entire motion picture field knows for a certainty what’s going to work. Every time out it’s a guess and, if you’re lucky, an educated one.”
He was referring to the fact that despite their market research, gut instinct and experience, no one in Hollywood has any idea how well a film will do before its release. He gives the example of one of the highest-grossing films of all time – “Raiders of the Lost Ark”. It was turned down by every studio except Paramount.
Film studios are not alone in not knowing everything. In 1994, a writer for Wired magazine noticed that the URL“www.mcdonalds.com” was still unclaimed, so he registered it and then tried to give it to McDonald’s. Their lack of understanding about the Internet was so amusing that it became an actual Wired magazine story. It proved difficult for Wired to give it to McDonald’s for free because they struggled to find someone who knew what the Internet was and whether they should accept a random journalist’s offer.
Kevin Kelly is the founder of Wired magazine. He is also the author of a wonderful book called “The Inevitable”, in which he writes: “Artificial Intelligence will define how things work and who we are in the future”.
Given the AI hype machine has been in overdrive, it’s hard to work out what is useful and how it could define how things work. There appears to be no formula or even guideline for working out where to place your bets.
In fact, the biggest problem with AI hype is that most of what is discussed under the term “AI” is not AI at all, but machine learning. Secondly, most writers about AI use the two letters as a shorthand for “what is ChatGPT usable for?”
AI use cases: from single point to the total customer funnel
The current use cases for AI in retail media revolve around automating personalisation, content creation, media planning and buying, audience creation and customer insights.
In the wider world of advertising outside retail media, WPP—one of the big advertising holding companies—worked on around five Super Bowl spots where AI played a role in elements like production, ideation, and media.
The clearest use case is video. One new example is from ByteDance, the Chinese company behind TikTok, which recently launched Omnihuman-1—a new AI tool that creates realistic videos from a single photo, of any length and style, combined with audio, with adjustable body proportions and aspect ratios such as portrait, half-body, and full-body formats.
You can see how such tools could be useful for Sponsored Brand Videos from Amazon Advertising and Walmart Connect. Sponsored Brand Videos allow you to share video content directly on the product detail page based on audience keyword searches. Given that the average ROAS for a Sponsored Brand Video targeting branded terms was 72% higher than other retail media ad types on Amazon, AI-powered video creation tools—by lowering content production costs for a high-performing ad unit—could increase conversions and revenue.
Amazon has now taken its AI-driven advertising propositions to the next level, where advanced AI works with humans across the entire marketing funnel. Amazon Ads now offers Brand+, which they describe as AI-optimised TV ad campaigns. The AI-driven optimisation platform is intended to simplify brand-awareness campaigns by delivering video ads across Amazon’s properties, such as Prime Video and Twitch, as well as premium video publishers like BuzzFeed and Fox.
Amazon claims to combine trillions of shopping, browsing, and streaming signals from across its ecosystem to model streaming television and online video audiences, predicting the consumers most likely to convert over time. Advertisers can apply AI-optimised automation to purchase TV advertising that reaches unique audiences of customers most likely to buy their products or services.
During Brand+ beta testing, some advertisers experienced more than a 10% increase in sales and over a 70% increase in website traffic.
Brand+ builds on Amazon Ads’ recent AI-driven programmatic media-buying innovations, including the launch of Performance+, which automatically optimises campaign performance for lower-funnel outcomes, and Ad Relevance, which uses advanced machine learning to match ads.
Amazon provides the example of a travel company looking to drive awareness among its highest-potential customers. The company could upload its first-party signals to the Amazon DSP, and Brand+ would analyse these signals along with Amazon Ads’ shopping and streaming signals to power machine learning models, identify patterns, and recognise customers who have searched for travel gear, bought travel guides, and streamed travel shows. Brand+ then delivers TV ads for the travel services to this audience of potential customers. The machine learning models continuously optimise the campaign by analysing new signals and engagement insights, helping reach high-potential customers early in their journey.
If “no one knows anything”, what should we do?
If, as Kevin Kelly says, “Artificial Intelligence will define how things work and who we are in the future,” what should people working in retail media actually do?
The first step is to carve out time and space to learn and experiment. Right now, as retail media continues its growth trajectory, short-term pressures dominate. Nevertheless, becoming more technical and proactive must be a strategic priority rather than an afterthought.
Here are some ways to think through applying AI in your retail media role:
Read More
You may also like
Subscribe to our email community