AI has taken center stage in many marketing-related discussions, but it took a supporting role at the fall Digiday Media Buying Summit, which concluded Thursday in Palm Springs, California.
During the three-day event, media agency executives discussed the potential of machine learning and generative AI both on and off stage. But they also confessed to having a lot of healthy skepticism.
Marilois Snowman, founder and CEO of media services agency Mediastruction and SaaS platform FutureSight, said AI can give you answers, but not necessarily the right answers. Snowman, who spoke on stage about media mix modeling, said AI can help extract, transform, and load data more quickly and efficiently. But she noted that it’s important to use the tool wisely to ensure the output is accurate.
“We’re still thinking more about machine learning algorithms because we can understand how they arrived at the answer,” Snowman said. “These outputs are only as good as the data they’re fed. So if you want to understand where things went wrong, find out the margin of error between what the model is predicting and what’s actually happening. And even if you find that your error bars aren’t improving, the ML model will help you better understand why.
In the keynote speech that kicked off the summit, Kamran Asghar, co-founder and CEO of independent media agency Crossmedia, said that clients said he is keen to test new AI tools.
“Once we have evidence that things are going well, we are using AI for predictive modeling to understand where to spend the next best money,” Asghar said. “So the standard step is to keep it around but not let it end up being everything.”
AI tools also have the challenge of potential platform bias. Asked about the revelations in Google’s ad tech antitrust case, Snowman said the loss of trust creates problems with measurement, saying, “We have to play in their sandbox (just kidding). But… We don’t have to rely on them for measurements,” she said. .
Warnings against over-reliance on automated ad buying were a key theme throughout the summit. Tucker Matheson, co-founder of digital strategy firm Marcacy, said the past decade of digital media has created platform bias. While large advertisers have the resources for media mix modeling, this isn’t necessarily the case for smaller brands. Matheson also notes that it may not be wise to lock in media budgets based on what a platform offers, in terms of traffic acquisition costs and return on ad spend.
“It’s so easy to go to Meta or Google, look at your data, and click a button,” Matheson says. “Platform metrics are still important, but I think we need to look at them a little differently. We all know, as we do, that creative testing is proliferating on Meta and TikTok. I think these platforms provide great click data to support creative testing.”
(Coincidentally, Google also announced new advertising features for P-MAX and Search campaigns this week, including new ways to optimize campaigns and track performance.) The Google Ads API version 18 update includes Includes new ways to query placement-specific data and content metrics for demand generation and video campaigns.
Meanwhile, the growing presence of AI in creative ad generation is being felt across the marketing spectrum. Kevin Rettig, Marriott’s senior director of marketing platforms and privacy, says balancing brand voice with accuracy is key to leveraging AI for customers. In a conversation about personalization, Rettig said Marriott hasn’t done much with generated AI ads yet. However, we are experimenting through a virtual assistant that was recently deployed using ChatGPT and open source outlets. Another challenge is maintaining the brand voice across the hotel giant’s dozens of brands. Each brand has its own unique brand voice.
“When you think about AI and content generation, brands really value their brand voice,” says Rettig. “There has to be a balance between automation and having the creative parts written and executed by someone else. That’s not to say we won’t do it, and I don’t think every brand should do that. I think we have a sensibility, but there are 30 brands that we really feel like have unique brand voices, so we really need to be mindful of that as well.”
Beyond AI, Rettig said the hotel giant is exploring different ways to personalize marketing through the Bonvoy loyalty program and other data sources. He said the company uses first-party data to understand customers’ hotel preferences, drive targeted promotions and enable advertising purchased through the Marriott Media Network. Marriott is exploring ways to expand its data pool through third-party data sources and cleanroom partnerships, while also using data validation tools such as Neutronian and Truthset.
Prompts and Products — AI News and Announcements
X is updating its User Terms and Privacy Policy for users, including changes to how data is collected, shared, and used to train X’s AI models. The new terms, which take effect in November, will also allow X to share user data with “third-party collaborators” for “independent purposes”, including training AI models “whether generative or otherwise”. Google has announced new AI shopping features for personalization ahead of the important shopping season. The company also announced that Prabhakar Raghavan, the company’s top search and advertising executive, will change his role to chief engineer, working directly with CEO Sundar Pichai. Raghavan will be succeeded by Nick Fox, vice president of product management at Google. This change is primarily due to Google moving the Gemini apps team under Google Deepmind. Amazon’s annual event UnBoxed introduced new machine learning tools for creating campaigns, finding relevant audience segments, and monitoring various metrics. The e-commerce giant also announced a new AI creative studio and audio generator to help scale your campaigns. Adobe announced new AI creative tools at its annual Max conference, including new GenStudio for performance marketing and a related partnership with Microsoft Advertising. Adobe also announced new AI video models trained on licensed content. Taboola has introduced a new generative AI advertising assistant called Abby that enables marketers to create and manage digital campaigns through a chat-based interface. Perplexity has announced the first version of its enterprise search, called “Internal Knowledge Search,” which enables employees to use LLM to conduct research at work. Mozilla and the Open Markets Institute have partnered on a new report on AI and competition.
AI-related stories DIGIDAY[Japanese version]
The AI revolution has increased the demand for AI librarians. (DIGIDAY) Publishers are receiving credits from AI companies to use large-scale language models through recently signed content licensing agreements. (DIGIDAY) Marketers aren’t enticed by new AI-powered search tools. (DIGIDAY) Marketers are grappling with the AI advertising puzzle as TikTok rolls out Smart+. (DIGIDAY) Agency executives are balancing strategies for automation and new ad formats. (DIGIDAY)
AI news from elsewhere
Security researchers have discovered a way to use malicious GenAI prompts to collect personal data based on chatbot users’ past conversations. (Wired) YouTube has taken a step towards labeling videos using the C2PA standard. (The Verge) The UK government has said new legislation may be needed to address uncertainties related to AI and copyright. (Lexology) Generative AI is creating a flood of fake AI voices that sound like Kamala Harris and Donald Trump (The Washington Post) Dane Stucky, former chief information security officer at Palantir, becomes OpenAI’s new CISO I joined as. (Tech Crunch)
Source link
