|The Privacy Sandbox Debate: Two Sides of the Same Coin
Image sourced from Shutterstock
The Privacy Sandbox launched a polarizing debate even before Chrome enacted its Tracking Protection feature on 1% of its audience. On the other side of the fence, voices from all sides of the ad tech industry raised concerns about Google's true motives for constructing the Privacy Sandbox. Critics wonder if privacy is the true goal or if it is a tool "used to further Google's data collection efforts."
Notably, The Trade Desk refuses to participate in the Privacy Sandbox initiative. The DSP called out big G, stating, "Google is self-serving and sandbox solutions are inadequate."
But what do the results say? Paul Bannister, Chief Strategy Officer at Raptive, posted some results on LinkedIn, and these are the significant stats:
• The Privacy Sandbox is not at a full 1% deprecation yet
• Cookieless Chrome users are monetizing about 30% worse than those with cookies.
• Chrome iOS users perform 15% better than Safari iOS users, even though they are identically unaddressable.
In response to the criticism, Google's senior director of product management, Victor Wong, addressed some misconceptions. Here are his noteworthy responses:
Objection 1: Privacy Sandbox doesn't provide one-to-one replacements for third-party cookie-supported use cases, so don't expect the same features or functionality as third-party cookies, or you will be disappointed.
Objection 2: Third-party ad tech solutions that recreate web-tracking IDs without cookies are easier to retrofit into existing products, but those solutions aren't a meaningful improvement' on third-party cookies.
Which side is right? Only further testing post-cookie solutions will give us that answer. However, the industry-wide agreement is to tailor post-cookie solutions to your business needs. There's no one-size-fits-all solution. - AB
|According to New Report, Big Brands Lack Preparation for Cookie Deprecation
|It could be Google's repeated delays or just plain wishful thinking. To assess the advertiser community's level of readiness, Adlook, an ad tech company, surveyed nearly 200 CPGs to learn about their plans for thriving in a cookie-free world. Surprisingly, more than half (55%) are still exploring their options or are just beginning to consider what those options may be.
Less than half (45%) are making "major" revisions to their data collection and usage pattern, which is concerning considering the number of consumer data privacy laws currently wending their way through state legislatures. That said, over one-third (38%) expressed concerns or skepticism about the impact cookie deprecation will have on ad targeting. - SS
|OpenAI Says The New York Times Isn't Telling the Whole Story
|The Times surprised no one when it filed a lawsuit against OpenAI, claiming the AI company violated its copyrighted materials by training ChatGPT on millions of its articles. This week, OpenAI launched a PR blitz to announce to the world that there's more to the story than meets the eye and that, despite the lawsuit, the company still wants to be friends with the publisher.
In a blog post, OpenAI says it collaborates with news organizations and that training its chatbot on published material falls under fair use.
"The principle that training AI models is permitted as a fair use is supported by a wide range of academics, library associations, civil society groups, startups, leading US companies, creators, authors, and others that recently submitted comments to the US Copyright Office." That said, OpenAI says it invites publishers to opt out because it's just what good folks do. The New York Times, according to the blog post, exercised that right this past August.
At issue is the notion of "regurgitation" (regurgitation, hallucinations, the language surrounding generative AI is quite colorful!). Regurgitation refers to LLMs memorizing data it has been trained on, which leads to regurgitating, rather than creating new text, in response to user prompts. It's a phenomenon that OpenAI claims is exceedingly rare; the Times claims otherwise.
"Interestingly, the regurgitations The New York Times induced appear to be from years-old articles that have proliferated on multiple third-party websites," writes OpenAI. To obtain those regurgitations, the NY Times manipulated ChatGPT prompts and even went so far as to include lengthy quotes from those articles.
While The Times may have manipulated its prompts, prompt manipulation is a crucial concern with generative AI (think ChatGPT DAN). OpenAI's position appears predicated on the notion that users will always play fair when prompting ChatGPT for responses. This is a key issue, considering that responses generated by ChatGPT are potentially copyrightable content, and citing the sources that led to the ChatGPT response is difficult to trace.
It will be interesting to see how this plays out. - SS
|CES: The Ad Products To Watch
|Over 100,000 people flocked to Las Vegas for the Consumer Electronics Show (CES), and we published a "how-to" for publishers and advertisers to make the best of their time at the conference. AI was a big pull for the conference, but what advertising products made a splash?
NBCUniversal's One Platform: The company boasts the Total Audience platform as its next step in cross-platform, audience-based advertising. Marketers can use NBCU's AI-driven planning and activation tech, leveraging the company's new investments in automation and data to provide deduplicated reach through a single buy across linear and streaming.
Shoppable Ads on Hulu (and Disney+ soon): The Gateway Shop, a new evolution of interactive ads, builds upon the current Hulu ads where users can scan a QR code or receive additional information on their phone by selecting the option with their remote. These new ads emphasize commerce, aligning with streaming's shift towards generating business outcomes for brands, moving away from the awareness metrics traditionally linked to TV advertising.
Monks.Flow: This AI suite for marketers offers a platform-agnostic environment tailored to a company's unique needs. Marketers can reduce costs by optimizing workflows with synthetic media, integrating tools and processes to de-silo the organization, building business intelligence through live data streams for informed decision-making, and maximizing impact by reaching new audiences and driving business outcomes for growth. - AB