CDP Institute

Traffic to ecommerce websites fell 1% year-on-year while ecommerce app sessions rose 13%, according to Similarweb. Shopper experience still needs improvement: 41% of shoppers feel like strangers to their favorite retailers and only 20% see frequent personalization. Convenience is king: 55% say they’d pay more to find the right item without scrolling through several pages of results.

Consumers continue to resist advertising: Clutch reports that 93% say they skip ads when they can (55%) or always (37%). Oddly enough, this surveys show somewhat higher satisfaction with what consumers do see: 76% say they see relevant ads monthly and 47% are comfortable with personalized content. Entertaining or interactive ads are most likely to capture attention (78%), followed by ads that are informative or inspirational (63%).

Consumers aren’t getting what they want from customer service, either. Although CallMiner found that brands say just 20% of consumers view digital interactions as their preference for most issues, 84% of brands require them to engage with a self-service platform during a typical support journey. Nearly 30% of brands limit escalation options, which isn’t friendly at all.

Media mix models (MMM), originally developed in the 1960’s, have undergone a resurgence as new technologies make them easier to execute and click-based attribution becomes less credible. But evaluating MMM brings its own challenges. Australia-based Mutinex offers a vendor-neutral Open Source MMM Validation Framework which is now available for free to North American marketers and agencies.

Dr. Doolittle could talk to the animals, but how about the animal crackers? Now consumers can, so long as the cracker box has a QR code connected to Digital Link’s AI-powered product assistant. “This is not a gimmick,” says Digital Link CEO Paula Rivero.

Most marketers are (probably) less interested in talking to cracker boxes than the consumers who buy them. But real consumers can be tough to reach. Market Logic Software is solving the problem by converting consumer profiles into chatbots that will answer marketers’ questions through real-time, natural language interactions.

Amazon would prefer that you buy everything through them, but if you insist on using another merchant, they’re still willing to handle the shipping and fulfillment. They’ve just made their Multi-Channel Fulfillment (MCF) service available to Walmart merchants and will soon add Shopify and Shein options. MCF clients place their inventory in Amazon-run warehouses, where Amazon will pick, pack, and shop customer orders.

Google continues to expand use of its Gemini AI agent within the Chrome browser. Starting now, Mac and Windows desktop users will be able to ask Gemini to analyze pages they are viewing. In a few months, AI agents will handle tasks such as ordering groceries or booking a haircut.

Not sure you trust AI to work on your behalf? OpenAI researchers offered new reasons to worry, reporting that “AI scheming–pretending to be aligned while secretly pursuing some other agenda–is a significant risk that we’ve been studying” although it’s “not an imminent behavior in our currently deployed systems.” Note that the quote contains emdashes without surrounding spaces, hints that the study report may have been written by an AI with good reason to downplay its true scheming abilities. Just saying.

Tools to help brands measure and optimize their AI search results are now too common to mention. But it’s still unusual for a CDP vendor to provide one. So let’s note that CDP Zeta Global is adding a Generative Engine Optimization solution to its AI Marketing Cloud.

Captify Technologies captures results of conventional search interactions – up to one billion per day, in fact. They use this behavior to build programmatic ad audiences without tracking consumer identities. It’s an appealing enough capability that Captify was just purchased by ad tech vendor Verve.

Model Context Protocol (MCP) is everyone’s favorite technology right now because it promises to eliminate integration bottlenecks, especially for AI systems. But unless you just fell off the turnip truck, you’ve probably suspected it wouldn’t deliver everything that’s expected. Sure enough, Microsoft researchers have just found there are significant limits to what MCP can do when multiple agents and MCP servers are involved in the same process.