CDP Institute

Google will pay to carry stories from Canadian news outlets under an agreement announced yesterday.  Details are sparse but the process for compliance with government’s C-18 law has been simplified to meet some Google objections.  Meta, the only other firm required to make payments under the law, has not changed its decision to block reliable news.

The threat that large language models will reach human levels of performance on common tasks is overblown, according to research co-authored by Meta’s top AI scientist Yann LeCunn.  A new benchmark of real-world problems found that humans succeeded at 92% while ChatGPT-4 with plug-ins solved only 15%.  A cynic would note that Meta has a strong interest in reducing pressure for AI regulation, and that other experts have argued the debate over human-like performance is a distraction from concrete problems that AI poses today.

Shopping on social media has had mixed success but this report from The Influencer Marketing Factory finds it’s making steady progress.  Nearly half (43.4%) of consumers surveyed had bought at least one product through a social app in the past year.  Still, many more would rather buy on an ecommerce platform (62%) than through social (22%), with trust in the social apps standing as the most important obstacle.

Do not confuse Snow Software, which helps companies track IT costs, with Snowflake, a cloud database often used for CDP-ish solutions.   Do take note that Snow has just been sold to Flexera, which also helps companies optimize their return on tech investments.  It’s a helpful reminder that cost-conscious IT departments are more than happy  to “help” marketers who are doing a poor job of controlling tech spend.

Sure, cloud databases are popular, but the really cool IT kids have moved on data lakehouses, which combine data warehouse functionality with data lake flexibility.  Or so says this survey from Dremio, which sells guess-what.  Lakehouse advantages over cloud databases include lower cost, self-service access, and easier support for AI.

Consumers may find passwords annoying, but IT security people dislike them much more intensely.  This Delinea survey finds that biometrics are the preferred replacement among IT security teams (59%), followed by multi-factor authentication (46%), one-time passwords (37%) and PIN codes (30%).  It’s not clear that consumers have quite the same enthusiasm for biometrics, which often feel creepy.

Amazon Web Services (AWS) and Salesforce announced a set of new integrations, including sales of Salesforce products on the AWS Marketplace and easier data sharing between Salesforce Data Cloud and AWS data services.  Most of the changes are fairly technical but the general lesson is that even mighty Salesforce is ultimately a guest on turf controlled by AWS.

Oh, the irony.  Just as streaming TV services recognize that selling ads is more profitable than selling subscriptions, European regulations are forcing ad-sales giants like Meta to offer ad-free subscription services.  Snapchat is the latest to roll out a paid, no-ad option in certain markets.  Snapchat’s post is cryptic so here’s a MediaPost article  with some perspective.

This Canto study isn’t entirely about AI but the most interesting tidbit is that B2B companies are twice as likely as B2C companies to use AI for content creation (49% vs 27%) and to search existing content (40% vs. 18%).  Feel free to speculate why.  On the other hand, pretty much everyone agrees they need more content and AI can help.

Mindbox, a Russia-based customer experience platform that qualifies as a CDP, set up a separate U.S. operation shortly after Russia invaded Ukraine.  They’ve now formally relaunched and renamed the new company as Maestra.  The focus is on mid-size retail.

Content generation is the most common marketing use of AI, right?  Not so: this Movable Ink survey places content creation dead last at 26%, trailing customer behavior analysis (59%), personalization (56%) and campaign optimization (55%).  Managing customer data (40%).  is the most commonly-cited challenge with AI integration, followed by data privacy (38%), consistent omnichannel personalization (37%), brand reputation (29%) and lack of skills (23%).  Interesting.

Our friends in IT are also eagerly exploring AI integration, with 35% even ranking it ahead of perennial obsessions security (31%) and cost (29%) for the past twelve months.  The gap grows even larger for 2024 (AI integration 35%, IT cost 28% and security 28%).  Meanwhile, 88% are moving towards platforms and away from point tools, perhaps as a way to rein in costs (although just 23% think they’re overspending on SaaS applications).