A roundup of the marketing and growth signals worth paying attention to this week: physical print as event strategy, AI shopping behavior, ChatGPT’s coming ad format, and a case for building instead of buying when your data is the differentiator.
Brand gazettes: print as a post-event asset
Brands are producing custom print publications tied to specific activations, giving attendees something tangible to take home. Grey Goose used its gazette to tell stories about the brand’s French heritage. Disney recreated the fictional Runway magazine from The Devil Wears Prada 2 as a real 90-page issue. QR codes and interactive elements connect the print content back to digital experiences.
The operator angle here is less about budget scale and more about intent. A purpose-built print piece extends the event past the room and past the day. If you run conferences, pop-ups, or brand activations, a gazette is worth modeling as a content artifact rather than just swag.
Consumers are stretching budgets with AI
Grocery prices are up 32% since 2020, yet spending on cars, clothing, and restaurants still increased in August. Some consumers are using credit or drawing from savings. Others are cutting elsewhere. One consumer reported using ChatGPT to compare grocery prices and saving about $100 a month.
For marketers, this signals that price sensitivity is real but spending hasn’t stopped. The buyers who feel squeezed are actively using tools to justify purchases, not just avoiding them.
ChatGPT visual ads are coming to the US
OpenAI is testing visual ads inside ChatGPT that show products and services through images during image generation. The format lets advertisers demonstrate products in use and link to more information. Conversion tracking and attribution are being added through partners including Hightouch, Tealium, LiveRamp, AppsFlyer, and Triple Whale. US testing starts later this month.
This is early, but the infrastructure signals intent. If you’re running performance campaigns, the attribution partners listed are names you already know. Worth watching for how CPMs and conversion rates compare to existing social formats.
➿ Growth loops and the slowest step
A post from Expo’s blog identifies three compounding growth loops: app installs improving store visibility, content informing the next piece of content, and product usage generating reviews and user-created content. The content loop is flagged as the slowest because it requires the most manual work, yet it can produce significant results. The recommended fix is to find the slowest step in each loop and remove unnecessary manual handoffs so each cycle completes faster.
The framework is transferable to solo operators running any content-driven acquisition channel.
️ When building beats buying: the custom Calendly case
A 3-person team built their own scheduling tool in 20 minutes after their AI agent identified a limitation with Calendly. The custom tool connects scheduling to Salesforce and prospect data, shows each prospect a personalized prospectus, and routes them to the right salesperson based on account ownership. The team considered integrating Calendly, but the agent determined that most of the custom logic would still need to be built anyway.
The takeaway the team landed on: the default remains buying software. Custom builds are reserved for narrow workflows where proprietary data creates an advantage that off-the-shelf tools can’t replicate.
Google update signal: AI content guidance just changed
Google recently rewrote its documentation on AI content and authorship, which historically precedes a major algorithm update. Under the updated guidance, publishing AI text at scale without manual review counts as little to no effort. Invented authors with AI headshots count as deception. GEO tactics built to win AI answers, including self-ranking listicles and paid mentions, are being compared to old SEO spam patterns that Google has previously targeted. The recommendation from SEO analyst Lily Ray: cut or consolidate scaled content before an update rather than trying to recover after.
AI shopping assistants: a three-way comparison
A hands-on test compared ChatGPT, Claude, and Muse for shopping soccer shoes. ChatGPT won by showing 3 options with a recommended pick and clear comparison. Claude offered useful product details but lacked images and purchase links. Muse made the experience more engaging through interactive questions and visuals but recommended the wrong shoes. The signal for anyone building or optimizing AI-assisted commerce: a small number of options plus a clear recommendation outperforms either a long list or an engaging experience that gets the answer wrong.

