A lot moved in marketing this week. Here is a fast scan of the signals worth tracking, with the operator angle on each.
Influencer Budgets Are Shifting to the Middle
Between 60 and 70 percent of influencer briefs now favor mid-tier creators over mega influencers, according to reporting from Storyboard18. Brands are prioritizing engagement and measurable performance over raw follower count. Creators in the 50K to 500K range are delivering lower cost per view, while historical watch time and audience fit are increasingly driving pricing decisions. Mega creators still command premiums when mass reach or cultural relevance is genuinely needed, but brands are pushing for fees tied to expected campaign performance rather than status.
For solo operators running lean influencer programs, this is a tailwind. The mid-tier tier is more accessible and now more defensible to justify internally.
ChatGPT Ads Hit $1B in Annualized Revenue in Under 200 Days
One analyst makes the case that OpenAI can build an effective ad model by keeping ads separate from answers and clearly labeling them, similar to early Google AdWords. The data point behind the optimism: ChatGPT Ads reached $1 billion in annualized revenue in under 200 days. Advertiser demand is already there before the format has matured. Running shoes, travel, and price comparisons are called out as especially promising categories because users now ask ChatGPT to research those commercial decisions instead of going to Google.

Customer Proof Works Better as a Story Than a Logo
A roundup of unconventional social proof tactics is worth bookmarking. The examples: Deel lists Snoop Dogg as a customer, Hinge commissioned novels from real couples’ stories, and Shopify lit up the Las Vegas Sphere with every merchant’s first sale. Smaller versions work too. Turning a “too advanced” one-star review into an ad, or showing “customer since” dates under logos, both carry more credibility than a standard logo wall.
Lead With the Human Outcome, Not the Tech
A piece on product storytelling makes a point operators building AI tools should read carefully: lead with the customer experience, not the technical capability. VisiCalc gave people a clear reason to buy the Apple II. Tesla could frame Full Self-Driving around making long-distance travel less exhausting rather than leading with autonomous driving as the feature. Turn complex technology into a recognizable picture of what life looks like after the product does its job.
5 AI Marketing Takes Worth Stress-Testing
A Substack piece runs through five claims about AI and marketing that cut against the standard narrative:
- AI may not reduce marketer demand. Leaner teams could be offset by the growing number of businesses that need distribution.
- Chat ads could become a major ad format as LLM usage grows.
- Shared repositories can preserve institutional knowledge and give AI better context to work from.
- Employee-generated content can extend brand reach through trusted voices.
- As AI handles more digital interactions, in-person events may become more valuable for human connection and closing deals.
OpenAI’s Ad Pixel Collects More Than You Might Expect
OpenAI’s ad pixel ties a one-year __obi cookie to ChatGPT accounts and sends data to OpenAI when an advertiser’s tag loads. It also collects hashed emails and phone numbers, unencrypted city and postal codes, and page paths containing medical and debt topics. The pixel works for logged-out users but appears limited to Chrome on Android. iOS browsers block it.
Usage Churn Drives 26 to 40 Percent of Subscription Cancellations
RevenueCat’s research puts usage churn behind 26 to 40 percent of subscription app cancellations. The root cause is usually a broken habit loop, not price. Defaulting to discounts when users churn misses the actual problem. The fix depends on where the loop breaks: weak triggers call for timely re-engagement, while poor onboarding or unclear value requires getting users to a meaningful outcome faster. Retention improves when the intervention matches why usage dropped.
