Subscribe | Free AI | CRO Audit | Cool Tools

Which Test Won?

For a SaaS company, which variation had more paying customers over time, and why?

This week’s sponsor is Ometrics AI

News This Week

AI

• Anthropic’s New A.I. Model Sets Off Global Alarms.

  Mythos has triggered emergency responses from central banks and intelligence agencies globally, as Anthropic decides who has access to the powerful model. (Link)

• Anthropic and Amazon Sign a $100 Billion, 10-Year Compute Deal

  Anthropic and Amazon inked a landmark $100 billion agreement securing up to 5 gigawatts of computing power over ten years, with new Trainium2 and Trainium3 chips coming online this year. Claude will become directly available within AWS to over 100,000 enterprise customers, cementing Anthropic as the dominant AI supplier in the cloud wars. (Link)

• OpenAI Launches ChatGPT Images 2.0 as the New Leader in AI Image Generation

  OpenAI released ChatGPT Images 2.0, delivering faster, higher-fidelity output with expanded layout options and the ability to handle complex multi-image tasks that previous models couldn't manage. It immediately claimed the top spot in every category on Image Arena with a record 242-point lead over the nearest competitor. (Link)

• SpaceX Secures Rights to Acquire Cursor for $60 Billion — or Pay $10 Billion for Joint Work

  SpaceX confirmed an unprecedented deal giving it the option to acquire AI coding platform Cursor for $60 billion later this year, or alternatively pay $10 billion for their joint development work. The arrangement combines Cursor's developer product with SpaceX's massive AI training compute as Musk's companies prepare for an IPO and compete against Anthropic and OpenAI in enterprise coding AI. (Link) (Link) (Link)

• Google Transforms Gemini Enterprise into an All-in-One Agent Platform at Cloud Next

  At Google Cloud Next, Google folded Vertex AI into Gemini Enterprise Agent Platform, a unified end-to-end system for building, deploying, and governing AI agents at scale. The launch adds low-code Agent Studio, long-running agent runtimes, simulation tools, and enterprise governance controls like Agent Identity and Model Armor, alongside access to 200+ AI models. (Link)

• OpenAI's Codex App Is a Preview of ChatGPT's Super App, and It Runs in the Background

  OpenAI's updated Codex app works as a true autonomous agent with its own cursor, running tasks in the background at the OS level without hijacking the user's screen. With persistent memory via Chronicle, 90+ plugins, and scheduled overnight workflows, it's the first unified AI workspace that lets multiple agents work in parallel while you keep doing your own work—a significant leap beyond anything Anthropic currently ships. (Link)(Link)

• Claude Launches Live Artifacts, a Real-Time Auto-Refreshing Dashboards Inside Cowork

  Anthropic released Live Artifacts, enabling Claude users to build dashboards and trackers that automatically refresh with current data inside Cowork. All artifacts are saved to a new Live Artifacts tab with version history accessible across any session, and the feature is now available on all paid Claude plans. (Link)

Digital Marketing

• OpenAI Launches CPC Ads Inside ChatGPT at $3–$5 Per Click — and Minimum Spend Has Dropped to $50K

  OpenAI opened cost-per-click bidding to select advertisers inside ChatGPT, with CPCs ranging from $3–$5 and CPMs cooling from $60 to around $25 since February's launch. Minimum spend has dropped from $250,000 to $50,000, signaling a rapid shift toward self-serve access. With thin competition and users in active research mode, early advertisers have a narrow window to test before the crowd arrives.

• 49% of AI Users Will Switch Brands If an AI Recommends a Competitor

  New eMarketer research finds that nearly half of active AI users say they are likely or very likely to try a different brand if an AI assistant recommends an alternative — even though only 46% fully trust AI recommendations and 86% still verify AI information before buying. The implication is clear: AI brand visibility is now a competitive moat, not a nice-to-have. (Link)

• Google's Dynamic Search Ads Are Retiring in September — Transition to AI Max Now

  Google has announced that Dynamic Search Ads will be sunset in September 2026 and automatically transitioned to AI Max campaigns. Marketers who switch proactively gain more control over targeting, creative, and landing page settings than those who wait for the forced migration — making now the right time to audit and rebuild DSA campaigns inside the new AI Max framework. (Link)

• The "Bland Tax": Generic Brands Are Being Erased From AI Search Results

  Search Engine Land reports a growing pattern: brands with vague, undifferentiated positioning are systematically disappearing from AI-generated answers. When AI models can't clearly summarize what a brand does or why it's different in a few sentences, they simply skip it. The antidote isn't more content — it's sharper, more specific brand language that gives AI models something concrete to cite and quote. (Link)

• Self-Promotional Listicles Face 29–49% Visibility Drops and $53K Fines Per Violation

  GrowthWaves documents what Google has been quietly enforcing: self-serving listicles — content designed to rank a brand's own products favorably under the guise of editorial review — are experiencing visibility drops of 29–49% and face potential penalties of $53,088 per violation. Google confirmed awareness of the tactic, calling it one of the most abused patterns in SaaS content. (Link)

Conversion Optimization

• AI Is Flipping the Purchase Funnel. Brands Must Now Win at the Bottom Before the Top

  Search Engine Land argues that AI-mediated discovery is forcing a "funnel flip" where brands need to convert high-intent, bottom-of-funnel users discovered through AI answers before investing in broad awareness. When an AI assistant surfaces your product in response to a specific need, that user is already pre-qualified — the conversion challenge shifts entirely to the landing page, trust signals, and frictionless checkout rather than awareness or consideration. (Link)

• Banks That Invest in CX Outperform the Stock Market, New Data Shows the ROI

  A new study finds that banks with superior customer experience ratings significantly outperform peers on stock performance over time. CX Dive's coverage frames this as a boardroom-level argument: CX investment is not a cost center but a financial return driver — which means conversion optimization tied to experience quality can be pitched in the same language as any other capital allocation decision. (Link)

Cool Tools

• Claude Design - Claude can finally draw (Link) (Link)

• ChatGPT Images - ChatGPT image update (Link)

• AI Image Extender - Ads image content to make images fit any size (Link)

• VidGenerator - create videos and images (Link)

Which Test Won Answer

Variation A will have the most sign-ups, but the reason for long-term revenue may not be so straightforward.

Sites where you have to enter your credit card have 80% less signups than those without the credit card. But the trial-to-paid subscription rate is 50% compared to 15% for no credit card required.

However, sites where you enter your credit card also have a higher churn. This can also cause bad branding, i.e., the client forgets they entered their credit card, received a charge, and then gets pissed at themselves and the brand.

In the end, three to six months out, a SaaS with no credit card required typically has a conversion rate of twice that of a credit card required.

The theory behind this reasoning is that you get more customers trying your service with a free trial, no credit card required. Because customers are making the decision to pay because they like your service, they will last longer compared to customers who are forced to enter a credit card to try it out.