Optimize - ChatGPT On Top For Now But For How Long? Button Test Results

Subscribe | Latest News | Cool Tools | Which Test Won?

Latest News

• Mistral claims the second biggest model behind ChatGPT (Link). Mistral is working with Azure with multi-lingual LLM. Mistral also just announced that it is coming to AWS Bedrock soon (Link). Mistral LLM usage is growing, they seemed to be focused more on offering models for developers and not tools for consumers to use as ChatGPT offers. The bottom line is whoever has the better LLMs will win. Switching LLMs is relatively simple.

• Meta is working on a mobile LLM that resides on the device, increasing speed. (Link)

• A new approach to reducing token consumption - Decoupling Reasoning from Observations for Efficient Augmented Language Models, detaches the reasoning process from external observations. (Link)

Comment: The above news only underlines how fast the race is for the LLM space. I imagine that soon there will be many LLM’s available, each for specific use cases. Once this is accomplished, the amount of computer power needed may change as the industry shifts focus away from creating LLMs. The industry will be more focused on providing quick access to LLMs, gathering data efficiently and securely, and system integrations.

• Google’s Genie can now generate interactive worlds based on 2D images. The data set is from publicly available internet videos (Link)

• Google adds AI to help you write to Chrome. This was first released in Gmail and is now being tested in Chrome. (Link)

• Through the Google News Initiative, Google is paying news publishers to use an unreleased suite of gen AI tools in exchange for their feedback. (Link)

• Slack adds AI to summarize conversations and improve search. (Link)

• Meta’s layoffs continue to impact advertisers as the company replaces account team members with AI. (Link) Klarna, the fintech company, is reporting that it laid off 700 customer service reps in 2022 and replaced them with AI.

• Ooops - The latest version of Google's Gemini artificial intelligence (AI) will frequently produce images of Black, Native American, and Asian people when prompted – but refuses to do the same for White people. (Link)

• Tumblr and WordPress to Sell Users’ Data to Train AI Tools. (Link)

MaxAI.me - Do more Faster with 1-Click AI

MaxAI.me lets you chat with GPT-4, Claude 3, Gemini 1.5. You can also perfect your writing anywhere, save 90% of your reading & watching time with AI summary, and reply 10x faster on email & social media.

Join 1M+ users on MaxAI.me to do more faster!

Cool Tools

• Adsby - AI co-pilot for Google search ads (Link)

• Stey - AI for web copy (Link)

• Saner - Scale and utilize your knowledge easily (Link)

• Superhuman - Released Instant Reply to speed up your emails (Link)

• TypePromt - AI content for social posts and blogs (Link)

• HeyEditor - AI video and photo editor (Link)

• Openmart - Selling to local businesses? Let AI easily build a list from Openmart’s Database Link

Which Test Won?

A/B testing is pivotal in conversion rate optimization. Every other week, I will share real A/B tests I’ve conducted for clients, and you can pick which variations performed better.

On a product page for a product that costs $300, which variation will get the most clicks to add to the cart and or to checkout?

Last week’s A/B test was kind of a trick question. The first clue is “product that costs $300,” which can indicate a higher probability of a one-time purchase. Typically, the fewer decisions a shopper has to make, the higher the conversion rate. Asking “Add to Cart” and “Buy Now” could hurt the conversion rate as well as the AOV. What is interesting about this test is the difference between mobile and desktop users. Here are the results for the test below:

Newsletter members: 61% voted for Variation A

The actual results for button clicks were:

Mobile: Variation A won with a lift of 70.58% @ 96% statistical confidence
Desktop: Variation B won with a lift of 123.76% @ 98.% statistical confidence
The combined results were not statistically accurate but showed a slightly negative result for Variation B because the site had more mobile traffic.

The reasoning: On mobile, “Buy Now” is a one-click step, which makes life easier for mobile users. On desktop, the shopper has more to explore and is more likely to use the “add to cart” button as a wishlist until they are ready to make a purchase. This can cause mobile to have a lower AOV but they did make a purchase.

Note that as with all tests, you need to test this on your site before implementing it. A test result from one site does not indicate it will be the same on another site due to many factors.

See you next week for another test!