Even though I pay the $299/year Black Friday rate, I just don’t use WA anymore, and I no longer find real value in the training.

One of the things I do want to do is transfer my domains out of WA. I’d love to hear from people here:

• What was your experience transferring domains away from WA?

• Which registrar/provider did you use — and would you recommend them?

• Any warnings, tips, or pitfalls I should be aware of (authorization codes, domain lock, renewal costs, etc.)?

Thanks in advance for any advice or personal experiences.

Mike

So, I had an email from Clickbank this morning flogging a course on the importance of good headlines. The email provided some clues as to how to write the best affiliate headlines (according to Clickbank - they should know, lol). 

I immediately thought of Chatty and pasted the email into the chat, asking my mate to create a 'perfect headline' generator ... essentially reverse engineer the email and fill in the blanks.

Here's what it came up with ... see how you get on!

✅ Final Headline-Generator Prompt (with Web Search + No Fabrication)

Act as a senior performance copywriter and a cautious researcher.
Your job is to produce 7 unique, non-formulaic, high-converting headlines, but only after gathering accurate information.

🔹 Workflow

Step 1 — Ask for one thing only:
“Please provide the product or service name.”

Step 2 — Perform a brief web search using that name to gather:

• Target audience

• Main benefit or main problem it solves

• Distinguishing features or positioning

🔹 Critical Rule (No Fabrication)

If the web search returns no clear, verifiable information about any of these items, you must stop immediately and say:

“I couldn’t find enough reliable information.
Please provide the target audience, main benefit/problem solved, preferred tone, and platform manually.”

You must not infer, assume, or invent details.

Step 3 — If information is found

Present the findings back to me in a short list and ask for confirmation or edits:

“Here’s what I found based on the web search—please confirm or refine:
• Target audience: …
• Core benefit/problem solved: …
• Distinguishing features: …
Also confirm your preferred tone and platform.”

Step 4 — After confirmation

Generate 7 original headlines that follow these rules:

• Speak directly to the audience

• Avoid clickbait and exaggerated claims

• Use clear specificity

• Use numbers only when they naturally fit

• Keep length scroll-stopping but not truncated

• Optional bracketed context (e.g., , )

• Include one curiosity-based headline that opens a loop without deception

• Each headline must use a different style (benefit-driven, emotional, question, contrarian, data-driven, short punchy, bracket-boosted)

• No templates, no repetitive structures, no detectable patterns

Here’s what I don’t get - where are they getting the numbers from?

As far as I know, OpenAI, Perplexity, Claude, and Google’s AI Overviews don’t release any kind of public query data. Are they just pulling from Google keyword volume and applying some kind of formula?

My gut tells me not to trust these tools. Anyone know how reliable (or totally made up) these “AI volume” metrics might be?

If it’s all predictive, does it even matter? Or should we be focusing more on whether our content shows up in AI answers, not how often people ask the AI?

Spending any time investigating this stuff makes my brain hurt 🤯

It’s only about seventeen quid, so not exactly a massive gamble, but I’m wondering if anyone here’s actually given it a go and seen any proper results? Not just the usual shiny screenshots.

I'm still finding my feet with SEO and socials, but YouTube’s always felt a bit scary. The idea of doing faceless videos with AI appeals to me though, especially if it can pull in traffic without me having to act like a vlogger.

Worth a punt? Should I blow my tea and biscuits fund on it?

Cheers,
Rita

Start by reviewing your analytics dashboard—Google Analytics is a popular tool for this—and identify the top referrer for the day. This could be a search engine, social media platform, email campaign, or direct traffic.

Once you’ve pinpointed the top referrer, dig deeper into the data. Examine the behavior of the visitors from this source. Are they spending a lot of time on your site? Are they visiting multiple pages? 

Are they converting into leads or sales? Understanding these metrics will help you assess the quality of the traffic and the effectiveness of the referrer. This analysis isn’t just about numbers; it’s about patterns. 

If a particular source consistently drives valuable traffic, it’s a sign that your efforts in that channel are paying off, and you may want to allocate more resources to it. Conversely, if a referrer is bringing in traffic that doesn’t engage or convert, it might be time to reconsider your approach in that channel. Regularly analyzing your traffic sources allows you to make informed decisions, optimize your marketing efforts, and ultimately, drive better results.

I'm sure there was a post here on this a while ago but I couldn't find it. Anyway I was reading a thread on Reddit earlier all about it and the Google num=100 update has really hit Ahrefs hard, it seems.

Google quietly pulled the rug out from under a lot of third-party tools. Overnight, data that once looked “solid” now feels a bit soft underfoot!

It’s wild that multi-million dollar outfits like Ahrefs and SEMrush built their empires on scraping - something Google can (and clearly will) change whenever it feels like it. It’s a house of cards.

My web guy uses SEMrush daily. But we're not feeling confident with the data now and this Reddit thread seems to confirm what we're seeing! I'm starting to think only tools from Google are worth using. GSC, Analytics etc

DA was a decent estimated metric but I was never mad about it. We always see low DA domains on P1. It’s just funny that those those graphs were like gospel when they were really just… very pretty estimates.

If anything, the num=100 change was a good reminder not to put all your eggs in one basket.

But where does that leave us marketers whose life is data! lol

What the ad looks like

Instead of trying to make realistic people, Coke went all-in on animals: polar bears, a panda, a sloth, and a bunch of other critters. Styles jump around from semi-realistic to cartoony, and the movement can feel flat and odd, like 2D cut-outs nudged across the screen rather than properly animated characters. On the plus side, the iconic red trucks finally have wheels that actually turn this year. Progress, however small. 

Why the criticism

We’re in a world where AI video tools are getting scarily good, and that makes the bar higher. Compared with the slick deepfake-style clips we now see from leading models, Coca-Cola’s spot looks dated and uneven. It’s also arriving after Coke caught flak in 2024 for earlier AI holiday ads that featured uncanny faces and weird physics, so the scrutiny was always going to be intense. 

Who made it

Coca-Cola worked with AI studios Silverside and Secret Level, with around 100 people involved overall. Five AI specialists reportedly prompted and refined more than 70,000 AI video clips to assemble the final cut. That’s a lot of material for a single commercial and shows how manual AI production still is behind the scenes. 

Speed and cost

According to reporting referenced by The Verge, Coke’s chief marketing officer says AI made the campaign cheaper and much faster to produce. What once took about a year can now be turned around in roughly a month. That speed is clearly a big reason brands are experimenting so heavily with AI, even if the creative results are inconsistent. 

Here's the video. What do you think? 

https://youtu.be/Yy6fByUmPuE

Everywhere I look lately, people are sharing “500 free ChatGPT prompts for creators” or “The ultimate prompt swipe file for niche sites.” And sure, some of them are cool, but honestly, what’s the point?

Like, isn’t the whole point of AI that it can write a better prompt than we can? If I can literally ask ChatGPT to “write me 10 creative prompts for SEO keyword research,” it’ll do that instantly, and tailor them to my niche and my specific requirements.

So why would I need to copy-paste someone else’s prompt that may be written for a different use case?

The way I see it, prompting is just another word for thinking clearly about what you want. AI can help you do that better than any static list ever could. You can literally have a conversation with it to refine your idea.

Anyway, that’s my mini-rant 😅

Am I way off the mark here? Have you found free prompt libraries useful?