@wordsthatrank·Oct 6I asked ChatGPT for blog ideas for a made-up bakery, twice. Same model. Only the prompt changed.
Seven words got me titles any bakery in town could post.
After The Prompt Improver rewrote it, every idea came back with a search phrase and a reader ready to order.
I didn't touch the model. I rewrote the brief.
Save this for the next time ChatGPT hands you something generic.
#ChatGPT #PromptEngineering #AIWriting #ChatGPTPrompts #BlogIdeas
@wordsthatrank·Sep 26Does anyone have advice on how to create UGC content completely for free?
I’m looking for a genuinely free end-to-end setup with no paid APIs, no watermarks, no hidden paywalls, and no trial limitations.
Any tools, workflows, or tips that have worked for you would be greatly appreciated.
@wordsthatrank·Sep 22224 clicks to 2,140. One professional-services site, first 28 days against last 28 days, 16 months apart.
Average position went from 51.7 (page five) to 9.8 (page one). Clicks per day went from about 8 to 76.
I worked on the site the whole time, which limits what I can claim. For the first twelve months the work was mostly publishing and building: posts, service pages, links. From about May 2026 I started running the site's Search Console data through the prompts in my book, and the chart bends upward from there. Comparing the three months before that with the latest three months, clicks went from 2,662 to 5,831, with position holding near page one.
I'm leaving impressions out on purpose. Google over-counted them until 27 April 2026, so the early numbers are inflated and the comparison isn't fair. Clicks were counted correctly throughout, which means the 9.6x holds exactly as it stands.
One page didn't make it. A service page moved from position 62.3 to 48.4 and stayed on page five: thin page, competitive search. That is in the write-up too.
The client asked not to be named, and there are no revenue or backlink figures here. It is one site, and other work happened alongside. This is a record of what the data showed, not a promise.
The full breakdown is Chapter 26 of my Search Console book.
What did your clicks do over the last 28 days compared with the same 28 days a year ago?
#SEO #DigitalMarketing #SearchConsole
@wordsthatrank·Sep 18What's your online income this month? 😅
Reply with your number. No judgment (some judgment).
@wordsthatrank·Sep 18Selling online on Whop. What's your biggest struggle right now? 😅
@wordsthatrank·Sep 17🚨 One prompt. Two very different results.
A one-line prompt and a 15-component brief.
Same model. Two minutes apart.
The one-liner returned 10 blog ideas that could fit any bakery on earth. 🥱
The brief returned 10 ideas that fit one specific bakery, with search intent attached to each. 🎯
I’ve worked in content and SEO since 2011 and trained more than 40 writers.
The pattern holds:
A model fills every gap you leave with the most average answer it knows.
So I put the framework into a free guide 📘:
• Why prompts return generic answers, with a before-and-after example
• The 15 components of a strong prompt, including what each one does and why the model needs it
• 5 mistakes that flatten output, and the component that fixes each one
• The Prompt Improver: paste your rough prompt at the bottom and get a rebuilt version back
• A worked example you can copy and use
📄 13 pages + the prompt as a text file.
🆓 Free to use.
Works with ChatGPT, Claude, and Gemini.
💬 Which component do your prompts skip most often?
Mine was stakes.
My output stayed polite and forgettable until I fixed it.
