These are things I have personally tried and tested and am happy with the results. You can always improve it. You can make things better. But it gives you a good taste of what can be done in vibration coding. These are things that I accomplished in maybe 15 minutes, half an hour. It’s pretty easy to take the first steps and say, “Oh, this works.” Maybe you want to make some improvements and refine the code and what you expect.
But I only ever ask ChatGPT. So if a page’s code doesn’t work, I’ll say, “I got this error. Can you fix it for me?” So you get a new piece of code, then you try the new one, and it goes on forever.
Tag matching
So the first one is tag matching. I wanted to match certain CTAs to a certain page.
I simply had a huge volume of pages that I didn’t know where to start. Then, I loaded the tags into one column and the URL and embeds into two other columns and asked ChatGPT to match them. Well, I uploaded the code to Google Colab and this was done via vector embeddings and cosine similarity.
Entity trust tracker
Then maybe you want to create an entity trust tracker. We all know that Google works with entities nowadays. They have a tool where you can just enter any word or phrase and it will give you an answer to say whether Google understands it as an entity and how strong the trust is in that entity.
Maybe you want to monitor your brand name. Or in my case, at one point, I was obsessed with keeping track of my name and checking if my knowledge panel was there and if the trust was increasing. So I created something in Google Sheets that would simply ping every day and see and save the security level that was directly on my Google Sheets.
I’ve done this once and it’s been working every day for a year.
hreflang match
Maybe you don’t want to do manual hreflang matching, or at least not from scratch. You can also create something via Vibe tagging and just load the embeds from the original source and the embeds from the pages I want to match.
The results are quite good. In my case, there are pages in many languages and it works. It doesn’t have to be just English or just the same language. It gives me a great draft. So I can do a spot check and say, “Does it make sense for this page to link to this page?” So it gives you a few steps out of the way right there.
Content decay tracker
Maybe you want to observe the decay of content.
It seems like everyone is feeling the effects of the decrease in traffic these days. Maybe you want to see which pages have lost the most traffic over time. So instead of going page by page and seeing, well, two years ago, it had this traffic, and now it has this traffic, you can just do all of this at once and get in bulk which pages are performing better, which pages are performing worse, and how much better or worse they’re doing over a certain period of time.
This is just manual labor. It’s just like processing data. The real work, the exciting part, is how you’re going to figure out how to get those pages or content back on the right track.
Find related pages
Finally, maybe you want to find related pages, which are also a combination of vector embeddings and cosine similarity.
Then you upload your list of pages and embeds and it will find many matches. There is an article on the Moz blog that explains how to do this and we can also link to the post.
That’s all we have for today. I’m Gus Pelogia and I hope you’re excited to program some of your own SEO tools too.
It will save you a lot of time on Fridays or any other day of the week to spend more time doing the things that really matter. Thank you.
