Slav Hurmanau

Trying to build something useful 😅

[Slav Hurmanau @lucky72o]

I measured my "blog vs no blog" experiment for the second time today.
Week 1 showed Google noticing the blog.
Week 2 is the first time an AI engine recommended it.

Two identical products, two weeks old: Gripelens has a blog with 8 posts, Quibblemap has no blog.

What happened in week 2:
- I asked Perplexity 33 questions about review analysis tools. It named Gripelens as "the closest fit" 4 times, and every time it cited the same blog post, never the home page.
- Google showed Gripelens 100 times and sent the first click. Quibblemap: still 0 impressions, 0 pages indexed.

ChatGPT and Google AI Mode still name neither product.
I will publish the whole "week 2" write-up tomorrow.

[Slav Hurmanau @lucky72o]

The video: https://t.co/KAtL2BMfHV

[Slav Hurmanau @lucky72o]

I spent the last few days recording the second video for my blog vs no blog experiment.
It still takes me forever to record the video. I was supposed to publish it on Friday with the results of the first week. It's Sunday evening, it's only going up now, and tomorrow I'm already running the next measurements.
Delayed a few days, not ideal.

Still, it took less time than the first video.
I hope that by video 10 the whole thing will be way smoother and the quality better too.
I was making this one for quite a while but still in a rush: the quality is not great and my speech is not great either.
I'm publishing it anyway, and I'll come back to it in a year or so to see the improvement.
Hopefully.

[Slav Hurmanau @lucky72o]

@mikedutuandu Partly. One of the two indexed posts links to both alternatives posts, the other one doesn't.
Good point, I'll add links from that one too to help the new posts get indexed.
Thanks for the suggestion, I'll keep an eye on it and report what happens

[Slav Hurmanau @lucky72o]

I published 4 blog posts this week for Gripelens, one of the two products in my "blog vs no blog experiment". The other product, Quibblemap, got none. Both launched on the same day, 19 October, with zero traffic and zero mentions anywhere.

Tomorrow I measure for the first time: Google impressions in Search Console for both sites, and whether ChatGPT, Perplexity or Gemini name either product when asked about their categories.

My guess is zero across the board. Six days is too early for either. But lets see

[Slav Hurmanau @lucky72o]

@TanyaDonska Thank you for the offer, maybe we can do this at some point in the future. Currently I'm focusing on other things

[Slav Hurmanau @lucky72o]

I measured my "blog vs no blog" experiment for the first time today. I expected zero across the board for some time. But surprisingly I can clearly see the difference already.

Two identical products, one week old: Gripelens has a blog with 4 posts, Quibblemap has no blog.

What Google did in week 1:
- Gripelens got its first search impressions on day 2. Quibblemap has none.
- Googlebot visited Gripelens four times as often.
- Asked "what is Gripelens", Google AI Mode describes it correctly. Asked "what is Quibblemap", it guesses a crypto tool.

In 33 questions about review mining tools to ChatGPT, Perplexity and Google AI Mode, neither product came up once.
Zero clicks and zero signups on both.
I will try to publish the whole "week 1" write up and stats tomorrow

[Slav Hurmanau @lucky72o]

@volosatovde Partly, yes

At day 7 Google had indexed 2 of Gripelens's 10 pages, both blog posts, with 2 more crawled but not indexed yet.
Quibblemap had 0, Search Console still said "processing data".

I will measure again next Monday

[Slav Hurmanau @lucky72o]

I published the week 1 write-up of my "blog vs no blog" experiment on my blog.

Yes, a blog post about a blog experiment.

I know people don't really read blog posts, but I think the numbers in it are quite interesting.
Not sure about others, but for me, seeing a new site get crawled by Google in real life, and seeing the difference between two sites, is quite eye-opening.

The number that surprised me most is Googlebot. The week before launch, with both sites behind a login page, it visited each of them 14 times. In the first week after launch it came 90 times to Gripelens and 21 times to Quibblemap.
Every other crawler, GPTBot included, behaved the same on both sites.

The full numbers, where each one comes from, and the one question I changed in my prompt panel:
https://t.co/3eZNvK5ItL

[Slav Hurmanau @lucky72o]

Can we please see the real examples of unexperienced people following the suggested steps and getting clients?

The theory sounds great! Would be great to see real examples

[Slav Hurmanau @lucky72o]

The two older ones, and neither of the alternatives posts.
The ranked list of 7 tools got 7 impressions, the "are G2 and Capterra reviews fake" post got 6.
Those two are also the only Gripelens pages Google has indexed so far.
The two alternatives posts are crawled but not indexed yet.
All the numbers for both sites are here if you're interested: https://t.co/3eZNvK5aEd

[Slav Hurmanau @lucky72o]

Agreed, and I changed it today.
Quibblemap's title, headline and description now carry the same key words as Gripelens: competitor reviews, complaints, counted, quoted.
The sentences stay different on purpose, so the two sites don't read as duplicates to Google.
Gripelens is unchanged, so the week 1 numbers still hold.
I'll note the change in the week 2 write-up.

Thank you for the suggestion!

[Slav Hurmanau @lucky72o]

Agreed on impressions first, that is exactly what I measured this morning, and they showed up sooner than I expected: 13 impressions for Gripelens on day 2 to 6, all on the blog posts, none for Quibblemap. Numbers here: https://t.co/t2IOS2ybmp

So far I published 4 posts, all aimed at founders researching competitors.

Two listicles aimed at the category, two "alternatives to X" posts aimed at people already searching for a tool

[Slav Hurmanau @lucky72o]

I measured my "blog vs no blog" experiment for the first time today. I expected zero across the board for some time. But surprisingly I can clearly see the difference already.

Two identical products, one week old: Gripelens has a blog with 4 posts, Quibblemap has no blog.

What Google did in week 1:
- Gripelens got its first search impressions on day 2. Quibblemap has none.
- Googlebot visited Gripelens four times as often.
- Asked "what is Gripelens", Google AI Mode describes it correctly. Asked "what is Quibblemap", it guesses a crypto tool.

In 33 questions about review mining tools to ChatGPT, Perplexity and Google AI Mode, neither product came up once.
Zero clicks and zero signups on both.
I will try to publish the whole "week 1" write up and stats tomorrow

[Slav Hurmanau @lucky72o]

@JohanesSurya You are probably right
Is there a way to make the boring stuff more fun? 😀

[Slav Hurmanau @lucky72o]

Every day I add more and more tasks to my already overflowing task board 😅

So many experiments I want to run and projects to build.
So hard to stop exploring new things.

What is the right balance between sticking with one project and experimenting with new things? 🤔

[Slav Hurmanau @lucky72o]

I run Claude for everything
Each step has its own agent, and I'm still playing with the models and effort levels per step: Opus 5.5 for the editor, writer and research, Sonnet 5.5 for the reviewer, low effort for the humaniser and the visual check.

The facts file comes from the editor

[Slav Hurmanau @lucky72o]

Fact checking is one of the most important steps in my blogging workflow

My workflow builds a facts file before it writes a post: every number, quote and named example from the research, each with its source URL.
The writer may cite only what is in the file.

On the image attached is a facts file for a Gripelens post on "fake reviews on G2 and Capterra" (Gripelens is one of the two products in my blog vs no blog experiment).

- Seven claims rejected
- Six are numbers from other websites with no source behind them
- The last one is different. Two vendors' reviews came in suspicious bursts. The plugin still would not let the writer call them fake: suspicious is not proof

[Slav Hurmanau @lucky72o]

This is my first ever attempt to record my voice and face. I fee extremely embarrassed to publish the video, but I will do this anyway

https://t.co/Yt0mpTHPwV

Just want to come back here in a year and see the improvements

[Slav Hurmanau @lucky72o]

Step 25 out of 25. My first YouTube video is up: the rules of the blog vs no blog experiment, where Gripelens and Quibblemap stood on day 0, and what I expect after six weeks.

Recording took a day, editing took two. Not sure how people do this every week 😅