[Journey] money sites experiments

[JSRLove]

Hey everyone!

After a period of silence, I've decided to embark on a new journey. Don't expect any groundbreaking techniques or inspiration from my outstanding results just yet, because I'm just starting this SEO path... again.

A bit about my background​

The first time I heard about making money with affiliate websites was back in 2011. Being a budget-limited student (honestly, there were times I had no money for food), I handled everything myself. Despite not achieving significant results, I earned around $500 after selling my three websites.
However, I discovered my passion for web development during this time. So, I started as a PHP/WordPress freelancer, making 5-10 times my university scholarship and feeling like I had conquered the world.
Then came my first offline job, followed by more. Projects became more serious and complex, and I grew both in salary and position.
Now, I find myself in a huge company that most of you probably know. However, I'm grappling with a 5/2 working schedule and don't see suitable options to significantly scale my income.
So, after a 10+ year break, I want to give affiliate marketing another shot.

My journey goal​

I don't expect my first, second, or nth project to make me a multimillionaire. But I'm confident that consistent learning and hard work will help me discover methods that work for me, and then I can scale them.
Due to the dynamic nature of the SEO sphere and time constraints, I find your questions and feedback invaluable.
I also learn from my mistakes, so you'll know what doesn't work

What I have now​

I'm running two websites with AI-generated content.

The first one was based on the principle: just do something and see what you did wrong.
I tried to explore the tech stack I'd use, built some infrastructure around it, and checked how difficult it was to index such content.
I found a simple solution that cost me $4 for two websites and dashboard (i can easily add up to 10 sites without cost increase). It supports something like a CI pipeline and shows good results according to page speed insights.

Currently (after 1.5 month), it has 500 pages in the index but generates almost no traffic. Why it happened?

• Sh*t content
• Huge competition (niche too general)
• Unclear website structure

In addition after 3 weeks after launch i found issue with url and google drop insights after i fixed it :facepalm


The second site was deployed Nov 17. What i did differently?

1. Fixed all tech issues before launch and perform additional check that everything good
2. Made niche research. Main keywords have average competition +-15 according Semrush. However i underestimate top 10 competitors
3. Because of 2nd point i have clear content strategy and site structure
4. I've rebuilt content generation strategy

Have no idea how to evaluate this result, but thats what i have for now:

Whats next?​

1. Until i have scheduled content i want to improve second site usability
2. I need to build next category articles structure to be able generate content
3. Improve internal linking for both sites
4. Fix all errors/warnings/notices from Ahrefs audit

Well. It's already a lot of text. Thanks everyone who read it. Feel free to ask any questions or share you thoughts

 

[GeeLark AI]

Goodluck with your journey dude

 

[onlinemarketer12]

$4 for two websites ? How much articles on the website ?

 

[JSRLove]

Goodluck with your journey dude

Thanks, mate!

$4 for two websites ? How much articles on the website ?

Seems i mislead you with pricing
4$ monthly - price for infrastructure. (shared BE + multiple FE)
First site has 3600 articles with 600 words average length. It was generated through 3.5turbo model and cost me 12$
Second has 80 articles (3.3k words average). Generated with gpt4 api - 10$

 

[JSRLove]

Hey BHW!

After spending almost 20 hours with my second website i finally ready to share with you some updates

In few words:

Never spend 6 minutes doing something by hands when you can spend 6 hours failing to automate it

It's not actually true about this week, but close enough

Dive into details​

New category​

The major part of the time took building new category for my posts:

• Page design
• Building prompt for ChatGPT that satisfy me
• Parsing the output and display it on page (i'm using GPT JSON output to have control over each article section)
• Testing, debugging...
• Building the prompt for images generation (they are rlly cool ^^)

Also i decided to add some interactivity for each article. To avoid niche spoiling lets say it something like advanced quize.
The biggest mistake was writing everything with Vanilla JS. I'm trying to keep my pages as thin as possible, so instead of adding phaser, react or even jQuery i decided to hark back and write everything with pure js.

Can't say i'm fully satisfied with the results, but it looks pretty good, especially on desktop.
It was an attempt to bring some "value" to users (and little hope that someone share the link )

Speaking of numbers:

• Articles generate - 20 (however the category design applicable to other similar categories, so i'm able to generate at lest 100 new)
• GPT 4 model costs - 1.42$ (including tests)
• Dalle - 0.82$

Usability improvements​

• Added table of contents for most of the pages
• Added button that scroll you page to the top
• Improved internal linking ( automatically map some keywords inside the text to existed articles)
• Added "similar" posts at the bottom of each page

Some technical "improvements"​

To be honest i don't know what happened with impressions. Maybe it's my improvements, maybe some Google update:

Next day after deploying my updates i've opened console and saw this ugly impressions drop.
Anyway. What i did:

1. Instead of having single template for page title i've created multiple ones to fit 70 chars length.
2. The same for description
3. Updated some titles to match the title Google choose for SERP
4. Corrected the order of heading tags (mostly replaced h4 to h3)

As a result i see green "excellent" badge in ahrefs

Other​

Added 2 links from wiki. One of them in "Refferences" section, the second in "External resources".

BTW: How long it takes to see them in ahrefs / GSC ? Do i need perform any actions?

Few words about first site​

Actually i have nothing to share, since i've opened it only few times to check everything fine. Everything fine with technology, but not fine with traffic. Here is screenshot from GSC:

No clicks, few impressions from 500+ indexed articles

What next?​

1. Because of there are low value AI generated websites with niches i'm not rlly interesting in i'm not going to buy "good links for 300$". I'll try to explore my competitors links profile and try to obtain for free as much links as i can
2. I still want to somehow rebuild the internal structure of first site. Perhaps i'll see onother drop of impressions, but who cares? The site has only few clicks per weak

I never did a linkbuilding, so can't expect any results. Probably most of the time i'll spend to research. Let's see

Thanks for everyone who read it
I also want to apologize for my english (especially for grammar). I've spent too much time with ChatGPT, so won't open it at least for few days.

See ya!

 

[Akhile]

Alright, good luck on the next phase. I'll be watching out for the outcome.

 

[TechEinstien]

Good luck. Don't loose hope.

 

[awesomewebsites]

Even the job you have is paying good so you can invest that money in different mediums like crypto and stocks. Invest in crypto as per risk appetite only. Like if you are young, you can invest more in crypto but if you are near retirement, stocks will be a better option.

 

[PennyLeonard]

It is the best choice to relax your body and mind, broaden your horizons and increase your knowledge. Come on, your future has endless possibilities! God bless you！

 

[BitSpider]

following

 

[JSRLove]

Me again

First of all - thanks everyone who comment or just watch. You, guys, make this journey much easier by adding motivation. Also i've started feel some commitment, so can't give up.

Even the job you have is paying good so you can invest that money in different mediums like crypto and stocks. Invest in crypto as per risk appetite only. Like if you are young, you can invest more in crypto but if you are near retirement, stocks will be a better option.

Fundamental advice. Dunno am i young, but i'm already dream about retirement
I'm 31. Regarding the crypto - i believe that it will be the major part of our future, but we are not ready yet. So i've put in some fundamental coins +-15k. It's not the amount that make me rich, but i'm not afraid to loose it. Waiting bull run to fix some amount of positions, but won't grow it more, cause i start feel uncomfortable.
Stock - awesome options. Though about it every few month, but it's kinda tricky to invest in my country :/ However after your message i started thinking about it again

Back to the projects​

I've started research competitors backlink. Found some websites allowed me post mostly nofollow links. However the main links mass come from guestpost for money :/

I've tried to google something like "High DR ******** backlinks lists". And holly crap...there are tons of sh*ty platform with awful UI, annoying popups and server errors. I've tried to put link for some of them and drop this "activity" cause i'll spend more for psychotherapist after such execution with dubious results.

However i found the platform with paid backlinks. And some websites looks like everyone able to just write the post, and receive the link.
I have some articles on medium and that experience much pleasant than using garbage dump like articles.h1ad for example

Corrected plans​

I don't like change the plans (to be honest i like it, but it leads to abandoned projects), but i won't suffer whole weak with such link building. I'll try to place link every day or two but will avoid the fanaticism.

Initially (before choosing the niche) i thought about localization mechanism. So ill translate whole site to my mother language. Want to check how it will rank in my country.

Also it should unlock me the future steps ( i though about micro SAAS based on this niche). But its an another story.

See you at Friday, with more detailed report.

 

[asasin]

what platform is this ?

 

[JSRLove]

what platform is this ?

icopify dot co

There is no advertisement, cause i can't say anything about link quality. Purchase few for initial deposit ( they ask you to deposit 10$ to be able see the domains )

 

[JSRLove]

Hi there

As promised here is my Friday report

Main focus (translations challenge)​

As i wrote in previous post - i've worked on localization.
Since i'm not using any CMS and my content consist mostly of JSONs objects it's not rlly trivial task.
What i've tried:

1. Of course my first thought was about ChatGPT. Since 3.5 model now also able to provide JSON output i've decided to start with it. It's not a new content, just translation, so i had an assumption that it's simple task. I was wrong....it works well whenever you try to translate whatever language to English. But in my case English -> Ukrainian it produced sh*t
2. Ok. 3.5 doesn't work. So i decided to increase costs and try to use GPT 4. A bit better, but after trying to read one article i've understand that spending dozen hours to fix content isn't my goal.
3. I've took an attempt to find someone who will do it for money. But profreading 2 000 000 symbols of translated AI content seems too crazy in terms of price
4. Than i've tried to use google/bing/yandex translators - total shit
5. The last option i had is Deepl. I didn't start with it because it's not supported in my country. So after some search i found the seller and purchase an account for 2$. It allows me translate 20 documents (20kk symbols)

But the problem with Deepl - it doesn't support json. It want word document or plain txt.
So i wrote quick script that decompose my json into chunks, marked by special codes. Deepl translated the text and than, but codes i composed txt back to json.

Total expenses: 2$ Deepl account + 2.5$ burned with GPT tests

Some tech work
Since i'm pretty satisfied with translation quality in case of good ranking i'll do other locales. So I spent some time to make sure that adding new one takes less than day for whole site.

Current state
I guess that 95% work is done. The only things i need:

• Find cyrillic font that looks similar to my english one
• Double (or even tripple) check that i correctly setup all meta data
• Fix some css styles, since ukrainian words usually longer that english ones
• Deploy and waiting the good results in SERP )

Linkbuilding​

I wasn't so consistent with linkbuiding as i though a week ago since localization took much more time than i expeced.
Anyway. I see first results


Also i've created pinterest account. Uploaded few AI generated images. Have few impressions and few click. I definitelly won't spend my time to upload everything manually so will try to find/create automated solution

Organic traffic​

Nothing to share. The results pretty the same as week ago.
The only thing i noticed - keyword difficulty is total shit. I have 2 articles generated the similar way, with similar structure and published almost at the same time. Article with KD 47 (according ahrefs) ranked at 43 place in USA. Article with KD 1 - 54 place.

Future plans​

1. Release localization and check how it ranks
2. Build POC for my "SAAS" idea
3. Find the test auditory for testing SAAS

Thats it for today!

Many thanks for everyone who following!

 

[Steptoe]

What i've tried:

1. Of course my first thought was about ChatGPT. Since 3.5 model now also able to provide JSON output i've decided to start with it. It's not a new content, just translation, so i had an assumption that it's simple task. I was wrong....it works well whenever you try to translate whatever language to English. But in my case English -> Ukrainian it produced sh*t
2. Ok. 3.5 doesn't work. So i decided to increase costs and try to use GPT 4. A bit better, but after trying to read one article i've understand that spending dozen hours to fix content isn't my goal.
3. I've took an attempt to find someone who will do it for money. But profreading 2 000 000 symbols of translated AI content seems too crazy in terms of price
4. Than i've tried to use google/bing/yandex translators - total shit
5. The last option i had is Deepl. I didn't start with it because it's not supported in my country. So after some search i found the seller and purchase an account for 2$. It allows me translate 20 documents (20kk symbols)

But the problem with Deepl - it doesn't support json. It want word document or plain txt.
So i wrote quick script that decompose my json into chunks, marked by special codes. Deepl translated the text and than, but codes i composed txt back to json.

Have you looked at the MADLAD400 or M2M100 models? https://textsynth.com/documentation.html#translations

 

[JSRLove]

Have you looked at the MADLAD400 or M2M100 models? https://textsynth.com/documentation.html#translations

Wow, nice share. I'll definitely play with it

However quick check shows that M2M100 rlly similar to google/bing/yandex output
MADLAD400 - something between GPT3.5 and GPT4

But what is cool - if i understand pricing correctly their Stable Diffusion similar to DALLE2 by quality, but 4 times cheaper )

 

[JSRLove]

A little update for today: localised version deployed today. Already see 2 pages in index by "site:domain.name/uk" request
Continue working with SAAS feature MVP