Creating content via template and Data driven content Pages (11)
Creating content via template and Data driven content Pages
Hello everyone,
I hope you had a nice week!
As in one of my previous posts discussed I am a coder. My strength is in developing solutions, setting up IT architecture and creative solutions to create products in a fast way.
I am currently creating an automated newsletter that focuses on investing and finance. The content of the newsletter is created everyday and also automatically send out to my list.
Atm I am learning all the marketing site of the business and listen to a lot of podcasts, besides reading the stm forum.
I found the following interview quite interesting: https://www.nichepursuits.com/matt-paulson-marketbeat/ (btw I am a huge fan of this guy!!!)
Matt Paulson is talking about data driven content pages. An example is: https://www.marketbeat.com/stocks/NY.../price-target/
After researching I found that he is also generating a lot of content pages algorithmically, for example: https://www.tickerreport.com/
All headlines & content is automatically generated. If anyone is interested in this topic, it's basically called in the academic field "natural language processing" --> data-to-text.
Here are my questions to you:
1. Automated Content
Does this still work in the field of SEO? Isn't google against automated content creation.
2. Translation
I read in another post that content, that get's translated from one language to another, such as translation with deepl still works and ranks in google. Have you seen any evidence in the wild?
3. Data-Driven Content Pages
Could these types of pages be used for all niches. F.ex. affiliate sites that are data driven? How does google compare these types of pages to completely written pages?
4. Acceptance of Google
As these pages seem to rank in google. How does these type of creation compare to spinning vs creating content with a template?
I really appreciate your replies!
Interesting stuff - right up my street.
While I can't give you any definitive answers, heres some things I know from research, and from testing this (albeit a few years ago).
1) While big G's black box algo says the penalize duplicate content, there is many examples the algo is not as strong as they'd lead you to beleive.
A few years ago you could get away with short snippets with backlink to original article, not sure if thats still the case though.
Who gets slapped with the duplicate content penalty isn't always the guilty party - larger sites with good domain authority can bully out small sites.
2) Spinning content was a big thing that worked years ago, and G algo got better at detecting it, but with the GPT3 algo appearing recently I can't imaging they could detect that.
GPT3 is definitely one to watch, especially on the next election/fake news front. Its limited access research for now, but I can't imagine it'll stay that way for too long.
A bot created with the API managed to go undetected to most on Reddit for a while, and a blog written entirely with GPT3 is pretty believable unless you knew.
On translation, it's possible to reuse content in different languages to target untapped niches.
3) I think the best strategy would be to employ a GPT3-level content surrounding scraped re-usable content and it'd be pretty difficult for G to tell the different.
Especially in e.g. Crypto field where BTC prices are posted on multiple platforms constantly.
4) Google has gotten way better at detecting spun content. I'm not sure if a newer spin platform exists, one that combines a grammarly-type element with word/sentence rewrites - but if it does, I'm sure it could fool Google.
Recent algo-update may have improved all this, but only time will tell.
However running traffic to your email opt-in and sending autogenerated content sounds like a solid plan.
Have any other podcast recommendations?
Creating content via template and Data driven content Pages
My goodness - what a fascinating topic! Thank you both for opening my eyes to what's possible!
I have little to contribute - last I did SEO with focus was years ago, when spun content and link spamming using software was enough to rank.
Except: I'm thinking how awesome it would be, for someone with good coding skills to build an automatic content generator and offer it as a saas. If it can fool google, it can turn into a huge cashcow.
I'm thinking: Specify a niche and some keywords, and the generator spits out an entire blog already scheduled for posting twice a week on wordpress.
You can practically spit out hundreds of such sites or more a day with next to zero effort. Back in 2008 I was using a similar service to make easy money. I fed the scraper a ton of keywords and it generated one page of scraped content per keyword. The content was garbage but back then that didn't deter them from being ranked and getting a tiny bit of traffic each. And when you have millions of such pages, revenue adds up.
The only costs were the scraper service, domains and hosting. Once I had enough sites with enough pages on each, they paid for themselves and more.
Back to the new, yet-to-be-built, state-of-the-art content generator: If the content quality is good enough to build a fan base, great! If not, it would still be pretty kick-ass to monetize these sites with adsense or clickbank or other affiliate offers as well.
Amy
Sent from my iPhone using STM Forums
I've seen quite a few auto-generated content sites ranking high. In some niches, it's actually quite natural to present content in such a way...
Finance is actually a good example... there are several sites that simply show the prices of stock, forex or crypto... all autogenerated and still ranking high. For example: https://coinmarketcap.com/
Another example would be blog aggregators, they generate the content from rss or api, then show snippets from the updates and categorize them... I've seen a few ranking well too. For example: https://blog.feedspot.com/food_blogs/
Then there are cold-data-based websites that pull all kinds of data from various sources and present them on a website, for example something like https://www.city-data.com/
There are many examples out there... but it might be hard to outrank these long running webs, the content is one thing, the other SEO factors is another. But generally speaking, autogenerated content can certainly rank high.
Creating content via template and Data driven content Pages
This “mass site” practice is pretty common with pay per call and lead gen offers. “Doorway Pages” are created for every city etc and filled with spun content to rank in google. “Best plumber in Austin Texas”, “Mould Removal in Dallas Texas” type keywords
Because it’s viewed by google as very low quality you’ll have some issues getting all of your site indexed, so you’ll need to build out a huge amount of sites (which brings a lot of other issues PBN builders encounter, like diversity in hosting etc)
I’ve had success with this approach, but it became almost a full time job keeping up with google updates and the technical aspects. Easier and more stable ways to make money IMO.
Sorry for bumping up an old thread. But I found I could add something to it. Moderators, I apologize in advance if this is bad practice.
Originally Posted by marpre001
Data to Text can be achieved by a simple set of conditional statements (if/else).
NLP goes far beyond that. NLP takes data and tells a story in Human Readable Terms.
And vice versa, it can take natural language inputs and turn them into Computer Readable Terms.
This is what is employed by a few of those "Content Generators" that promise to generate "Human Readable" content.
I am not sure if you've been following OpenAI's GPT3 - but you should. It has a lot of promise.
Now moving on to your questions -
__________________________________
1. Automated Content
Does this still work in the field of SEO? Isn't google against automated content creation.
__________________________________
There's a difference here.
The industry saw a spike in what was "Aggregators" such as Digg and Delicious and decided that automated content works. They did work for a while.
When you use the term "Automated Content" - you mean content generated using Data. When the term is used as an umbrella term, it generally means "Auto Blogs" or ones generated through a content generator in subpar language.
So let's say you scrape a directory of Plumbers in California
Instead of a regular directory - your content would be.
<Name> is a plumbing solutions provider in <CityName>, California. They have been in business since <Year> and are pretty well known in the area. They can be reached at <Phone> and <Phone2>
Acme Plumbers is a plumbing solutions provider in Los Angeles, California. They have been in business since 2003 and are pretty well known in the area. They can be reached at 222-222-222 and 333-033-333
<Name> is a plumbing {solutions|service|facility} {provider|firm|company|business} {located in|situated in|in|from} <CityName>, California. They have been {in business|operating|active} since <Year> and are a pretty {well-known|respected|highly-rated} business in the city. They can be {reached|contacted|called} at <Phone> and <Phone2>
2. Translation
I read in another post that content, that get's translated from one language to another, such as translation with deepl still works and ranks in google. Have you seen any evidence in the wild?
__________________________________
Translation works, yes and no. It depends on your target audience.
Look at the language speaking demographics and you can decide on English to > Whatever Language
For instance, Canada has a huge French speaking population. So it should technically work great.
__________________________________
3. Data-Driven Content Pages
Could these types of pages be used for all niches. F.ex. affiliate sites that are data driven? How does google compare these types of pages to completely written pages?
__________________________________
Theoretically yes.
You're only limited by the data you have. If you're able to get remotely relevant data to any niche, you can generate a site with pages.
Google cares only about the value you add. So when building a page, it should be something that adds value. If you can construct an API that updates regularly - then it's what you should do.
I gave you the directory as an example. It can be used for any large data set. Business Listings. Finance Information. Crypto Information.
Coupon sites do it too. Take a coupon feed, isolate the brands, fetch/scrape more information about the brand, create brand pages, implement the coupon feed. BAM! Value Added!
__________________________________
4. Acceptance of Google
As these pages seem to rank in google. How does these type of creation compare to spinning vs creating content with a template?
__________________________________
It comes down to user-intent - nothing beyond that.
In most cases, you're showcasing data. And getting interested people on the site for you.
Acceptance of Google is not the question, acceptance by people is.
My core advice to you would be data enrichment. If you find a single data source, merge it with other sources and enrich that source with more and more information. So that your key validations are enhanced.
Originally Posted by contentistic
Thanks for your recent contributions, really interesting!
It's perfectly fine to bump older threads when there is solid info do add, just as jeremie already posted above. And I have to say you really posted a valuable reply, thanks a lot for that.
GPT-3 is very interesting, just did some reading about it as a result of your suggestion. And looks like Microsoft spotted the opportunity :
Originally Posted by jeremie
Originally Posted by matuloo
Originally Posted by matuloo
Consider this, the GPT-3 Neural network model is so large that it cannot be easily moved from a cluster of machines it was originally trained on. It needs a ginormous infra costing millions of dollars.
But the output - damn!
I got a chance to play with it with a friend doing her CS in Stanford - who had access to it. We built a text to CSS, SQL, JS thing and it was fun while we had access.
To give you a live example. Here's an app called thoughts by a guy on twitter.
https://thoughts.sushant-kumar.com/startup
https://thoughts.sushant-kumar.com/affiliate
https://thoughts.sushant-kumar.com/money
Just replace the word at the end and it will generate a unique tweet. Isn't that friggin' cool?
Soon enough, we won't be able to tell whether we're talking to a person or a bot, not sure if it's a good thing though Reminds me of an episode of the "Black Mirror" series : https://en.wikipedia.org/wiki/Be_Right_Back
Have you seen that?
Originally Posted by contentistic
Home > Free Traffic Sources >