Improving the indexing of pages in Google using GTP3

Improving the indexing of pages in Google using GTP3

Often, the owners of online stores think that after investing a lot of money in the development of the site and filling it with products, they will immediately receive hundreds of orders per day. The profit will be calculated in millions, or even billions. But, as a rule, it is difficult to get a positive result without proper work on the site.

Why do some young online stores immediately get good search traffic, while others do not have results even after a few years? It’s all about indexing pages with the Google search engine. If the documents do not get into the index, the site does not bring traffic and customers.

Today we will try to figure out what site indexing is, what it depends on, and how to properly use the free GTP3 tool to improve the index of an online store.

I, Ivan Zhdanovich – SEO and PPC. I manage the IT startup Seo House[ссылка удалена модератором]. In my portfolio there are many cases in which I managed to bring IT projects to the first place in competitive topics. I run my blog, teach at universities and regularly hold webinars.

What is site indexing?

Site indexing is the process by which search engines such as Google scan and analyze the content of web pages for inclusion in their database (index). When a site is indexed, it becomes available to display search results for relevant user queries. Indexing allows search engines to efficiently find and offer relevant search results.

Everything seems to be quite simple. I created a site, installed Google Search Console and Google Analytics counters, and all pages were indexed. But this is, of course, an ideal case. Unfortunately, this happens very rarely. When working with clients, Seo House often sees a huge number of unindexed pages in search, like the image below.

Image indexing pages for an online store.

One customer’s image shows over 4,860 pages that were not indexed by Google. And less than half – 3,430 documents – are indexed.

Reasons for poor page indexing

If your site is indexing poorly in search engines, it could be due to several reasons. Here are some of them:

1. Incorrect configuration of the robots.txt file: The robots.txt file is used to tell search robots which pages should be indexed and which should not. If the robots.txt file contains errors or is not configured correctly, it can cause pages to be indexed incorrectly.

2. Duplication of content: If your site has duplicate content, such as the same text on different pages or using other sites to copy your content, search engines may have trouble determining which page to index.

3. Errors in the code and structure of the site: The presence of errors in the HTML code, the inaccessibility of pages to search robots or the incorrect use of meta tags can lead to poor indexing.

4. Not enough links to your site: If your site has few external links, search engines may refer to it less often to index new content.

5. Low quality content: If the content on your site is not unique, of low quality, or does not meet user requests, search engines may lower its priority in indexing.

6. Missing XML sitemap: An XML sitemap helps search engines understand the structure of your site and index all its pages. If you don’t have an XML sitemap or it’s not set up properly, it can make indexing difficult.

7. Slow page loading speed: If your site loads slowly, search engines may have limited time to index your pages, which can cause incomplete indexing.

The reasons listed above are usually eliminated with the help of a programmer or seo specialist, and only one of them with the help of a copywriter. This reason is the low quality of the content on the pages.

How can you improve the quality of the content on the site pages

The most common answer is to use a copywriter to write texts on all pages. This solution is suitable only for small sites. And what to do for online stores that have more than 100,000 pages. The price on English-language exchanges for 1000 characters with spaces ranges from $20 to $100. If you multiply the minimum amount by the amount of text, you get a significant cost of $2,000,000. And in the future, there is a question of the profitability of such a business. So what to do then?

The way out of this situation is content generation using artificial intelligence or GPT3. At the moment, there are quite a few services, but I will show one free service,, where you can try your hand at writing content in English without registration.

The image above shows an example of generating answers in English.

A case of content generation for an online store

The company Seo House has an interesting case about a foreign online store that did not want to be indexed. Seo specialists of this site have made many manipulations regarding the indexing of pages. They connected the Google Cloud API service, changed the domain to a new one, but it did not give results. In the image below, you can see the deplorable state of the site during the period of the client’s contact with our company.

Image of online store indexing

The image above shows that more than 175,000 pages are not indexed, and only 113 are in the search engine. At the moment, the site was more than 16 months old, and the “Google Sandbox” filter did not have to be sinned. The banal problem of this site was that the structure of the online store was copied from a popular marketplace. Categories and product cards were not unique, and Google did not want to waste resources on another clone of a popular site.

Work on the uniqueization of the site

In general, the site consisted of product and category cards. These pages did not want to be indexed due to non-unique content, so our team was faced with the task of making two types of content unique. To solve the problem with indexing, it was decided to add blocks with unique information. Content creation was done through the Openai Chat API. The tool itself has free limits, but for a large number of pages I had to use paid data packages.

Unification of product cards

The Seo House team started the first task, namely the uniqueness of the product cards. During the analysis of competitors, the non-uniqueness of the description and characteristics of the goods in the search results was found, but at the same time, everything was well indexed in large stores. It was decided to leave the product description and focus on creating additional unique blocks on the product card pages.

1) Adding feedback block to product cards.

The developer was tasked with adding a feedback block to each product card. Each page must have at least 3 reviews, which were generated automatically through the Openai Chat service API.

The task in the Openai Chat service for all reviews should be: “Form the text for the review. The review should include product features. Too watery constructions should not be added. More specific facts about the product. The review should be between 300 and 500 characters. Add rare words. Use rare words. don’t use “I recently” words. Create reviews from real people“.

The query text should be: “Review + “H1 Header”” where the value of “H1 Header” is the dynamic value of the H1 header of each product card.

As a result of the optimization of the product cards, a block with reviews was created.

An image of the feedback block on the product card

Reviews were generated both positive and negative to make them seem more natural to the user.

2) Adding the “Questions and Answers” block to the product cards.

The client had a useful “FAQ” page with popular questions and answers from users on the site. Therefore, it was decided to create a “Questions and answers” block on each product card. Mark up the FAQpage micro-markup so that the structural data is correctly displayed in the Google search results. And with the help of the Openai Chat API service, generate unique questions and answers for each product page. The number of generated questions had to be at least 3 for the markup to work correctly.

The Seo House team prepared a task on the Openai Chat service for a programmer to automatically generate content through an API for an online data store. The task sounded like this: “Compose in Ukrainian in English is based on the facts of the text, not using the most common words and phrases, but using words and phrases that are partially used and appear incomparably, unique, relevant information. Use text formatting. Body text itself goes on after the colon:”

The result was a FAQ block with unique information.

Image of the “Questions and Answers” block on the product card

Due to the large volume of data generation, it was decided to fill at least one answer to the question, which added uniqueness to the page and improved ranking in search engines. As a result, more than 300,000 unique questions and answers were generated in 7 months.

Unification of categories

In order to make the categories unique, a decision was made based on the competitors to add text from 1000 to 2000 characters with spaces. To improve the quality of the content, it was decided to add the parameters of each category to the task, which could be automatically obtained quickly and without using people. The parameters were as follows: the name of the category, the possibility of delivery to one or another US state, the number of products, minimum and maximum prices. The result is pretty good text in the image below.

Image template text for one of the categories

After the introduction of the “Questions and Answers” block on the product cards, users began to use this functionality. The client suggested adding such a block to all categories of the site.

Image of the FAQ block for the category.

Image of the FAQ block for the category.

The image above shows the implementation of a question and answer block for a category.

Results from unique content

Work on unique content lasted 7 months. The following worked on the project: a programmer, an SEO specialist and a copywriter. After 7 months, our Seo House team managed to index 80,900 pages on the site. Pages that lacked unique content were not included in the index.

Image of page indexing after 7 months of operation

If we sum up the content generation, we can safely say that our work together with artificial intelligence brought a positive result, and saved the client $ 2 million on content alone.

If you look events in Mykolaiv –

Related posts