Automatic content generation for Google SERP

Table of Contents:

Today we will discuss a fairly new, but already quite common, way of filling a site with relevant content. This material will be useful both for optimizers, who are looking for solutions to fill the site with large amounts of content, and for marketers/site owners - in terms of assessing the possibility of implementing the described solution on their own site in order to increase the share of low- and mid-range traffic.

When it comes to content - the first thing SEO-optimizers think about is where to get a quality copywriter who understands the subject, and how much it will cost 1000 characters. If the second question - depends directly on the quality of the found copywriter, then with the first question - there may be problems. Finding copywriters who deeply understand the topic of your business - can be a daunting task. At the same time, the inert market of "SEO-copywriting" is still a "conveyor" of SEO-texts that successfully fall under filters.

In this case, if we talk about writing content for large landing pages - from 10 000 to 1 000 000 - copywriting can take years and cost a fabulous financial investment. Where do you get high-quality content in large volumes from in this case?

Warning! The method of content generation described in this article is suitable for many sites, but not for all. To evaluate the depth potential and the possibility of generating content on your site - use the recommendations at the end of the article.

Content generation is not as bad as it sounds

Speaking of content generation - the first thing that comes to mind for experienced marketers is UGC (User Generated Content). UGC, as you can see from the abbreviation, is the process of user-generated content.

UGC (User Generated Content)

User Generated Content (oruser-generated content) is original content created by the brand audience - the users of your site. This is one of the basic concepts of Web 2.0, which was talked about a lot back in the 2000s, but now almost nothing is heard about. Nevertheless, this method of generating content is being used successfully by many brands.

There are several types of UGC, which are most in demand in today's online world:

• Questions and Answers (FAQ);
• Reviews;
• Reviews;
• Comments;
• Forum;
• Photos (custom);
• Reviews;

As practice shows, for companies from small and medium-sized businesses - it is not always profitable to use the ability to generate content by users on their projects, because:

• It is necessary to introduce an additional staff member whose responsibility would be to monitor and pre-moderate the content generated. Relevant for "Reviews", "Reviews", "Reviews", "Photos", "Comments", etc.
• This increases the load on the company's specialists who have expert knowledge and are able to give a complete and useful answer to users' questions. Relevant for "Q&A (FAQ)", "Online consultations", etc. - where user participation in the process implies feedback (response) from a specialist.

In addition to UGC there is an alternative way to generate content, we called it - YouGC.

YOUGC (You Generated Content)

You Generated Content is a method of generating content, which is reasonable to use in generating pages based on large amounts of patterned and structured data. In its pure form, YoGC is the output of an array of structured data according to certain pre-prepared templates. Officially, there is no such term, but it is quite relevant when compared to UGC, especially for SEO.

Possible types of YouGC:

• Generation of the main cascade of pages (structure) on the basic semantics;
• Generation of subsections by different types (devices, services, etc.) with linkage to the main node;
• Generation of subpages for each product or product group according to a set of characteristics (sizes, colors, and other characteristics).

This type of content generation has the following MUST HAVE requirements:

• Systematizability of data. Content generation by YouGC methodology is suitable only for systematizable data, which in addition to the "secondary" characteristics, presents a number of "primary" ("structure-forming");
• The presence of a large volume of "secondary" characteristics. The large volume of characteristics involved in the generation allows you to generate content in greater depth, to cover a larger part of the low and low-frequency semantics, to get a better quality content at the output;

These requirements are described in more detail in the examples below.

Advantages of this method:

• Page generation for commercial sites.
  Suitable for generating the main cascade of pages for commercial sites (basic product or service offers).
• No response from the brand.
  No need to interact with the user (as in UGC).
• Tight deadlines.
  Allows you to implement in a short time a fairly large layer of thematic landing pages for different types of queries (geo-dependent, geo-dependent, commercial, partly informational).
• Template elaboration!
  It's enough to make a change in a single template to implement edits on thousands of pages.

UGC

• wide coverage of topics (low frequency queries)
• topics of interest to other users, community building
• involvement, by encouraging other users to participate
• more suitable for generating the information segment of pages
• often does not integrate well into a commercial site (requires a response)

YouGC

• Template generation of commercial pages
• the possibility of greater depth of generation within a single subsegment
• Template elaboration of conversion elements
• Template LSI and backlighting
• Template development of low-frequency (and even HF-HF queries)

Examples of content generation

Consider some examples of topics that allow a LARGE or LESSER degree to implement the structure of the site in whole or in part.

Please pay your attention! Before you start, you need to clearly formulate your offer (product offer or service) and describe all the possible characteristics of your offer.

Topics "Tourism": airline tickets

The main product offering in the subject is a document confirming the fact of the contract between the passenger and the air carrier. This document is called "air ticket". And this product has a lot of properties which we can use in generating content.

Properties of an airline ticket:

• Country of departure
• City of departure
• Destination country
• Destination city
• Airline
• Number of connections (or direct flight)
• Transfer airport(s)
• Date of departure
• Month of departure
• Season of departure (summer, winter)
• Departure airport
• Arrival airport
• Type of aircraft
• Flight number
• Meals on board
• Service class
• One way / round trip
• Price
• Paid luggage
• Departure time
• And this is not a complete list...

As you can see from the list, there are a huge number of properties of an ordinary airline ticket. Highlighted in the list of properties - are structure-forming, that is, using these properties we will work out the structure of the landing pages of the site.

Pay attention: the more properties of your product offering you can formulate - the more high-quality generation templates you will be able to work out.

Initial elaboration of the structure

We assign a name (variable) to each structure-forming property:

• Country of departure = FROM_COUNTRY
• Departure city = FROM_CITY
• Destination Country = TO_COUNTRY
• Destination City = TO_CITY

For example:

• FROM_COUNTRY = from Germany
• FROM_CITY = from Berlin
• TO_COUNTRY = in the Netherlands
• TO_CITY = Amsterdam

In order to properly implement a sufficient structure, a semantics analysis must be performed. We will not go into the details of semantics analysis, because even a not particularly deep study of thematic queries gives an idea that most of the country occurrences refer to the destination (TO_COUNTRY), but among the queries there is also a share of occurrences to the departure point (FROM_COUNTRY). To minimize the number of generated pages with "weakly useful content" - we combine items 1 and 3 in one segment of the structure.

It is highly not recommended to generate a huge number of pages for all intersections of properties, as it will lead to a very inefficient use of crawling budget (from the word "crawling" - quota to bypass your site pages from a search robot), increased time reindexing the main cascade of pages, as well - difficulties in subsequent analysis of the effectiveness of the work performed. Also - in the race for "quantity" may suffer "quality", which will lead to the mass exclusion of landing pages from the index (in Yandex - with the notation "insufficient quality").

Sufficient structure in the subject "Airline tickets" looks like this:

• Main page - site.ru
  Main page - is a splitting page for all internal sections and subsections.
  • Countries - /countries/.
    Section which contains all countries with which we have flights.
    • Country - /countries/*TO_COUNTRY*/.
      End pages for each country.
  • Flights - /airline tickets/.
    Aggregates all cities withwhich there is flight service.
    • From city - /airline tickets/*FROM_CITY*/.
      The node of the structure that aggregates all the cities of departure (point "A" in the "from A to B" route).
    • To city - /airline tickets/*FROM_CITY*/*TO_CITY*/.
      End pages of the "from A to B" route.

Distribution of requests and working out of templates

Each generated page on the site must be relevant to a separate group of queries.

IMPORTANT: when working out the structure of the generated section - avoid semantic duplication.

In this example, the distribution of semantics on landing pages is as follows:

1. Home Page
   • The main page
     is a splash page for all internal sections and subsections.
     • buy airline tickets;
     • buy airline tickets online;
     • airline tickets online;
     • cheap plane tickets;
     • etc.
   • Countries
     an aggregation page free of semantics.
     • Country (e.g. Germany)
       • airline tickets to germany;
       • buy a ticket to germany;
       • airline ticket prices germany;
       • etc.
   • Airline tickets
     aggregation page, free of semantics. General queries related to the purchase of airline tickets can also be allocated to the page.
     • Airfare from the City (e.g. Detroit)
       • Cheap tickets from Detroit;
       • Buy a plane ticket from Detroit;
       • etc.
     • Airfare by route (for example, from Detroit to Berlin)
       • Detroit berlin ticket prices;
       • buy a ticket detroit berlin;
       • airline tickets to Berlin from Detroit are cheap;
       • etc.

As you can see from the examples of semantic groups allocated to the generated landing pages, the main potential in traffic generation is the section on routes - "from city to city". The template of this type of landing pages will require the most detailed elaboration.

Data collection for generation

In the subject of airline tickets - the main work before generating the structure is to collect the names of cities with links to airport IATA codes - all reservation systems work through IATA codes, and each city and airport has its own IATA code.

IATA code bases with the names of their respective cities can easily be found through a Google search, for example:

• Detroit = DTT
• Berlin = BER

It is also necessary to collect all the ISO 3166 country codes. This classifier provides a list of all countries with their two-letter (alpha-2), three-letter (alpha-3) and numeric codes:

To generate page content - short and full country names will be required, and alpha-2 or alpha-3 codes can be used to generate URLs, for example:

• http://site.ru/strany/aus/ - Australia
• http://site.ru/strany/aut/ - Austria
• etc.

It is necessary to save all the data in tables in the database, which will later be used to generate content and customize the page addresses (URLs).

Intent processing

The main commercial user intent is the request "to buy", which expresses the user's desire "to conclude a contract with the carrier". Immediately after working out the structure - look for the characteristics of this intent.

The characteristics of the intent always have an overlap with the properties of the product itself.

They can be divided into three groups according to their intension:

Why is it important to work out all possible intents in combination with their characteristics? It is the detailed elaboration of query intents that will allow you to implement higher quality data generation patterns.

Setting up routing (routing)

Immediately after the development of the structure, analysis of semantics, grouping queries by intents - it is necessary to implement the technical part of the generation:

• Setting up correct address processing according to the structure discussed above (CNC),
• Configure 404 page handling when requesting non-existent cities and countries (so that no useless pages are created),
• Implementation of loading different templates depending on the requested URL.

At this stage, you need to make a clear and transparent TOR for the programmer, which should take into account:

• Each generated page - must have its own content template;
• Depending on the address requested by the user, the generation involves either data from the "Countries" table or data from the "Cities" table;
• Each generated page - must have its own unique set of meta-data: title, description.
• Each end page (e.g., the "Flights from Detroit to Berlin" page) should be accessible by a unique address (there should be no duplicate content at different addresses).

Upon completion of the technical part - the section should be a kind of "blank", for the operation of which - you need to work out the patterns for each type of landing pages.

Generating templates

We will not describe in detail that each landing page should have an "associated" functionality (have a form of order or search for airline tickets for the viewed destination), have a quality design and convenient functionality - because it's obvious.

Emphasize the fact that each template should be sengenyvat text blocks to increase the degree of uniqueness of pages and increase relevance to relevant queries.

An example of a fictional text:

On SITE.ru you can buy cheap airline tickets from Detroit (USA) to Berlin (Germany) at a cheap price. We compare airfare prices among 700+ airlines flying between Detroit and Berlin and let you choose the most affordable flight. We sell airline tickets to Berlin without commissions, including fuel and airport taxes...

Substitute variables in this text:

On SITE.ru you can buy cheap airline tickets from %FROM_CITY%(%FROM_COUNTRY%) to %TO_CITY%(%TO_COUNTRY%) at a good price. We compare airfare prices among 700+ airlines flying between %FROM_CITY% and %TO_CITY% and let you choose the best priced flight. We sell airline tickets to %TO_CITY% without commission, including fuel and airport taxes...

To improve the quality of your texts - we recommend using not only the original country names, but also their declension in the generation.

Depending on the functionality of your site - accompany all the elements (price calendars, search forms, lists of airlines that operate flights on the route, etc.) with independent text blocks using the names of the cities of departure and destination.

Additional data sources

Always be on the lookout for additional data sources with which you can expand the number of your landing pages, the number of key HF phrases accounted for, and improve the quality of your landing pages.

Additional data for the subject of airline tickets:

• Flight schedules,
• Weather,
• promotions and special offers,
• related topics (hotels, tours, insurance, etc.).

Topics "Tourism": sale of tours

In this subject, as in the subject of airline tickets, you must first clearly articulate the description and characteristics of the product, which is the main product. The final product in the theme of "Tours" is a "travel package". "Package" is a set of services combined into a single product offering, formed by a tour operator, distributed through a network of its own offices and through travel agencies.

Once we have clearly described - what exactly is our main offer, it is necessary to collect the characteristics of this product:

Tour properties

• tour operator
• city of departure
• destination country
• resort
• departure date
• vacation month
• holiday season
• number of days off
• number of adults
• number of infants
• charter flight or not
• terms of insurance
• package price
• price per person
• type of holiday (beach, skiing, etc.)
• type of accommodation (hotel, apartment)
• type of accommodation (building, bungalow, etc.)
• hotel star
• Food type (RO, BB, HB, FB, AI)
• hotel features
  
  • line by the sea (if it is a beach holiday)
  • area
  • attractions (entertainment in the area)

Selected properties are structure-forming, and it is on these properties will build the future structure of the site / section.

Initial elaboration of the structure

We assign a name (variable) to each structure-forming property:

• departure city = %FROM%
• destination country = %TO_COUNTRY%
• resort = %TO_CITY%
• month of rest = %MONTH%
• holiday season = %SEASON%
• hotel stardom = %STARS%
• Food type (RO, BB, HB, FB, AI) = %BOARD%

Since any travel agency works through the booking gateway - let's assume that all the tours on all of the above characteristics we will be able to show the user on the landing pages.

Sufficient structure

As in the previous example - it is necessary to study the main semantic groups and on the basis of the analysis design the future structure of the section. A sufficient structure for the topic "Tours" may look like this:

• *.site.ru - city of departure (if there are many cities of departure - we recommend using subdomains)
  
  • /*TO_COUNTRY*/ - destination country
    
    • /*STARS*/ - hotel category
      
      • /*BOARD*/ - type of food
      • /*SEASON*/ - holiday season
        
    • /*TO_CITY*/ - resort
      
      • /*STARS*/ - hotel category
      • /*BOARD*/ - type of food
      • /*SEASON*/ - holiday season
        

Keep in mind that each node of the structure - must have its own unique template for generating content, texts and meta-data.

Data sources

Most travel agencies that sell their services on the Internet use a gateway (or intermediary) that offers certain sets of widgets for the site.

Depending on the widget provider for the site - the set of data available for use in content generation will vary. Use the data provided in your partner's tools to generate landing pages.

Integration with widgets

As a rule, most "widget" vendors have taken care of the flexibility of their tools, and allow you to customize filters in search forms depending on the parameters passed. This allows you to display on the landing page "Tours in Greece" - filled widget with the country, list of resorts in Greece and available hotels. If the widget is connected via IFRAME - this is usually done via a set of GET parameters in the iframe link.

In generating the structure - we start from the filtering capabilities.

1. We take all countries, resorts and other parameters from the widget form.
2. Save the necessary values to the database:
   • country name (with cases);
   • country id in the booking system;
   • city name (with cases);
   • city id in the reservation system;
   • other parameters involved in the generation of the structure (with cases).
3. Set up routing based on the collected parameters.

Further actions are identical to the above example:

1. We implement the technical part of the generation: processing the URL, substitution of appropriate templates depending on the requested address;
2. Set up 404 error handling when requesting non-existent parameters.
3. Set up the generation of meta-data: title, description.
4. We work out a template landing pages: search tools, the layout of widgets.
5. We work through text generation templates using variables from the database.

Generation of landing pages in the theme of "Tours" allows you to implement the structure on all countries, destinations and resorts that you offer your customers in a short time. If you do not work with one destination - then manual elaboration of such structure on the basis of one of the CMS - will take a lot of time, and as a consequence - will lead to excessive costs of development and copywriting. When generating - you get thousands of pages, while spending significantly less time.

The benefits of content generation

Content generation saves not only your time but also your money. This method of filling the site is great for sites of many topics, not only those described in the examples above. After implementing the basic structure and working out the templates - regularly analyze the number of pages in the index of search engines, traffic to the pages of the generated section, and if you find flaws - make adjustments to the generation templates. Please note that in order to make changes on 10,000 pages (e.g., resort pages) - you will need to make changes only in 1 generation template, which significantly saves your time.

The main problems you may encounter are

• Poor development of templates and texts.
  Can lead to pages dropping out of the index for "Not good enough" reasons.
• Excessive generation depth.
  Long reindexing of changes with a large number of unnecessary generated pages.
• Technical errors. Poor quality or incomplete TOR for functional development (technical part) can lead to a huge number of errors: duplicate pages, wrong server responses, etc.

Tips and Tricks

If you've decided to implement a generation section on your site, here are a few recommendations:

1. Do not outsource the development of the content generation template to third-party copywriters. If you decide to use the services of a copywriter - make sure that the copywriter really knows your topic, otherwise you risk getting templates with low-quality texts, which will significantly reduce the quality of landing pages and have a negative impact on their ranking in search engines.
2. Carefully study semantics and demand, gather all possible characteristics of your product or service, and estimate the generation potential.
3. Do not generate separate pages for low-frequency queries. Most of the HF semantics can be allocated to intermediate generation nodes.
4. Do not start the work on the generation of content with the technical part. Primarily - analysis of semantics, demand, highlighting structure-forming properties, and only after that - technical implementation of the generation (routing and elaboration of text generation templates). If you change the order - you may find unaccounted semantics groups later, which will lead to a change in the logic of page generation and URL processing, and to unnecessary financial and time costs.
5. Constant analysis and timely adjustments are the key to success.

Have you tried content generation yet?

More Posts