Time to change your approach to keywords

In this post, we’ll touch on a topic that isn’t talked about much in Polish SEO yet, and it’s one of the most important elements in SEO today.

The change in approach to keywords was influenced by two updates: Hummingbird and Panda 4.0. Both have changed how we should view keywords today and how we should use them.

Hummingbird update

.

GRAPHIC

.
In 2013, Google introduced an update called Hummingbird. It didn’t cause much of a stir in Poland (it didn’t have a big impact on results), but it really changed a lot.

What exactly?

Google, until this change, analyzed content by the keywords in it. With this update, the way content is analyzed has changed. Currently, each text is assigned a topic. This update is an introduction to semantics, with the help of it Google tries to recognize the intent of the user who enters a search query.

Therefore, should we stop using keywords in the article and focus on the topic of the article?

Not really. The algorithms are not yet perfect, so it is worth combining modern and old approaches to achieve optimal results.

Panda 4.0 – Topical authority update

.
In May 2014, there was another update that affects the perception of keywords. As you know, Panda is an update that affects content: all sites that have low-quality content are exposed to the negative effects of this algorithm. Update 4.0 was one of the most important in the whole cycle of these updates. It determines how much a site covers a given topic – the so-called topical authority. Let me explain it using the diabetes topic as an example.

A site that has only described type 1 diabetes in its entire structure is less likely to rank high than a site that has thoroughly described every topic within diabetes. This doesn’t just apply to this one article about type 1 diabetes, but to all content about diabetes. The more we cover the topic of diabetes with our content, the higher rankings we have for all diabetes queries. However, it is important to know that a good practice in this case will not be to describe all issues about diabetes in one very long article – Google does not like that. Articles that describe one fairly narrow topic in depth are better treated. When creating content about diabetes on the site, you should break this topic into smaller topics, as in the image shown below.

GRAPHIC

Next, each topic should be reflected in the site content. If each topic is described in the site, the site receives high topic authority from Google – the site content largely covers the topic. It is therefore important to identify all side topics within the main topic (such as diabetes), but also how the content was created (more on this later in the article).

The structure of the site is also an important element from a topical authority perspective. Content within a topic should be related to each other in an appropriate way. The Architecture of information should be built in such a way that content from a given topic intermingles in the structure of the site. If a given topic has been covered with many articles, but they are scattered on the site in an unstructured way then the SEO effect will be much worse.

Assumption

.
According to this theory, thorough coverage of a given topic should make the site rank high for the topic (not just for single words) in a search engine. Thus, when describing the topic of “diabetes,” one would expect an increase in position proportional to the increase in topical authority, or topic coverage. Such an assumption is partly confirmed by a recent study conducted by baclinko on a sample of 10,000 URLs. Content covering a topic thoroughly tends to rank higher in search results than content covering topics more poorly. This is illustrated by the chart below.

GRAPHIC

Examples

.
We have made an assumption. It’s time to see if it is reflected in real search results. For testing, we used the “diabetes” theme already mentioned above, and the largest medical websites in Poland were used for analysis.

• medonet.pl – 2nd place for the phrase “diabetes”

.

• poradnikzdrowie.pl – 5th place for the phrase “diabetes”

.

• abczdrowie.pl – 6th place for the phrase “diabetes”

.

• zdrowie.wp.pl – 17th place for the phrase “diabetes”

.

These are 4 medical sites that, in theory, should have high coverage of the “diabetes” topic. Let’s check how big this coverage is and how it affects the position of the sites within the analyzed topic.

Medonet.pl

.
For the word “diabetes” medonet.pl has 3,090 indexed results in Google.

GRAPHICS

This is quite a sizable number considering that we are analyzing only one topic . Currently, the site is visible on 670 keywords related to diabetes (such words that contain the word “diabetes”).

GRAPHICS

The average position of the site within these 670 keywords is 18.52. The 670 keywords are spread across 103 different URLs so one URL is seen on average for 6.50 keywords. The total potential of words within “diabetes” on which the site is visible is 86,110 searches.

Poradnikzdrowie.pl

.
For the word “diabetes” poradnikzdrowie.pl has 72,300 results indexed in Google.

GRAPHICS

This service is visible on 1204 keywords related to diabetes.

GRAPHICS

The average position of the site within these 1204 keywords is 8.12. The 1204 keywords are spread over 298 different URLs so one URL is seen on average for 4.04 keywords. The total potential of words within “diabetes” on which the site is visible is 106,580 searches.

Abczdrowie.pl

.
For the word “diabetes” abczdrowie.pl has 36,700 results indexed on Google.

GRAPHICS

This website is visible for 1267 keywords related to diabetes.

GRAPHICS

The average position of the site within these 1267 keywords is 8.44. The 1267 keywords are spread across 312 different URLs so one URL is seen on average for 4.06 keywords. The total potential of words under “diabetes” for which the site is visible is 110,460 searches.

Health.wp.pl

.
For the word “diabetes,” zdrowie.wp.pl has 6,910 indexed results in Google.

GRAPHICS

GRAPHICS

Summary

.
The following table presents the metrics described above. The summary of this data is intended to test the relationship between the coverage of the topic “diabetes” and the visibility of the analyzed sites on this topic.

Conclusions

.

• A high correlation can be observed between the number of URLs in the topic “diabetes” and the average position of the site

.

• Services with a high number of address URLs in a given topic are visible for a higher number of keywords

.

• The sites with the best information architecture (links in a given topic), i.e. abczdrowie.pl and poradnikzdrowie.pl, are also the ones with the highest average position and highest visibility in the diabetes topic.

This analysis seems to confirm the assumption we made. The conclusions seem to be clear. Of course, visibility or average position is also affected by external links and the authority of a given site, but for the analysis we chose sites with similar parameters in this regard.

Content creation

.
With the Hummingbird update, the way in which content should be created has also changed. We have described below two approaches to content creation. The old one and the new one, which differentiates between how to look at the use of keywords in content.

The old approach

.
The old approach paid homage to the principle: the more closely aligned landing pages the better. Typically, the text covered one very narrow topic, e.g. “Gestational diabetes – diet” and variations of the word were placed in the content, as shown in the graphic below.

GRAPHIC

This resulted in short text forms covering a very narrow topic, but not exhausting it. This way of creating content made it:

• There was a lot of content similar to each other thematically

.

• These contents were generally short, and as is known, longer content affects positions better

.

Of course, using this approach is nothing wrong. Google will not punish us for it, but the effects of such action will not be as effective as they once were.

New approach

.
With the hummingbird update, Google no longer focuses only on keywords and their number of occurrences when analyzing content, but tries to attribute a topic to the content. The content should therefore describe the topic in depth. However, you should not focus on a very narrow topic (like diabetes in pregnancy vs. diet), but focus on a broader topic (not too broad) and describe it in depth which will allow the search engine to more likely determine its topic, but also assign it a higher topic authority. The method of such division is shown in the image below. The topic of the content in this case is “Diet in Diabetes” and not a very narrow topic like “Diabetes in Pregnancy – diet” – this topic is only a side topic.

GRAPHIC

Content created in this way:.

• Describe a fairly narrow topic like “Diabetes Diet,” but do so in depth which is beneficial from a topical authority standpoint

.

• They tend to be much longer (Google likes long content)

.

• Totally, they take less time (there is less content to write)

.

As I wrote above – the algorithms are not yet perfect. Just because Google assigns a topic to a given content does not mean that we should completely cut out keywords. Just describing a given topic will not ensure high search engine rankings. Combining this technique with properly selected keywords will give the best results. This issue is well explained by Rand Fishkin in one of the episodes of whiteboard friday (video below).

.

So when creating a text about “Diabetes Diet” the content structure should look like this.

GRAPHIC

In this particular example (diabetes) it could look like this.

Topic: Diabetes – Diet.

Subtopic 1: Gestational diabetes – diet

• Pregnancy diabetes – diet – recipes

.

• Pregnancy diabetes – diet – menus

.

• Pregnancy diabetes – diet – fruits

.

Subtopic 2: Type 2 diabetes – diet

• Type 2 diabetes – diet – recipes

.

• Type 2 diabetes – diet – what to eat

.

• Type 2 diabetes – diet – Mediterranean diet

.

There is a main topic (Diabetes – Diet), side topics, but they still have keywords in them, which should be used in the content using the standard content creation model. This combination of two techniques will provide optimal results.

Conclusion

.

• Standard keyword saturation is something that is slowly ceasing to work.

.

• The assumption that higher topical authority has a positive impact on positions for the entire theme was confirmed not only by the baclinks test, but also by the cases we analyzed

.

• Semantics is a topic that is still in its infancy. Contrary to many opinions, keywords should not be abandoned, but they should be used in a different way.