How to Optimize the User Experience to improve SEO
For Google , user experience (UX) is a positioning factor and increasingly important. Due to this, it is necessary to know how to analyze the behavior that users have on our page, which areas of our website have better results and which areas we should improve.
Aabhas Vijay is a Digital Marketing consultant and runs his own blog about SMTP with more than 5 years of experience helping large companies to position themselves at the top of Google. In this article he talks about the relationship between user experience and Google is and how it affects the positioning of a website.
SEO is not optimizing sites for search engines, but for people who use search engines.
In order to better understand how the user experience influences SEO , In this article we will cover the following topics
1. Understand the goals of users and Google
Many people and companies think that in positioning from the following perspective:
A user does a Google search with an intention> my landing responds to this intention> I give the answer to the user> finally, I direct the user towards conversion.
This is a half truth; which is the same as saying that it is a lie.
► Why is it a lie?
The only one who knows the intentionality that users have when performing a search is Google. Many users may have different intentions and do multiple searches, but end up on the same page.
A search is not the same as an intentionality.
Also, a landing page does not respond to a single search, but can respond to many different searches. Therefore, by saying that the intent of the users is multiple .
► How do we respond to multiple intentions?
The solution is clear: with a cluster of landings that respond not only to that search, but also to future searches.
► With this we are going to take the user to the conversion / purchase?
Possibly not. It depends on the customer journey . If a user takes two weeks to buy a product, It is assured that with an informational search it is very difficult to get that user to convert.
What is the most relevant result for Google?
The most relevant result for Google is the one that responds to the current search and future searches . To show this training, he used the evolution of Google results as an example.
► Google results in 2016
In 2016 when you were doing the search “how long is the Matrix” the following information appeared:
In that year, Google released what is known as Google’s responses, where it directly gives you the solution to the search you have performed. In this case, I was telling you that the Matrix movie is 136 minutes long .
In addition, on the right side it already offered the Knowledge Graph with information about the film and actors who participated in it.
Finally, at the bottom you could see SEO results .
► Google results in 2018
The first thing is that Google does not say 136 minutes, but results in 2 hours and 16 minutes . This result is more humane, since this is how we measure time. In addition, it not only gives the time in the movie you are looking for, but also other movies in the saga.
On the right side, the Knowledge Graph remains.
The other great variation is the disappearance of the SEO results for other questions that users ask in addition to the one you are asking.
Conclusion : Google is trying to answer your question and future questions with a landing that groups content from different pages. In other words, Google is responding to you with a cluster of landings and content.
2. How does the user experience affect SEO?
One of the metrics that Google uses to see the quality of a page, at the level of user behavior, is Pogo Sticking .
Pogo Sticking : a key metric in SEO that indicates the percentage in which users visit a page through Google results and after spending x time leaves it by hitting the back button of the browser.
We can measure Pogo Sticking with Google Tag Manager . We can create an HTML5 event called Story.pushState () that allows us to put our URL in the user history with the variables that we want.
In this way, when the user clicks on the back button, they undergo a redirection of milliseconds to a page on our site that allows us to know if the user clicks on the back button of the browser and thus collect relevant information to be able to analyze it later.
Using this strategy does not affect the user, since it is imperceptible to him.
Google Analytics is a very useful tool as a database, but as the only analysis tool it does not work. This is because it only offers 17% of the total data and must be complemented with other tools.
Other metrics to complement Analytics data, in addition to Search Console, are:
No. of Queries vs Pogo Sticking
- X axis : middle position
- Y axis : Percentage of pogo sticking
- Circle color : Cluster to which it belongs
- Circle size : No. of queries / searches
In this graph there is a pattern: when pogo sticking is greater than 20%, the number of circles is smaller and smaller, that is, that Google thinks that this user behavior is bad and reduces the number of keywords for which it positions that content.
However, below 20% Pogo Sticking, there are a lot of circles and the size is larger. This means that for Google there is a positive behavior , that is, the user finds what he was looking for and, in addition, future doubts.
Impressions vs Pogo Sticking
- X axis : middle position
- Y axis : Percentage of pogo sticking
- Circle color : Cluster to which it belongs
- Circle size : No. of prints
The effect is the same as in the previous case, except for the behavior of a cluster. This means that there is a cluster where user behavior is different from that of the rest of the web.
Does this mean it is wrong? No. User behavior can be different for each website, and even different for each cluster on the same website.
Each section of a website can have different behavior thresholds
Whiskers Box Diagram
To know the point / threshold from which the results are good or bad, we need to make a “Whisker Box” diagram.
We are not only interested in optimizing pages, we are interested in having a KPI that tells us from what result something is right or wrong.
The Whiskers Box diagram is used to visualize data with asymmetric values , since, in the end, the distribution data of the pogo sticking are asymmetric.
Aabhas shows in the example, in this case the median (eye, not the mean) of the behavior indicates that above 12% of pogo sticking the results would not be good.
Each cluster on our page works differently, therefore, and as in the previous examples, there are two clusters above the median whose behavior is different.
We cannot apply a user behavior rule to our entire website.
In addition, we also see that outliers appear (those that leave the graph) that are the first to attack. To do this, we must combine these results with the number of impressions that these data have and look for patterns to improve.
3. How can we optimize our clusters?
Web Architecture
First it is interesting to understand how the architecture of the web in question works. Generally, the architecture is as follows:
This architecture is fine, but is it real? Does Google send your users to the home page or internal pages? How do users behave? This architecture does not reflect the true behavior of users on a website.
For Aabhas, the relationship between clusters is more important than the architecture of the web, since clusters serve Google to measure authority.
Links outside the cluster harm us.
If on a website, we link everything to everything is a problem. Everything inside the same cluster is well linked, but the rest of the links are not positive.
EXAMPLE:
The journalists of readings, linked to each famous person that appeared in the news. Every time a famous person was named, he was linked to even if it had no relevance in the news.
► What was going on? That Google understands each celebrity as a cluster.
► What did they do? They relabelled each famous person with those news that were really related to him and thus not mention more famous people who could appear indirectly.
In this way, they went from having a visibility index of 0.5 to 3 on the celebrity tabs.
Page Rank
The Page Rank is the probability that a user clicks on a link.
The Page Rank is not distributed equally in all your links on your website.
Clicking on a link has a different Page Rank weight depending on:
- Size, color and shape.
- Position within the document
- Number of words.
- Anchor text
- Cluster relationship
- User behavior
Also, depending on the position of the link on the website, it has a different value and this is not measured by SEO tools .
- Architecture = Header
- Contextual = Content
- Secondary = Sidebar
- Secondary architecture = Footer
Conclusion
► What interests us?
Control, measure and improve the clusters of those projects in which we work. The problem is that SEO tools do not give the necessary information, since they cannot measure a different page architecture for each project.
► What do we have to have?
- Relationship between URLs and number of cluster
- Link position in HTML
- Anchor text
- Import all the data to generate a data visualization.
** TRICK: Google Search console only gives 1,000 results, but if you link it with Data Studio you can export all the data in an excel.
You generate a database with the previous data, and data such as traffic source, sessions … etc. that you want to add to your analysis.
Finally, with all this data you can create a directed acyclic graph with which to see the true behavior of users on your website.