What does Markov chains mean in SEO?

In SEO Markov chains can be used to see the way users move through your website and then take actions to improve their experience so more visitors achieve your goals (buy a product, sign up to a newsletter, etc.).

I'll show you Markov chains in action on eCommerce websites, then on blog or article sites and then how you can see this for your own site as well as how to actually leverage them to your advantage.

A Markov chain example for an eCommerce website

Here are three statements about the website.

Statement 1: 30% of the visitors who arrive at the site after searching 'Product A' in a search engine come to the site via the page product-a.html. All these visitors then click on a link advertising the premium version ('Product A Plus') which takes them to the page product-a-plus.html. They place this in their cart and go to checkout (checkout.html).

Statement 2: 30% of the visitors who arrive at the site after searching 'Product B' in a search engine come to the site via the page product-b.html. All these visitors then click on a link advertising the premium version ('Product B Plus') which takes them to the page product-b-plus.html. They place this in their cart and go to checkout (checkout.html).

Statement 3: No one who arrives at the site via the page product-a.html ever buys Product B Plus. No one who ever arrives at the site via the page product-b.html ever buys Product A Plus

This website has two Markov chains:

• product-a.html -> product-a-plus.html -> checkout.html
• product-b.html -> product-b-plus.html -> checkout.html

Why is a Markov chain important to an eCommerce site?

Many eCommerce webmasters make big mistakes with devastating consequences. They look at their sales and think "I never sell 'Product A' or 'Product B', they're just taking up space in the warehouse and costing me money when I advertise them".

They drop the products from their websites and their sales of 'Product A Plus' and 'Product B Plus' collapse because they hadn't seen or understood the Markov chain.

A Markov chain example for a blog or article website

Same approach, here are three statements about an imaginary website.

Statement 1: 30% of the visitors who arrive at the site after searching 'Keyword Phrase A' in a search engine come to the site via the page article-a.html. All these visitors then click on a link in the article which takes them to article-b.html. After reading this they sign up to the newsletter (newsletter.html).

Statement 2: 30% of the visitors who arrive at the site after searching 'Keyword Phrase X' in a search engine come to the site via the page article-x.html. All these visitors then click on a link in the article which takes them to article-y-plus.html. After reading this they sign up to the newsletter (newsletter.html).

Statement 3: No one who arrives at the site via the page article-a.html ever reads article-y.html before signing up. No one who arrives at the site via article-x.html ever reads article-b.html before signing up.

This website has two Markov chains:

• article-a.html -> article-b.html -> newsletter.html
• article-x.html -> article-y.html -> newsletter.html

Why is a Markov chain important to a blog or article site?

Many webmasters make big mistakes with devastating consequences. They look at their analytics and think "No one who reads Article A or Article X ever sign up to my newsletter".

They drop the articles from their website and newsletter registrations collapse because they hadn't seen or understood the Markov chain.

How to see if a page is in a Markov chain?

Google Analytics will tell you this for every page on your website so long as you have set up goals correctly. With eCommerce it is well worth setting up your analytics so it knows the value of each transaction. For content driven sites give your goals a nominal value (e.g. $1 for every sign up).

After you have done this and left Analytics a few days or weeks to gather data you will see, for every page on your website, a 'Page Value'. This says that if you can get a visitor to this page you will earn $x. Its a really simple way to make sure that you don't rock a Markov chain you may not be aware of. If Analytics says a page on your site has a value its because it plays a role in achieving your goals.

That doesn't mean you should never touch it again, just don't do anything dramatic like deleting it.

Leveraging Markov chains

You can also use Markov chains to optimize your user experience. Let's say, for example, you sell bed linen and notice, by looking at your sales, that 10% of people who buy red duvet covers also buy red pillow cases and 10% of people who purchase green duvet covers also purchase green pillow covers. But red duvet buyers never purchase green pillow covers and vice versa.

There is a chain here that could possibly be strengthened. Try adding a focused suggestion to the site. When people put a red duvet cover in their baskets specifically ask them if they want to purchase red pillow covers at the same time.

This really helps people like me who forget the obvious when it comes to purchases for the home and end up doing a face palm later. You might then find 60% of your duvet buyers purchase pillow covers dramatically increasing your sales.

Here you have not just been an observer of a Markov chain but actually taken action to change the dynamics of that chain.