Cut your data analysis time in half with correct URLs architecture

September 4, 2020

Magda Founder & Head of Strategy
Read more Read more Magda started marketing for friends by stealing candies from her mom’s shop to distribute. She now does the grown-up thing and studies data science at Oxford University. She’s quite overfocused on marketing, statistics, and data. Between that and her hobbies in learning genetics, understanding psychology, doing exercise, and watching pandas, she doesn’t have much time left for the latest movies. Throw a Harry Potter her way – and super-educated Magda won’t get it at all.
Alex Head of Data Engineering
Read more Read more After working in web development and analytics for 5 years, there is NO WAY Alex will let you do anything before he’s got a whole data architecture engineered into your website. His motto: “In God we trust. All others must bring data.” Don’t be intimidated, though. He’s super patient and human about it
  • Introduction
  • Faster segmentation
    Including 2 scenarios
  • Better Website Tracking
    How to build an effective URL architecture
  • Conclusion

Introduction

The bedrock for conversion optimization is data.

However, data takes time! Extracting it, double-checking it, understanding it and most of all, organizing and segmenting it so that it tells you the story behind the user. Too often, we found ourselves wasting precious time manually organizing the data instead of investing that time in thorough analysis and valuable insights. Why?

Mainly, because of a website’s messy URLs Architecture or incorrect analytics implementations.

Today we’ll focus on why a proper URLs Architecture saves online businesses a lot of the most precious resources known to man – time and money.

Reading this article will be well-invested time if you are:

An entrepreneur who wants to increase the revenue your online business brings in [WITHOUT struggling and investing weeks and weeks in order to understand what’s happening on your site so you can make informed decisions]

A web developer who often builds sitemaps and you don’t want frustrated clients when later on they’ll want to use data analysis to optimise their site.

A data savvy marketer and you want to educate your customers by sharing with them a better way of designing the URL Architecture of the websites they build.

LET’S UNCOVER THE BENEFITS OF A PROPER URLs ARCHITECTURE

Simply put, these would be:

1. Faster Segmentation

2. Better website tracking

Ok, that’s vague. No worries we’ll get to examples and real scenarios.

Faster Segmentation

A good URL architecture can be the difference between 2 minutes to build a segment and 20 minutes to build a segment.

Generally, for large e-commerce websites, we create up to 250 custom segments, many of them containing URLs.

60 extra hours of work.

I’ll say it again: 60 hours of your life or $10k out of your pocket if you are working with top consultants or web analytics agencies charging around $170/hour.

Ok, but is segmentation that important for insights extraction and optimizing conversion rate?

YES.

If you are not doing segmentation you are not doing analytics. “Segmentation: Do Or Die, There Is No Try!” as Avinash Kaushik

Understanding how segmentation can benefit a business and help increase revenue

Segmentation is such a vast and essential subject, that’s why we’re working on a comprehensive guide on segmentation, which we will publish soon.

For now, I want to emphasize the fact that:

  • Segmentation is an essential part of any conversion rate optimization process. Segmentation facilitates the understanding of various characteristics of subsets of your audience and it helps you determine levels of engagement, reveal funnel leaks and identify optimization opportunities.

The story goes like this:

• Moreover, by segmenting your traffic based on which sections of your website they interacted with, you’ll be able to create well defined remarketing audiences and customize ads based on their preferences and previous website interactions.

Therefore, by making segmentation easier you are optimizing the optimization process.

Understanding the difference between a good and a bad URL Architecture with examples

A good URL architecture should make it easy to answer questions such as:

• Do female visitors who read your blogs articles with tips and trends from top designers convert better than those who read no blog articles at all? IF yes, do they also have a higher average order value?

• What are the top highest converting filters for dresses category pages/t-shirt category pages?

• What are the specific characteristics of the audience browsing through sportswear pages?

Let’s say that you want to answer the last question but your website URL structure doesn’t include categories or subcategories, it just looks like this www.site.com/p/blue-lng-joggning-pants.

Aannd, ladies and gentlemen, we’ve arrived at our destination ….the moment of truth.

How do you answer the questions above if your URL Architecture is, to use the professional term, messy?

Will you be able to divide the sportswear from other clothing articles? How about differentiating between sports pants and other sports articles?

Well, depending on how large the website is, there are 2 options available but none of them will make you happy:

A. For a large website, it’s impossible to answer

B. For a small website, it’s possible but it will take so much more resources than it should (time, money, frustration)

But what if your URL would look like this:

www.site.com/p/sportswear/pants/running/blue-long-jogging-pants

Answering the question would take you no more than a few minutes.

Now, the examples I promised:

Let’s assume that you are in charge of an online e-commerce business which sells dresses, shoes and t-shirts and you want to answer the following question:

“What % of site visitors reach summer dresses product pages and what’s their behavior, based on the channels they are coming from?”

A. SCENARIO 1 [Bad URL Architecture on a large website with thousands of product pages]

Suppose that the URLs of dresses product pages from your site look like the ones below and you are using Google Analytics to track the website traffic. Let’s say you have over 1000 dresses on your site, meaning 1000 unique URLs with nothing in common within their structure.

So here’s the challenge:

How do you bring all the site visitors who reached one or more dresses product pages under one Analytics Segment if there is No Common URL section?

Well, you can’t because:

• A Google Analytics segment can’t fit in so many different URLs.

As you can see, the OR option is no longer available after you introduce 20 URL fragments. And even if it would be possible, how much time would that take you to do that?  [even by using robotic process automation tools, that would still take you additional resources]

• There’s also the possibility of using regex if they at least have the category in common, however, that is also limited to a number of characters.

Now how do we tackle this challenge?

Clearly, you’ll need

Or, you know, one common URL fragment to rule them all.

Would the second version of URLs infrastructure help you find the answer to your question?

Yes. And fast. All these URLs corresponding to summer dresses products have something in common /p/dresses/summer/, so by setting a segment like the one below, you group all the sessions /users that are interested in summer dresses.

And there you have it, in less than 30 sec we found out that 18.150 users (or 23.016 sessions) have reached a summer dresses page within the selected timeframe. Now you can save this segment and apply it to various reports.

B. SCENARIO 2 [Bad URL Architecture but a smaller website with tens/hundreds of product pages]

If you are only selling 100 different types of dresses, with 100 different URLs.

Now that’s…. well…. doable, you could manually introduce each URL in segments and maybe even make use of regex. Very time consuming compared to the ideal situation, but at least you get the answer you are looking for.

[DISCLAIMER] – CONTENT GROUPINGS

You might know that content grouping is another Google Analytics feature to segment/group the traffic of a website so you easier understand what’s the visitors’ behavior without building a ton of segments.

The good news is that it can save the day with content groupings if you Don’t Have a logical URL structure of your website. But….. hold your horses, cause that’s not going to be easy peasy. [that’s how Google punishes you for not having a good URL structure ]

Not only you’ll have to arm yourself with technical knowledge [you’ll need data layer and GTM] but comparing to segments, content groupings have some weaknesses.

• You’ll get data for those groupings starting the moment when you created them [alike segments where you can get as much historical data as needed]

• It takes more steps to filter data on the view level by Content Grouping

• There is a limit of 5 content groupings

Don’t get me wrong, I’m a big supporter of Content Groupings as they save a lot of time IF they are implemented from the very beginning and complemented by the magic of segments.

Only that depends on the content.

If you want to optimize your revenue next week, but you have to wait for 3,6 months for your content groupings to register a reliable amount of data, how would that feel?

My suggestion is to build a proper URL structure from scratch [or correct the one you have now], group your main site sections through content groupings and complement with additional segments when you’ll start digging deeper in your data.

Better website tracking

URL structure can also become an issue when you are implementing tracking on your website.

Let’s assume you have a blog and the bounce rate is quite high because the users read your article but exit the pages without any other interactions. However, you don’t want Analytics to consider as bounces those users who read blog articles up until the end because they were clearly engaged.

Therefore, you decide to create an event in GTM, which should trigger when users scroll up to the end of an article. Then, you tell analytics to consider this event as a website interaction, meaning users who trigger this event will no longer be considered bounces.

Now the simplest way to do this would be by using the URL fragment which is common to all blog articles:

Normally, this would be a one minute task, but not if your URLs are like this:

www.domain.com/name-of-the-article, instead of:

www.domain.com/blog/name-of-the-article

Sure, when the URL structure is bad and blog articles have no common fragments, it can still be done with various workarounds, but the time investment increases significantly.

How to build an effective URL architecture

Don’t do it halfway.

Sure it’s a great first step to include category names in URLs, however, if you really want to understand your customers and get actionable insights, it’s not enough. Not by far!

Most websites tend to track things only halfway, building their URLs to include only the product categories, ignoring all subcategories.

An ideal URL architecture would include everything:

www.ecommercesite.com/large-home-appliances/embedded/plates/name-of-plate

www.ecommercesite.com/large-home-appliances/embedded/furnaces/name-of-furnace

www.ecommercesite.com/large-home-appliances/embedded/hoods/name-of-hood

www.ecommercesite.com/large-home-appliances/embedded/dishwashers/name-of-dishwasher

If you are just starting out to build your website, take advantage of sitemap building tools, such as Slickplan. Besides the fact that they give you an overview of your website and help you make the website structure as intuitive as possible for the user, they also provide you with URL structure, including all categories and subcategories.

Conclusion

Me and my team have gone through some very frustrating moments when we brainstormed all the things we could find out and draw insights from, only to realize that we are not able to properly look at data because there was no way to go around the poor URL infrastructure and build the segments we wanted to.

The struggle is real.

To each and every business out there, currently building their own website, you should be invested enough to have the foresight to do get it right from the very start! And trying to optimize your website later might be very difficult!

Author

Magda Founder & Head of Strategy
Read more Read more Magda started marketing for friends by stealing candies from her mom’s shop to distribute. She now does the grown-up thing and studies data science at Oxford University. She’s quite overfocused on marketing, statistics, and data. Between that and her hobbies in learning genetics, understanding psychology, doing exercise, and watching pandas, she doesn’t have much time left for the latest movies. Throw a Harry Potter her way – and super-educated Magda won’t get it at all.
Alex Head of Data Engineering
Read more Read more After working in web development and analytics for 5 years, there is NO WAY Alex will let you do anything before he’s got a whole data architecture engineered into your website. His motto: “In God we trust. All others must bring data.” Don’t be intimidated, though. He’s super patient and human about it
Download our FREE URL Architecture Template
Do you have a question related to
Cut your data analysis time in half with correct URLs architecture
We'd love to hear it: