A Bullseye View of Web Analytics

Target is an international discount department store retailer that now has 1,797 brick-and-mortar stores spread across the United States, an additional 124 stores in Canada, as well as locations in the growing market of India (Corporate Fact Sheet, 2013).  Additionally, the company has thirty-seven distribution centers, 361,000 team members around the world, and of course its online ecommerce site, target.com (Corporate Fact Sheet, 2013).

Since the first store opened in Rossville, Minnesota in 1962, Target has grown to become one of the top retailers in the United States.  The brand has made a pretty good name for itself over the years by providing high quality merchandise at discounted prices, which has attracted a customer base consisting mostly of middle-class families with an annual household income of roughly $64k (Corporate Fact Sheet, 2013). 

With technology continuing to develop at a rapid rate in the U.S., shopping online has become less frightening than when it was first introduced to the consumer market.  The popularity of online shopping has given rise to numerous online-only retailers such as Amazon.com and has also led to the vast majority of brick-and-mortar retailers expanding their brands via an ecommerce site.  But even with all of the competition, especially is the department store industry, Target has managed to become the second largest general merchandise retailer in the United States, with its ecommerce site, target.com, being ranked as one of the most visited websites (Corporate Fact Sheet, 2013). 

Target launched its first online website in 1999 and partnered with Amazon to help perfect the customer shopping experience online.  But, in 2009, Target decided to end its decade long relationship with Amazon, despite the fact that Amazon had spent years developing technology to help perfect the online customer experience (Zmuba & Patel, 2011).  Because the company was growing at such a substantial rate, Target decided it would be best for it to develop its own ecommerce technology and create and manage its own website (Zmuba & Patel, 2011). 

First Target Website

So, instead of using Google Analytics or other Analytics tools as a primary way to track its web analytics, Target decided to do things a little differently and develop its own website and tracking systems.  To help Target in its endeavor to gain control for the first time over its online website and sales, the company recruited help from more than 20 vendors, which included IBM, Endeca, Infosys, and appointed SapientNitro as its partner and lead integrator (Zmuba & Patel, 2011). 

At the beginning of this new venture, things didn’t seem to go exactly as Target might of hoped.  According to the AdAge article published shortly after the redesigned target.com was launched, Target's Site Plagued by Glitches, Friction Between Marketing and Tech Teams, the company faced a long list of problems only six weeks after the site launch which included broken links, missing registries, and shopping carts that seemed to be doing more of the shopping than the actual customers. 

Although the company may not have been up to speed on how to create a website and use analytics in a timely fashion to gain a better understanding of where on the site customers were having the most problems, Target was already ahead of the game when it came to tracking customers data online.  Target statistician, Andrew Pole had already developed a system for how Target would collect customers (or “guests” as Target refers to them as) information in order to effectively market to them efficiently as possible. 

According to a blog post from Avinash Kaushik, the diagram created and presented by Pole at a predictive analytics conference showed “the variables in play that Target collects (across online and offline touch points, mobile and desktop) and ties to the Guest ID (you) in order to do better marketing.”  Below is Pole’s diagram titled Bringing It All Together: Guest ID:

As you can see, this model has various paths that data can take in order to link a user to their unique Guest ID. 

If you ask me, that’s a pretty incredible amount of data Target is able to collect about each customer and store it under each customers own Guest ID.  This model allows the company to analyze every point of interaction the brand has with every customer and combining customer behaviors both online and offline to determine the best way to market to them.  Back in 2010, Pole also stated, “that Target was able to associate half of its in-store sales, nearly all of its online sales and about a quarter of all online cookies with specific Guest IDs” (Hill, 2012). 

In his presentation at the predictive analytics conference, Pole explained the personal information Target would collect for each customer in order to associate them with his or her Guest ID.  “It starts with name, address and tender (the credit card or debit card you use) and expands from there to a history of your store purchases, online purchases, mobile phone ID, actions taken in response to Target emails, and Internet browsing activity if you click on a link in one of those emails” (Hill, 2012).  Additionally, according to Hill (2012), Target uses its combined analytics system and location data to steer customers living near a competitor to shop at target.com instead. 

Target also has a potential value model that allows the company to determine how much each customer could potentially spend.  Also in his conference presentation Pole explains, “With data mined from past years’ spending and demographic databases — showing whether Antonia’s married, whether she has children, what her job is, the average income for her neighborhood — ‘we think she should be spending $5000’” (Hill, 2012).  So, if Antonia’s data reveals she was only spending $1,000 a year, Target will continue marketing towards her.  However, if a customer is spending $1,000 a year and the data leads Target to believe that is all he or she can spend based on discretionary income, then Target will stop marketing towards that customer in order to save the company money (Hill, 2012). 

Target also uses customer data both online and offline to determine the coupon value they will offer to each customer.  If the data for a specific customer from both online and in-store coupon redemptions reveals that a $1 coupon is enough to get that person to make a purchase, then the company will likely continue offering coupons for that amount. Target personalizes customer coupons so the value is equal to the amount that the data predicts will push them to make a purchase.

In its current Privacy Policy, Target states that the company may automatically connect information they already have about you in order to identify your Guest ID.  Below is the Automated Information Collection section of the Privacy Policy:

Automated Information Collection

We may connect information collected automatically with information we already have about you in order to identify you as a Target guest. If we are able to identify you as a Target guest we may, for example, link your activity on our website to your activity in a Target store. This allows us to provide you with a personalized experience regardless of how you interact with us – online, in store, mobile, etc.

Automated Information Collection

We and our service providers use cookies, web beacons, and other technologies to receive and store certain types of information whenever you interact with us through your computer or mobile device. This information, which includes the pages you visit on our site, which web address you came from, the type of browser/device/hardware you are using, purchase information and checkout process, search terms and IP-based geographic location, helps us recognize you, customize your website experience and make our marketing messages more relevant. This includes Target content presented on another website or mobile application, for example, Target Weekly Ad. These technologies also enable us to provide features such as storage of items in your cart between visits and Short Message Service (SMS)/text messages you have chosen to recieve. We also use Flash cookies for fraud prevention purposes.

In order to provide the best guest experience possible, we also use these technologies for reporting and analysis purposes, such as how you are shopping our website, performance of our marketing efforts, and your response to those marketing efforts.

Overall, Target has developed and implemented a sophisticated methods and technologies to track customers behaviors both online and offline to determine the most efficient and effective marketing efforts.   In terms of creating a better experience online for customers (and that I’m sure Target is already doing) A/B testing could be used to help with effective site design.  I would also suggest using a heat mapping tool because from my own personal experiences I think the homepage can become a little cluttered, especially during the holidays, making it harder to navigate. 

References:

Hill, K. (2012, February 24). Target isn't just predicting pregnancies: 'expect more' savvy data-mining tricks. Retrieved from http://www.forbes.com/sites/kashmirhill/2012/02/24/target-isnt-just-predicting-pregnancies-expect-more-savvy-data-mining-tricks/

Kaushik, A. (2012, April 12). Retrieved from https://plus.google.com/ avinash/posts/f5K1ueN9Tk1

Target. (2013). Corporate fact sheet. Retrieved from http://pressroom.target.com/corporate

Target. (2013). Privacy policy. Retrieved from  http://www.target.com/spot/privacy-policy 

Zmuda. , & Patel (2011, October 06). Target's site plagued by glitches, friction between marketing and tech teams. Retrieved from http://adage.com/article/news/target-faces-hurdles-site/230188/

Seeing the Big Picture

As I’ve mentioned in previous posts, Google Analytics definitely takes some time getting use to.  It’s easy to become wrapped up in all of the numbers and graphs when first setting up tracking for a website, which is exactly what I did when I was first introduced to the tool six weeks ago.  But, each week I have found it easier and easier to navigate the dashboard and have become much more comfortable with the overall layout of GA.  I have also gained a much better understanding of the valuable information each of the different measurements and metrics has the potential to provide and how different metrics can be used together to provide even more in-depth data. 

This week’s lesson examined Google’s customizable goals, funnels, and filters, which really helped me to put the last few missing puzzle pieces together and see the “big picture”.  Understanding the basic concepts of page views, unique visitors, bounce rate, etc., is definitely important in order to use the tool effectively, but those metrics alone can’t provide you with the data needed to determine the success of your website.  Plus, the critical factors for determining success are unique to each business, which is why using goals, funnels, and filters are so important. 

Goals

Google defines goals as being “a versatile way to measure how well your site or app fulfills your target objectives” (About Goals, 2013).  In other words, goals allow you to tell Google what to keep track of based on what’s important to your business.  GA provides four ways for users to track goals:

1. URL destination
2. Visit duration
3. Pages/visit
4. Events

URL Destination

This goal allows you to track specific URL’s by specifying a page’s unique URI (Uniform Resource Identifier), which are the characters that come after the domain name.  For example, if the URL is www.mybusiness.com/aboutus, the “Goal URL” you would enter into Analytics would be “/aboutus”.  Each time someone visits that URL, they will trigger the goal (Lofgren, 2012).  This goal type is often used for tracking sales by using the confirmation page URL or for tracking new subscription or account sign-ups via a Thank you for registering webpage.

Visit Duration

The visit duration goals allow you to track how many people stay on your site for a specified amount of time.  This goal gives you the option to track whether visitors stay "less than" or "greater than" the amount of time you specify.  For example, a support site looking to answer customers’ questions as quickly as possible might want to track how many visitors stayed less than 5 minutes.  On the other hand, a company that has just made design changes to its site may want to track how many users stay longer than 2 minutes to help determine whether or not the changes made are leading to a higher level of customer engagement.

While this goal type can be very useful, one thing to keep in mind is how Google tracks visit durations.  GA only creates a timestamp when a visitor loads a page but not when a visitor leaves a page. So, in order to calculate the time spent on one page, two timestamps (one when the first page visited is loaded and another when the second page visited is loaded) are needed.  Ultimately this means Google Analytics is unable to calculate the visit duration for exit pages.

For example, if a visitor enters a website on the homepage, spends a minute or two browsing, then clicks to another page on your site, Google will calculate the time spent on the homepage by subtracting the timestamp created when the second page was loaded from the time when the homepage was loaded. But, if a visitor only visits one page and leaves, Google doesn’t have a second timestamp to calculate the time spent on the first page.  If a visitor visits 3 pages, only the visit duration for the first and second pages can be calculated. Google documents the time for page visits with only one timestamp as 00:00:00 regardless of how long the visitor actually remained on the page.

According to the KISSmetrics blog post, 4 Google Analytics Goal Types That Are Critical To Your Business, this will likely result in the time on site metric being very different than the actual amount of time people spend on your site.  However, you can still gain valuable from this goal type by comparing the metric over time and looking for any trends in how the metric changes month-to-month or over an extended period of time (Lofgren, 2012). 

Pages/Visits

This goal type is very straightforward and tracks the number of pages a visitor visits before leaving your site.  Similar to the visit duration goal, you have the option of tracking greater than, equal to, or less than the number of pages visited that you will specify during the goal set-up.  The “greater than” condition is typically used to measure visitor engagement and the “less than” condition is often used by support sites to measure the site’s effectiveness (Lofgren, 2012).

Events

Google defines events as “user interactions with content that can be tracked independently from a web page or screen load” (About Events, 2013).  Examples of actions you can track as events include downloads, video plays, mobile ad clicks, widget usage, social media buttons, etc.

However, before you are able to track an event goal, you must first take the time to set up event tracking and have created at least event category.

Funnels

A funnel can help take a URL destination goal to the next level by tracking the path visitors take towards the end goal.   Because there are typically several other pages preceding the goal page that are designed to push visitors towards the goal, funnels can help pinpoint any problems along visitors’ conversion paths, giving you the chance to make any necessary changes. 

According to Fettman (2012), funnel tracking can help:

• ·      Identify any steps creating trouble or confusion for customers
• ·      Determine if there is any language or copy that is causing changes in customers’ emotional         behaviors during sign-up or checkout
• ·      Detect any bugs, browser issues, or any other technical troubles

Additionally, funnels can help you see how often visitors abandon a goal and where these visitors go after abandonment.

Funnels can be especially beneficial for ecommerce sites because visitors typically follow only a few paths before landing on the confirmation page.  By using funnels, you can see if there are any pages along the path to conversion that consistently result in shopping cart abandonment.  If a trouble page is identified, you can revisit the page and make any necessary adjustments in order to make the checkout process easier and more streamlined.  By monitoring the funnel tracking, you can then determine whether or not the changes made to a specific change are resulting in more or less shopping cart abandonment than before.

Filters

In her article Going Beyond Standard Reporting with Google Analytics Filters, Segments, Reports, and Dashboards, Hines (2013) explains, “View Filters allow you to include or exclude specific information in your Google Analytics reports, focus on a specific subdomain or directory, or rename URLs to make them easily recognizable.”

There are 3 predefined filters, which include:

• ·      Traffic from domains:  This filter allows you to include or exclude traffic from a specific domain.
• ·      Traffic from IP addressed: You can use this filter to include or exclude clicks from certain sources.  This can be a single IP address or a range of addresses.  This filter is commonly used by companies to exclude traffic from internal IP addresses in order to avoid measuring traffic from employees or an IT person conducting site maintenance.
• ·      Traffic from subdirectories: Through this filter you can include or exclude traffic from a specific subdirectory (such as www.example.com/education).

In addition to these predefined filters, Google Analytics also provides numerous custom filters.  Some of these filters include the Lower/Uppercase filter, which converts all field contents to either all lower or all uppercase letters, and the Search & Replace filter, which allows you to search and replace a confusing pattern within a field and replace it with an alternate form that is easier to recognize.

Conclusion…

Goals, funnels, and filters can help you go beyond the standard reporting offered by Google Analytics. 

• ·      Goals allow you to track the data most important to your company
• ·      Funnels are an extension of Destination goals that allow you to see the number of people that move through each step of your specified marketing process to help determine pages that may need to be fixed
• ·      Filters allow you to modify your traffic data by using the 3 predefined filters to include/exclude traffic from a specific domain, IP address or addresses, and particular subdirectories 

These traffic parameters can greatly benefit businesses by providing valuable information that can be used to track the most important factors for a company’s site, quickly identify problems in a key process to help create a better user experience, and to customize reports in order to effectively analyze and share in-depth data.

References:

Fettman, E. (2012, November 26). The google analytics conversion funnel survival guide. Retrieved from http://blog.kissmetrics.com/conversion-funnel-survival-guide/

Google. (2013). About events. Retrieved from https://support.google.com/analytics/answer/1033068?hl=en

Google. (2013). About goals. Retrieved from https://support.google.com/analytics/answer/1012040?hl=en&ref_topic=1007030

Google. (2013). About view filters. Retrieved from https://support.google.com/analytics/answer/1033162?hl=en

Google. (2013). Goal types. Retrieved from https://support.google.com/analytics/answer/3046660?hl=en&ref_topic=1007030

Hines, K. (2013, August 27). Going beyond standard reporting with google analytics filters, segments, reports, and dashboards. Retrieved from http://blog.kissmetrics.com/beyond-standard-reporting/

Lofgren, L. (2012, May 01). 4 google analytics goal types that are critical to your business. Retrieved from http://blog.kissmetrics.com/critical-goal-types/