A common urban legend claims that we only use 10% of our brains. The validity of that claim may be debatable, but here's one that isn't: Marketers use less than 10% of the rich visitor segmentation information available from their web sites. How can you push your web analytics brain to the limit?

Each site visitor brings a unique set of characteristics that can help you better understand the performance of your online and offline marketing. Web analytics beginners struggle to grasp the significance of attaching properties to visitors, opting instead to look at top referrers, and the behavior of the average visitor. Don't make this mistake. Experienced marketers know how to group and compare visitors for analysis that lends itself to marketing decisions.

It's not about a single visitor

When considering visitor segmentation, the newbie often asks if a web analytics tool can track the path of a single visitor. They aren't asking the right question. In fact, the question has nothing to do with segmentation. Individual visitor behavior, while useful for lead qualification, is not useful for making marketing decisions. You should make marketing decisions based on your comparison of aggregate visitor behavior and nothing less.

It's about comparing group behavior to a baseline, or others

Imagine your favorite line graph. On most web sites, basic statistics make attractive line graphs, charting metrics up and to the right. More visitors, more page hits, more referrers - more, more, more. While this looks good to the less-savvy CEO or investor, it doesn't empower marketing decisions. The key to good marketing decisions is the ability to extract identifying information that suggests the behavior of discrete groups, and then respond to the behavior of these groups with improved offers and product representation.

A good web analytics tool will be able analyze and segment visitors based on:

1) Where did the visitors come from and what were they looking for?

2) What did visitors do while on my web site? Where did they go?

3) What information did visitors reveal about themselves?

Common trends in visitor behavior lets us segment, measure and compare the behavior of each group. Let's take a closer look at each characteristic as it applies to a group:

Where did the visitors come from? What were they looking for?

Web analytics tools can detect the referring site of a majority of visitors. The referring site provided a link to your site. But do you know why? Perhaps you're blogged on the site. You may have an advertisement or a mention in a forum. Is this a partner's web site? Did the press cover you? With analytics you can stop guessing and track it down.

When you have control over the URL used to direct traffic to your site, we encourage you to add a parameter to indicate the source of each link. For example, a link to yoursite.com might look like this: www.yoursite.com?source=3384LblArt. Your web analytics tool can be set to label all visitors who had a parameter containing 3384.

In the case of radio, television or print promotions, a unique landing page or URL provides a simpler way to use web analytics to segment visitors. For example, a radio or billboard ad might use the URL save.yoursite.com. Your web analytics tool would then be set to label visitors whose first entry page was save.yoursite.com as visitors who came from this particular print or offline campaign.

The referring site information also reveals which search engine referred the visitor, and can often reveal which keywords the visitor used to reach your site. Segmenting and comparing visitor behavior based on keywords can help you to optimize your site for particular keywords.

In the example on the right, the keywords used by buyers are not the same keywords used by the average visitors. The experienced online marketer knows to optimize for keywords that bring buyers and visitors who spend a longer than average time on site. Without segmentation, this information is invisible.

Good segmentation tools let marketers compare the behavior of visitors referred from a particular site to the "average visitor" in terms of time on site, revenue generated, conversion to leads, newsletter sign-ups, pages visited, number of follow up visits and other measurements that can influence your marketing decisions.

What did the visitor do while on my web site?

Another common newbie request is the "most popular path". Once a web site has more than 500 visitors and more than 10 pages, a "most popular path" becomes statistically insignificant. It's much more meaningful to note which pages were visited most frequently, which pages led to certain behavior, and where certain types of visitors spent the most time.

Your web site exists for a purpose (or more likely, a number of interconnected purposes), and certain pages or parameters can indicate to you when visitors behaved the way you wanted them to. For example: We can, using JavaScript values or URL parameters to measure the total revenue generated from visitors who read our e-newsletter. But it's more meaningful for us to compare revenue (or some other conversion or time-on-site metric) between visitors who read the newsletter and visitors who respond to a particular promotion. Once we know which promotion is generating higher quality visitors, we can confidently determine where resources should be allocated.

Corporate sites, or sites that are aimed at lead conversion, reveal qualifying information through completion of a request form, newsletter sign up, click to another site, etc. If these are your goal pages (or behaviors) then you've got to label visitors who reach them and compare behavior to the average visitor, or the visitors who leave the site without converting.

A special note about buyers and window shoppers: Believe it or not, revenue isn't always the best indicator of the value of a visitor to your site. Sometimes, revenue may be delayed for weeks or months - too long to wait to make marketing decisions. Instead, marketers should learn to use and compare average time on site as an indicator of marketing effectiveness.

When a visitor exits the site before reaching one of our goals, the last page visited offers a clue as to why the visitor decided to leave. Did the visitor abandon a complex shopping cart? Could the visitor find your newsletter page? What were the top pages seen by visitors who exited the site on the shipping page? Often, exit stats lead to content modification and navigational improvements.

What information did visitors reveal about themselves?

Beyond keywords that brought visitors to the site, the search keywords that visitors use within the site often reveal what they are interested in. If your site sells fruit, and visitors are most frequently searching for balloons, there may be a discrepancy in what they expected, or better yet, an opportunity to expand merchandise for increased revenue.

In one case, a web site owner noticed that many visitors were searching for "FedEx" and exiting the site after the search came up empty. Addition of overnight shipping increased revenue by 18% in the first month after adding the service.

Your site can also reveal form parameters when a visitor fills out a form or checks a box. For example, a form that asks for a zip code might reveal zip=95060 to the web analytics tool. A submitted form that required age or gender may reveal age=35&gender=F. This provides an opportunity for segmentation based on revealed information. Now our web analytics tool can answer questions such as: Do women spend more time on site before they buy? Do men over 30 from the Western United States respond to magazine ads better than PPC ads?

In the example below it is clear that female buyers spend more time on site than the average buyer. The next step would be to learn what interests them by analyzing which pages they linger on and what keywords they use in their internal searches. These discoveries will help you make better, more relevant offers to targeted groups.

With visitor segmentation, it's clear that men over 30 from the Western US overwhelmingly respond better to PPC ads (which also yield a better return on advertising spend than print ads). However, you'd probably also notice that groups of men from these two ads also come in slightly under the average return on advertising spend. It might be time to look at keywords used through the PPC ads and cater to what these gentlemen are looking for.

 

Choosing which segments to measure

So how do you choose which segments to measure? Typically, it's one of two ways: Either you will want to perform a purposeful analysis of a marketing event, or you're reacting to a surprising event. Let's look at both.

Purposeful Analysis: This is what happens when you set out to measure the success of a particular campaign or marketing activity. Simply select parameters and attributes that create a visitor profile. This allows you to see the behavior of visitors that match the profile. Examples of purposeful analysis include:

• Comparing the behavior of visitors from Yahoo! to the average visitor after optimizing the site for Yahoo! placement. Typical action: Label all visitors who were referred froun Yahoo's domains, but did NOT come to the site through PPC.
• Comparing the average time on site for visitors who received a direct mail postcard to the average visitor. Typical Action: Label all visitors that entered the site on the landing page indicated on the postcard.
• Determining the most important search keywords: Typical Approach: Rank keywords time on site and generated revenue, create labels for natural search visitors and PPC visitors, and then compare each ranked list between natural search visitors and PPC visitors.

Response to Surprises: This is what happens when you notice an unexpected event or anomaly in your analysis and set out to investigate the source of that odd visitor behavior. Examples of surprises include:

• A disproportionately high exit rate on the shipping page of an e-commerce site. Typical action: Label visitors who exited at the shipping page AND spend more than average time on site, compare their internal search phrases to the average visitor (or average buyer), and then adjust content to meet visitor expectations.
• A sudden spike in visitor count from a particular PPC campaign while others remained steady. Typical action: Label visitors who came from this particular campaign and compare time on site and page visits to the aggregate of visitors from all other PPC campaigns. Compare country of origin and ensure that your ad is or isn't targeted to that country. Use a fraud reporting tool, like the one found in Lyris HQ Web Analytics (formerly ClickTracks), to analyze the ad for fraudulent activity.
• A sudden drop in referred visitors from a partner site. Typical action: Label visitors who were referred from this site, then check the top pages visited by this group. Ensure that the expectations set on the partner site match those of the pages most visited by this group.

By now, you should begin to understand the unlimited nature of analysis based on visitor segmentation and comparison of visitor behavior. Be sure to implement your online and offline marketing in a way that lends itself to analysis through web analytics. Group and compare visitor behavior to make better marketing decisions, and ultimately, use 100% of your web analytics and your web analytics brain.

Comments (0)

Subscribe to this comment's feed

Name

Email

Title

Comment

smaller | bigger

Add Comment