Archive for the 'Web Analytics' Category

We all share the mutual goal of getting traffic to our website. But what do you want to come of this traffic? You want action!!! You want this traffic to buy a product, join a mailing list, request more information, and complete a contact form, of course.

Wouldn’t you like to know how these actions are coming about? What part of your wonderfully user-friendly, well optimized website is inducing these actions to be completed?

Here’s where goal configurations come in. Many people are using Google Analytics on their site, however they are not taking advantage of the goal configuration component. The concept is straightforward – capturing when and how a visitor to your website is completing a stated action.  This is especially important if you are selling something on your site, because this is how you will determine your return on investment.

Now I’m going to skip over the Google Analytics basics (If you need to brush up on the basics, check out Google’s online tutorial) and take you straight though to the goals process.  In the main Dashboard, click on the Edit link at the right and you will come to the Conversion Goals and Funnels screen.

Google allows you to define 4 goals. This is a great restriction, because if you have more than 4 actions that you want your visitors to complete, well then your website is a big unfocused confusing mess. In my experience, I never really configured more than 2 or 3 goals anyway.

What you need to do is:

1. Activate the Goal.

2. Put in the match type (this will be different for a static site and a dynamic site). An EXACT match means that the Goal URL you enter is exactly as it will appear on the site. A HEAD match means that you have a dynamic site, so the goal URL will be different every time - except the beginning, or HEAD of the URL will be the same. Just the end part will be different.

www.yourwebsite.com/?Session_ID=9a0f0559368fb1a39dac93b3b50ace50& would be an example of when you put in HEAD match. There is another option called REGULAR EXPRESSION MATCH, which I discuss below.

3. Enter the Goal URL.  This is the page that visitors reach AFTER clicking the submit/purchase/join button. For example, your Goal URL might be: www.yourwebsite.com/thanks.

4. Name your goal: Contact Form. Mailing List. Download. Inquiry. Lead.

5. Set a goal value. If you are not selling something that has a fixed value, you can leave this blank. But perhaps your lead is worth $500 to you. You can put that number in here.

Now you are done setting up goals.

One thing I want to mention is the importance of having separate Goal URL pages. Say a visitor submits a contact form on your site. After he clicks the submit button, is he taken to a brand new page that says “thank you for your inquiry/comment/request”?  Or, do the words “thank you” simply appear on the screen, on the same page you were on before?

If you do not have a separate THANK YOU page, create one or have your developer create one immediately. Otherwise, the goal configuration is not going to work.

Say you have more than one action to complete on your site. Take a look at the screenshot below.

In this example, we have a simple contact form, and then we also have an appointment request form. You want to make sure that the “thank you” URLs are different from each other, because otherwise, Google Analytics will not be able to tell which goal is being completed. So just make sure you have two (or three or four) separate Goal URL pages. They can look like this:

• www.yourwebsite.com/thanks1
• www.yourwebsite.com/thanks2
• www.yourwebsite.com/thanks3

or

• www.yourwebsite.com/contact-thanks
• www.yourwebsite.com/appt-thanks
• www.yourwebsite.com/signup-thanks

As long as you distinguish one from the other.

Now on to Goal Funnels. This is where you establish each step your visitor must take to ultimately achieve his goal.

Let’s determine the possible steps of purchasing a pair of shoes online.

1. After you find the shoes you like, you Add to Cart. (www.yourwebsite.com/add-to-cart)

2. You are done shopping, so you want to Checkout. (www.yourwebsite.com/checkout)

3. You will be asked for your Customer Information. (www.yourwebsite.com/customer)

4. Then, you will be asked for your Credit Card Information. (www.yourwebsite.com/credit-card)

5. You will choose a Shipping method. (www.yourwebsite.com/shipping)

6. Review Your Order. (www.yourwebsite.com/review)

7. Complete Purchase. (www.yourwebsite.com/THANK-YOU), where this last URL is your GOAL URL.

The above are just examples and of course vary based on the type of website you have. Below is an example of a funnel, where all of the checkout information is on one page of the site, so of course you see only two steps.

That’s it as far as the set up process goes. It’s very important to establish a funnel if you are selling something on your site. It is not uncommon for visitors to abandon shopping carts in the middle of the checkout process. Don’t you want to know when this is happening, so you can figure out why? Maybe many visitors leave at the Shipping method page. Guess what - that sounds like your shipping rates are too high or you don’t offer enough shipping options.

A few months ago, I was contracted to implement SEO on an e-commerce site run on Miva Merchant. When it came to configuring the Analytics, we had quite the challenge. Whoever had set up the site had incorrectly installed and configured Analytics the old fashioned way - cut and paste the snippet of code, and sloppily defined some goals. However, the actual sales reports were clearly not matching the data recorded in Analytics - the numbers were way off. So we did some detective work, went through the entire buying process on the site to determine the various steps and URLs in the funnel.

As it turned out, each URL during each step of Miva Merchant’s checkout process was the same. We learned that goal configuration for a Miva Merchant Site is clearly a different process than it is for a regular site, and we had to contract a Miva specialist to help us out. In fact, even a different Analytics tracking code was necessary to install in the site-wide footer.

Take a look at the screen shot below to see how we had to set this up. We could not use HEAD match in this instant, because of the nature of the URLs generated through the Miva CMS. So we had to use REGULAR EXPRESSION match.

After this was properly set up, the data recorded began to accurately reflect the products sold.

Online Marketers spend a good amount of time and money designing websites and campaigns to drive visitors to the site. When a visitor clicks though from a banner, marketers need to ensure that the visitors just doesn’t remain a visitor but becomes a customer. In order for a visitor to become a customer the value proposition needs to be clear and the landing page, the page visitors enter the site from, needs to provide a compelling reason for visitor to stay on that site and become a customer. The actual process of becoming a customer might involve few more pages on the sites but the landing page is the first impression that a visitor has about the site.

Marketers need to constantly optimize their landing pages (and other pages) to ensure they can get more and more visitors engaged and drive them though the conversion funnel.

The key to any landing page optimization campaign is to ignore your own biases and listen to your customers. Your customers are your best landing page designers. Their actions on the site tell you how they feel about your it along with what is working and what is not. They will tell you which landing page design is best convincing them to become customers or do what you want them to do.

Everything that you have on a page impacts visitors actions. Every little thing contributes the page’s success or failure. Which means, you can test pretty much everything on your web page. After testing for while you will figure out things that do make impact and things that do not have any significant impact and are not worth testing going forward. Examples of things that you can test are:

• Headlines
• Images
• Text Size
• Font Color
• Place of images and text
• Call to Action
• Buttons/Text/Images – Size, Color, Font etc.
• Promotions – example: Does 20% of $100 works better than $20 off on $100?

Believe me, simple changes can lead to very significant gains, I have seen and experienced it over and over again. Only your customers can tell you what will work on your site and you won’t know this till your start testing.

If I told you that I recently increased the conversion rate by 65%, just with few minor changes, will that get you excited? Won’t you like to have that kind of increase in your conversion rate? Do the math and find out why you should be testing right now. (I will share the example a little later)

Convinced? Ready to know how we can start testing?

Get your web analytics tool in action. As I have written in past on my blog, don’t just do reporting, analyze the data and see what your customers are telling you. Start testing the landing pages which are bleeding the most visitors, say your home page for example. If it is the top entry page and you see that it has a bounce rate higher than 30% then that might be the low hanging fruit that you might want to go after first.

The Process of Testing

A/B Testing:

A/B (or A/B/C/etc.) testing is the simple and easy way to explore the world of testing. A/B testing is a process of testing multiple versions (version A, B, C etc) of a page to see which page is better than others (statistically significant) in driving user take action that you want them to take e.g. drive user to signup for email newsletter.

In A/B testing you (the tool you implement on your site) randomly sends the visitors to one of the various versions of the pages. The tool then tracks the number of visitors and those who take the specified end action as defined by you based on your business goals.

Once enough visitors have been exposed to the test to get a statistically valid result, the tool will declare the winning page version. (Note: Depending on the complexity and goals of your business there might be more data that you will need to look into to determine if the winner declared by the tool is actual winner or not and if there are additional test you will need to run before you can find the winner.)

Multivariate Testing:

Multivariate testing is powerful way of testing specific elements within each page that you are optimizing. When your optimization tool declares a winner of an A/B test, it tells you which version of the page is best. It does not, however, tell you which elements within the winning page caused the increase in conversions. This is where multivariate testing can help.
In a multivariate test, you test multiple elements on a page (headlines, buttons, text, images) against different versions and combinations of those same elements. I know it sounds confusing, so let me give you a simple example:
Say you are testing a headline and a button. Headline 1 says, “Sign-up to get access now.” Version 2 of this headline might say, “Become a Member and Get Instant Access.” You might have version 1 of your button say “sign-up.” Version 2 could say “Get Access.”
Your multivariate test would test Version 1 of your headline with Version 1 of your button against version 1 of your headline with version 2 of your button, and so on for each of the possible combinations of headlines and buttons.
One of the major drawbacks of the Multivariate testing is that you need to have a lot of traffic to get statistically valid (accurate) test results. Depending on which tool you use, you will need a different amount of traffic. A/B testing does not require as much traffic for valid results.

I hope this helps get you started with optimizing your website.

Ready for an Example?

Here is an example of the page that I ran on RoommateHub.com.

Original Registration Page

Version 1 of Registration Page (Can you spot the differences between original and this version)

Google Website Optimizer Declares the Winner. 65% Increase in registrations, how cool is that.

There are few good books, websites, and other resources on the subject. Here are two books that I recommend:

Further Reading

1. Always Be Testing
2. Landing Page Optimization: The Definitive Guide to Testing and Tuning for Conversions

Happy Testing.

Questions? Comments? You can reach me at batraonline at gmail.com

—————————————————————————-
Looking to fill your Web Analytics or Online Marketing position?
Post your open jobs on http://www.web-analytics-jobs.com/
—————————————————————————-
Site: AnilBatra.com
Twitter: http://www.twitter.com/anilbatra
Blog: webanalysis.blogspot.com

Guest post by Carmen Sutter. She’s on my team, and specializes in Web Analytics.

So you finally implemented web analytics. Now what? What metrics should you look at? What do those metrics mean? As always with questions like this the answer is: It depends. It depends on what type of site you are running. It depends on your goals for your site. It depends on the marketing around your site, if any.

Over the next couple of analytics posts we will help you explore some of the metrics you should know.

For any type of site, bounce rate is probably one of the more important metrics. The bounce rate tells you the percent of visits that only look at one page. Literally, how many people “bounce” right off your site. This is a metric you want to observe in aggregate for your entire site as well as on a page level.

There are a number of reasons for high bounce rates.

1) Maybe your content is not as fresh as it could be. Evaluate how often you update your content and see if you can improve on that and if that makes a difference in the bounce rate.

2) Maybe your site is not the site your visitors are looking for. Take a look at the search engine keywords that drive traffic to your site. Make those keyword sense for your site? Let’s say a large number of visitors come to your site via the keywords ” cookies history”. You could be a tech blog or a baking site - make sure your search listings are optimized to help potential visitors know what to expect when coming over to your site. And then evaluate if your bounce rate changed at all.

3) In sticking with the search keywords topic, the bounce rate is even more critical for your paid search campaigns. No matter if the goal for the PPC campaign is a purchase or lead generation, if your visitors don’t make it beyond one page and bounce right off your site, you are wasting money. Make a sure that your PPC campaign is targeted and brings in qualified traffic.

You’ll probably wonder what constitutes a good bounce rate and my answer, like for most metrics, is to benchmark against yourself. If you have a bounce rate of 80% today, aim for 75% next month. If you have a bounce rate of 10%, aim for 5%. We all have the tendency to want to compare against others, but there are so many factors affecting benchmark data. On the other hand, if you know what’s going on with your site, you can control the content and see how it impacts your own metrics over time.

In the next post we’ll take a look at important metrics for commerce sites.
———————
Carmen Sutter has extensive experience with Omniture, Google Analytics & Coremetrics. She is experienced in implementations, customizations, integration services, and Reporting & Benchmarking.

By streamlining reporting processes, performing in-depth analysis, and understanding customers’ level of engagement, Carmen helps measure and improve our search marketing campaigns.