Here’s an important question that many search marketers never think to ask:
Q. How many different types of website visitors do you have?
The answer is easier than you thought, right? We should qualify that statement: For the purpose of this test we can divide your users into two camps:
- Those who know your site (or business)
- The “newbies”
Newbies are users who have no prior knowledge of your site or indeed your business proposition. These people are immensely important to you—they are the ones you’ve spent so much time and money getting to your site. They are future loyal customers, the drivers of growth and prosperity and healthy looking ROI charts. But not if you lose them at first touch.
This is where those often quoted statistics about how many seconds you have to engage a user when they land on your site come into play. Whether the official figure be 2, 3, 5 or 0.5 seconds, we’re all united in the belief that it’s not long. We also all know that we need to do everything we can to keep these people on the site long enough to channel them towards behavior that both satisfies them and helps achieve our goals, turning them into customers.
Whether your site’s purpose is ecommerce, data capture, whitepaper download or something altogether different, it’s almost certain that there is some kind of action you want users to complete. Channelling as many of your prospects as possible to this action should be pretty high on your “list of stuff to do on the site.” First step, make sure you have some way of tracking these actions and that you know what your conversion rate is.
Segment & conquer
There are problems inherent in defining a “new” or “returning” visitor in all web analytics solutions. For that reason we use segmentation to refine the definition somewhat. Here we want to see site usage data for those users arriving at our site completely that have never visited previously. All they know is that they have just clicked on a link (be it paid or free), most likely from a search engine result page, and voila, here they are! With that in mind, the segment we create must exclude all users who we can reasonably assume have some experience of your website or prior knowledge of your business. For our website we can define this as anyone who fits (or doesn’t fit) these criteria:
For your site there may be more criteria. Take some time to get this right and make sure you really have excluded all unwanted traffic. For example: subdomains, or any strongly linked sister sites, email campaigns, etc.
As an exercise, this is a good intro to the immensely powerful “advanced segmentation” feature you’ll find in the top right of your dashboard in Google Analytics these days. If you’re running a different package, fear not: most web analytics solutions will allow you to segment traffic in some way.
Identify the right pages
Start by gaining an understanding of which site content is being accessed most frequently by our prospects. Apply the segment you just created to the “top landing pages” report. You might be surprised by what you see in the top landing pages report. We have seen many times that one page is optimized better for natural traffic while another is preferred for paid search traffic. Remember we’re talking about first impressions here so we’ll base this analysis on landing pages.
Once we know which page(s) they are, we can set about understanding two things:
- How these visitors are reaching the site
- What they are doing (or not doing) once they are on there
Answering these questions will help you to decide if this is good quality traffic that should intuitively be converting at a good rate. For example, are the sources either good quality and/or keywords highly relevant? If the answer is yes then this page is worth optimizing for conversions. If no, you can refine your segment to exclude these elements and re-evaluate.
Next, get an idea of how often users landing on those pages are converting. A good tip here is to add a landing page qualifier to your segment and duplicate the segment, one-per-page analyzed. To get a full picture, use a mixture of conversion rate, bounce rate, time on site and average page views to form this opinion. One of these metrics alone may not give a true picture. You may decide to weight it in favor of one or other metric depending on what is of value to your business. In most cases, goal conversion rate should be the focus here.
Analyze & theorize
To recap, you have segmented your traffic into newbies and old-hands; you’re focusing on the behavior of these newbies. Now you even know which pages they land on. Hopefully, you’ve worked out at what rate they are converting into customers and decided whether you think this is reasonable or not.
Now the hard part: being objective with your own website is truly a difficult thing. Putting yourself into the shoes of one of your new prospects doesn’t make this any easier. So have a stretch, make a cup of tea, crack your knuckles and, when you’re ready, load new prospect landing page number one. So, what is your first impression? Within 5 seconds can you identify and satisfy the answers to the kind of questions your new prospect is asking? For example:
- Do I know what this company does?
- Do I know what their price/quality proposition is?
- Have I identified that they have the right product/service/information for me?
- Most crucially: do I know what to do next? (Channelling towards the goal)
Be honest! You might have spent days crafting that synopsis of your professional life to give your site an air of authority. But does it answer the above questions? Your SEO-optimized homepage text might be great for search engines but does it really provide a succinct synopsis of what your business does? Likewise, another issue we’ve seen frequently is the confusion of different possible actions. Imagine a user is satisfied with the answer to the first three questions above. All they need now is a quick, easy way to proceed to the next step in the process. Does your website provide this for them? If you are dealing with internal sensibilities relating to the website design, this process will help you build a case for change.
You should now be left with a few informed assumptions/hypotheses about your website. It’s good to hypothesize so long as you don’t confuse an assumption with an empirically proven truth.
Test these assumptions, prove or disprove your theories. Depending on what these are, you will test different things. Here are some common examples:
- Think your signup process may be too long? Cut the form short and observe conversion rate, ideally test side by side.
- Think your homepage may be too wordy? Create a B version and A/B test.
- Think the call to action isn’t clear enough? Add a clear call to action to the landing page
There are some great tools to help here. We’re big fans of the free Google Website Optimizer. There are plenty others out there. Use a program like this to test different versions of key elements in either an A/B or multivariate test.
Whatever testing you conduct, remember to be objective with the results. It’s also crucial to have the right measurements in place. Once you’ve made an improvement, it doesn’t have to stop there. Move on to another segment or another portion of your site, measure, hypothesize and test again.
Opinions expressed in the article are those of the guest author and not necessarily Search Engine Land.