Hack89.Measure New and Returning Customers


Hack 89. Measure New and Returning Customers

One of the keys to successful web optimization lies in effective segmentation of your visitor population. One of the most obvious areas of segmentation is simply evaluating behavioral and conversion differences between new and returning customers.

New customers are exploring. Returning customers often know what they are looking for. Does your site make it easy for both groups to find what they need? Or by optimizing in aggregate, have you made it ideal for nobody? The only way to know for sure is to effectively differentiate between these groups, and then explore page-by-page differences in click and conversion behavior between the two segments.

So, how can you optimize your chances for accurate differentiation of new and returning customers?

At least for now, the best option is to ensure that the persistent cookies set on your visitors' machines are set in your domain (in other words, first party), even if you're using an ASP analytics provider. There are several options for making this happen.

6.10.1. Serve Your Own First-Party Cookie

The first and best approach to implement a first-party cookie via an ASP script is to simply serve the cookie from your own web site. The cookie must contain a unique visitor identification code, typically a random number or alphanumeric string that is passed back to you with each page viewed on your site. To then pass that ID value to your ASP as a parameter with the image request, simply make a small modification to your tagging script. Something like this (where MYFIRSTPARTYCOOKIENAME is the cookie you'll set and MYSESSIONPARAMETER is the parameter that you'll use for sessionization):

 function fpc(MYFIRSTPARTYCOOKIENAME){ var cookies = document.cookie; var pos = cookies.indexOf(MYFIRSTPARTYCOOKIENAME + "="); if (pos != -1){ var start = pos + MYFIRSTPARTYCOOKIENAME.length + 1; var end = cookies.indexOf(";", start); if (end == -1) end = cookies.length; var cookievalue = cookies.substring(start,end); cookievalue = unescape(cookievalue); MYSESSIONPARAMETER = cookievalue; } } 

Finally, identify the new parameter as the tracking method with your analytics application.

Of course, before you make any changes to the tracking script of your analytics vendor, you might want to double-check with them to ensure that it will work. It's probably also important to note that not all analytics solutions allow you to specify backup sessionization techniques, so ensure that yours will.

Because the cookie is now being served by your own domain, it is less likely to be subject to removal by anti-spyware programs, and returning visitors will be accurately identified. Note, though, that anti-spyware programs that prevent 1 x 1 GIF image [Hack #29] requests to specific domains (such as those of the ASP analytics vendors) may still prevent visitor tracking. If eliminating this problem is important to you, you must use one of the two following techniques, whether or not you also implement your own tracking cookie.

  • Use web server logfiles and, if using an ASP, talk to them about the ability to upload those logs for analysis

  • Implement your own client-side data collection point, served from your own domain, and then upload that data to the service for analysis

Under these scenarios, you maintain the interaction with your customers and you serve your customers a legitimate first-party tracking cookie, while still maintaining the flexibility of using a hosted analytics solution.

6.10.2. Have Your Application Provider Serve a First-Party Cookie for You

As an alternative to managing your own cookie, hosted service providers can generally serve a first-party cookie instead of the default third-party cookie for unencrypted (non-SSL) traffic. This solution requires you to select a hostname, configure this hostname in your DNS servers as an alias to another host provided to you by your analytics vendor, and modify the tracking code on your web pages to reference the new hostname within your domain. The result is a first-party cookie in your domain that is served by your hosted analytics vendor.

This process involves creating a CNAME record that would look something like this:

 my.internal.network.name. IN CNAME my.vendors.network.name 

Sites that have secure pages, such as shopping cart checkout procedures, require additional procedures. Because secure pages are generally encrypted using SSL (secure sockets layer), an SSL certificate must be present for each fully qualified domain name object referenced on the page. For a hosted service to properly serve the tracking image for these SSL pages, you must acquire an appropriate SSL certificate and provide it to them. However, the potential security implications of this approach require serious consideration.

Now that we've maximized our collection through the use of first-party cookies, we need to be sure that we have a place to store that visitor information so that it can be put to good use. Ensure that your vendor employs a visitor history database that will track key events like visits, campaign responses, conversions or purchases, segment IDs, and, of course, a unique customer identifier. Ideally, you'll be able to access the data in this database directly, as you may simply want to mine in an ad hoc wayfor example, you may think, "Give me a list of all of the customers who have spent more than $500 with me, in key segments, who haven't visited the site in more than three months."

6.10.3. Using the Information

Combining new versus repeat buyer segmentation with behavioral information on your web site can yield valuable information that you can put to very tactical use. For example, generate a report that shows a product drill-down by new versus repeat buyers that will tell you your most popular, highest-revenue-producing products for each segment (Figure 6-9).

Figure 6-9. Product categories and sales metrics


Browser-overlay tools [Hack #62] can also be used to evaluate link and conversion success for the different segments (Figure 6-10).

Assuming that you are setting your own cookie as discussed earlier, ensure that the cookie on the order confirmation page is set to count the total number of purchases. Then, using your analytics solution to understand the most popular, most profitable product preferences of each group, use that information to display different promotions on your home page for each of the different segments. By serving more relevant content to different segments, you can expect to increase conversions, average order size, and total revenue.

Jeff Seacrist and Eric T. Peterson

Figure 6-10. Browser overlay for the "New Buyer" segment




    Web Site Measurement Hacks
    Web Site Measurement Hacks: Tips & Tools to Help Optimize Your Online Business
    ISBN: 0596009887
    EAN: 2147483647
    Year: 2005
    Pages: 157

    flylib.com © 2008-2017.
    If you may any questions please contact us: flylib@qtcs.net