5.1. Hacks 6880: IntroductionThere was a time when most web data analysis was done simply to understand the technographics of web site visitors: which browsers and operating systems they used, how much bandwidth they were consuming, which loadbalanced servers pages were being served fromreally pretty boring stuff. Fortunately the fields of web measurement and web data analysis have progressed sufficiently that we're now able to ask truly interesting questions of the data in an effort to continuously refine the user experience online. Still, one should not forget one's roots, and so, inevitably, there are still a handful of good reasons to turn to technographic data from time to time. Demographic data, on the other hand, is cutting edgethe idea that you can know not only what your visitors are looking at but who they are, where they live, etc. Many of the recent advances in web measurement have revolved around attempts to tie web and CRM (customer relationship management) data together to create a more "holistic" view of Internet visitors. In fact, there is an often abused notion that the combination of web and CRM data will help companies create a "360-degree view" of their customers, detailing everything about how they interact with your organization, enabling you to better serve and sell to them. This notion has yet to be convincingly proven. Some companies are able to integrate data [Hack #32] and create a "180-degree view" of the customer and others simply say they dobut when the rubber hits the road, the data is usually almost meaningless. In this chapter, we'll deal with practical uses for technographic and demographic data, leaving the fantasy of a complete view of the customer to analysts and other dreamers. Technographic data is often referred to as the "ugly details" underlying a visitor's visit to your web site. This includes data points such as the type of browser visitors use, the operating system they're browsing from, the level of JavaScript they support, their acceptance or denial of cookies, and so on. Some of this data is actually pretty useful, and some clearly is not [Hack #74]. The trick is figuring out what data is useful to you and how to put it to work. The best uses for technographic data almost always revolve around your web development and quality assurance programs. Regardless of whether you're responsible for a simple web site or a complex web-based application, you have roughly the same commitment to qualitya page error is a page error, and no web visitor looks kindly on page errors. The essential elements of using technographic data are:
While any good web QA manager will likely scoff and say, "duh, of course!" the reality is that most companies completely botch step 3, either drawing the line too low, too high, or not at all. You don't want to test everything, and you don't want to test nothing; the challenge is to determine your tolerance for visitor complaints and work backward from that. If every visit that fails because of a JavaScript error costs $30, then you should be more aggressive than when the same error costs only $0.03. 5.1.1. Why the "Quotes" on Demographics?I say "demographics" in the context of web site measurement because Wikipedia has this to say about demographics:
This is a fairly pure definition of demographics, covering the standard characteristics that marketers are interested in when they're trying to target an audience. It is not unheard of that companies will purchase advertising opportunities in an effort to target, for example, "18- to 25-year-old men with annual incomes over $32,000 and at least a bachelor's degree who are either homeowners or in the process of buying their first home"which is pretty specific targeting when you think about it. Because the level of targeting for web visitors is rarely as granular as the most basic offline targeting efforts, I prefer to refer to "demographics" in an effort to convey that what we're trying to do is similar but not same. Demographics in the web measurement world will usually be confined to relatively simple and binary demographic segmentsfor example, male versus female, geography, or age groups. The one exception is designated marketing area (DMA) [Hack #78], which is a traditional demographic element that is making its way into some measurement applications. Because the data is difficult at best to get, and even more difficult to meaningfully integrate, our treatment in this chapter will remain both light and practical; for a more complete treatment of how demographic data can meaningfully be combined with your web data, I recommend contacting your measurement vendor. |