BACK TO BASICS: What is Web Analytics?
You already measure the results of your PPC campaigns and you keep a close watch on how Google and the rest rank your pages. Then, along comes something called "Web analytics" and people figure you know all about it. You probably do - you just don't know how to put it into words. Here are a few that will help you explain it to others.
First of all, there are two types of Web numbers so allow me to set the record straight and make life just a little easier for those of us trying to explain ourselves to colleagues, clients, friends, family, neighbors, pets, ceramics classmates, and carpool members.
Web metrics is the measurement of what's happening on the Internet itself. Had I only known that way back in 2002 when I wrote a book called "Web Metrics" (John Wiley & Sons). Fortunately, I was allowed to add a sub-title: "Proven Methods for Measuring Web Site Success". That, in a nutshell, is a wonderful definition for Web analytics.
Web metrics focuses on the number and types of people online, the number of broadband vs. dial-up connections, advertisers, advertisements (shapes, sizes, level of annoyance) and all things related to the Internet as a whole. How many Web sites are there? How many searches? How many emails? How many are spam? Does it make sense to promote items online for sale to certain countries or to seniors? How many people search at Google vs. Yahoo vs. MSN?
Web metrics are studied by companies with large panels of surfers whom they follow around online. These firms report which Web sites are the most popular and can ask their panel members to run comparative tests of your competitors' sites. Companies like Nielsen/NetRatings and ComScore lead the pack here.
Other Web metrics firms track the hits and clicks at Internet nodes: the ISP's. They watch how the great unwashed and unidentified masses surf from site to site, reporting how people research cars, get dates and track news stories. Look to Hitwise to dominate that niche.
A third sector measures the responsiveness of popular Web sites. Can the entertainment sites stand up to the demand for information on Oscar night? Which presidential site is the fastest and which gets bogged down by demand? Keynote Systems and TeaLeaf are serious players here.
A final group tracks online commerce. How much is being spent on advertising? What percent of consumer spending is happening on the Internet? Is B2B growing faster than B2C? Look to eMarketer as an 800 lb. gorilla here.
These are the ones who study of the Internet as a whole - the Internet from the outside.
On the other side of the thesaurus, Web analytics concerns itself with what goes on inside a particular Web site.
How successful are you at attracting valuable visitors? We're looking beyond how many people clicked on your search engine listing or keyword and asking how many high-value visitors showed up?
Once an individual becomes a visitor, what do they do? Where do they go? How easily can they accomplish what they want and what you want them to? Companies need some way to measure whether their Web site design and development efforts are worth the trouble.
Do your visitors drill down to the product information? Do they put things in their shopping carts? Are they less costly customers because they use the online customer care tools and services?
Web analytics offerings range in price from free to very not free. On the free side, the most well known would be Google Analytics. How good can it be if it's free? Very.
Google Analytics provides everything you're going to need in a tool in your role as a search professional. Google is putting everything they can think of in this tool in order to show you just how important it is for you to keep buying more keywords. Brilliant move. So why are there all those other tools out there doing so well if they're up against free? Because there's more to running a Web site than attracting people through the search engines.
Companies like WebTrends, Visual Sciences, and Omniture capture an enormous amount of information about what people are doing on your Web site and then let you dig in, while Google provides lots of pre-formatted, albeit very flexible reports. Google won't report on downloads of files like PDFs, jpgs or Flash. You want to know about server error messages? You have to look to the pay-to-play vendors.
The more sophisticated the tool, the more sophisticated the analysis. What does a sophisticated Web analysis look like? Here's an example:
Let's say you want to know how those who bought from you found your site. That's easy. But what if you want to compare those who bought over a period of several weeks cross-tabbed against the path they took through your site and the time of day they showed up? And then you wanted to see how many people came from the same source (banner ad, keyword, press release) but dropped out of the persuasion path? Here's where sophisticated Web analysis comes into its own. Each answer begs another question.
If you have a high-end analysis tool, you can perform multidimensional, ad hoc queries and then schedule the results to be periodically emailed to your teammates. Things get very rarified when you integrate Web analytics data with your CRM data, your accounting data and the rest of your business intelligence.
If you're running a smaller Web site, then knowing how many people showed up when, from where and at what they looked might be enough. But if a 2% jump in conversion rates translates to tens of thousands of dollars, Web metrics will help you find more prospective customers and understand the competition, while Web analytics will drop serious dollars right to the bottom line.
Web metrics and Web analytics; two very different areas of inquiry, both of vital interest when you are trying to make a go of it in an online world.
Jim Sterne is the Founding President of the Web Analytics Association and produces the eMetrics Marketing Optimization Summit. He's offering all SEO Newsletter readers a 10% discount for the upcoming conference in Washington D.C., October 14 - 17. Just use Promotion Code: BCLAY at www.emetrics.org.