Wise online merchants are always looking to improve sales, boost conversions, and better understand their customers. Fortunately, there's a single way to achieve all three, through the science of tracking your site's visitor behavior, known as Web analytics.
While there are exceptions, the general rule has been borne out time and time again: Those sites that invest time and money in Web analytics take a greater share of their market; those sites that neglect it tend to fall behind.
Web analytics is an umbrella term that refers to the tracking of many different indicators: total site usage, which products are selling (and which aren't), trends in usage, type of visitors (age, income, location), where visitors come from, and other factors. Web analytics data guides site managers as they improve their site, and their bottom line.
Running a Web site without such data is like steering a ship without reliable indicators of wind level and sun position: there's no way to find your direction.
Inherent in Web analytics are two key concepts: First, that a site manager examines her data on a regular basis, perhaps weekly or even daily. Second, that a site manager changes her site based on that data, then continues to track that data over time to watch the effect of those decisions. If the new data indicates the change was profitable, stick with it; if not, try a different direction.
The central idea is to trust the data more than your preconceived notions of shopper opinion, or what other sites are doing, or the ego of the CEO. In short, the numbers don't lie. And only by truly following the numbers can a site's revenue be maximized.
To go inside the ever-evolving world of Web analytics, we talked with Jupiter Research analyst Eric Peterson, whose recently published book, "Web Analytics Demystified," is recognized by industry experts as a leading guide to the topic. (Jupiter Research is a unit of Jupitermedia, the publisher of this Web site.)
Getting Started: Tools
The first piece of advice Peterson gives to e-businesses about Web analytics is: tracking and responding to your data is critical to your revenue. "A lot of companies that should have this stuff dialed actually don't," he says. "If you don't have at least a nominal analytics package in place, you're missing a piece of the puzzle." About 40 percent of e-businesses use analytics on a consistent basis to improve their sites, Peterson says.
Even those companies who think they've addressed the issue may not be maximizing the potential of analytics. Some successful e-businesses say, "We have some software in place, and we're probably getting the right reports." But, Peterson says that online merchants can slip behind competitors when they fall into the trap of thinking, "'we're probably getting the right reports."
An investment in analytics is highly beneficial even for companies who already do well. "Maybe a site sells a lot of a given product. But are they selling a lot of it because it's looked at a lot, or because it's looked it infrequently but every time someone looks at it they buy it?" In other words, what is that product's browse-to-buy ratio? And if it's favorable -- according to your data -- it should be moved to a more visible location in your site.
To get started -- or to move to the next level -- a site needs high quality analytics software. There are at least 14 vendors that offer enterprise quality packages, and perhaps an additional 30 to 40 vendors who offer inexpensive packages for smaller sites.
Peterson points to Urchin as a good package for the small to medium size online merchant.
While its functionality is not top tier, it offers a good value for a lower price. "They really have a good ability to track search engine keywords," he says. He also likes WebTrends Small Business edition and ClickTracks as economical packages. "They've done some interesting things with their interfaces."
The entry price point for an economy-class analytics package is under $1,000, but important add-on features can quickly push them into the $3-5,000 range. To small merchants who balk at spending $5,000 for software, Peterson stresses the potential. "What if you're able to significantly increase your profits or cut down on expensive calls to your call center? -- and that kind of stuff happens all the time."
The other option for a small merchant is to pay for a hosted solution, for which they pay a recurring monthly fee. For example, WebSideStory's HitBox Professional, which starts at about $30 a month. This can be a good choice, Peterson notes. In his experience, "It was pretty common for a business to get such a program, and have that epiphany -- 'gosh, I never knew what people were doing.'" But these hosted solutions may or may not have the robust capability of full-featured packages. And Peterson notes that long term it still may be a better deal to purchase a program for a higher upfront cost but no monthly fee.
There are, of course, higher end full-featured packages, which can cost five, six or even seven figures for deluxe solutions. Some of the top vendors on the high end are Omniture, Coremetrics and Visual Sciences.
Choosing a Package
When shopping for an analytics package, a merchant must look to see that the software has certain key functionality. Perhaps the most important of these is the company's tech support. If you have a question -- and you will -- does a vendor have dedicated account support? "You'd be surprised how important your relationship with your vendor actually is," Peterson says. "Getting a reasonable answer in a reasonable amount of time is one of the key contributors to success in using analytics."
In addition to account support, a good package will provide in-depth information about the following:
Commerce metrics: These are the basics of Web analytics: which products are selling, how much revenue is generated, and total units shipped.
Visitor Segmentation: Merchants understand that some of their shoppers are more valuable than others. A good analytics package will break down your visitors into all the major categories. For example, returning shoppers are much more valuable than first time shoppers. "So you should have visitor segmentation tools to let you understand activity of previous visitors and purchasers," Peterson says. Other good segments to track include: people responding to an e-mail offer, shoppers coming from a pay-for-click campaign, and users drawn from advertising partners like Google or Yahoo!. For example, "You want to be able to know more about people who came from Google who responded to a particularly expensive keyword -- helping you further quantify your return on investment."
Attraction to Content: This data tool tells you what products and pages are being looked at. "How often are people looking at your shipping and return policy -- and is that an issue?" Peterson says. This data determines the all-important "browse-to-buy" ratio of any product. If a product is viewed often, many businesses chose to move it to the top of the page. "Responding to browse-to-buy ratios is a good strategy for getting a quick lift in product sales."
Process measurement: These tools answer questions like: can users get through your checkout process easily? Where is the highest abandonment in your checkout process? This data can then be examined on a weekly basis to see if your improvements are working. "All of the top tier vendors have their own solution for process measurement, called 'conversion funnels' or 'fall-out reporting,'" Peterson says. "It's a very hot topic in analytics today: how to help people visualize through-put and abandonment in multi-step processing. Because if the data can be used to reduce abandonment at the start of your check-out, you can significantly increase the number of orders you get at the bottom of that funnel."
Continued on Page Two: Using Your Program.
IE 7
Firefox 2.0
