Web Analytics - Who's Mining the Store?
Before you can effectively analyze customer behavior, you need to determine who at your company will be responsible for tracking the data.
At this point in the evolution of e-commerce, virtually all online businesses have some form of Web analytics package. These applications, which allow you to track your shoppers' on-site behavior, are used by mom-and-pop e-commerce operations as well as Fortune 500 companies.
The issue now, experts say, is getting the most out of this software. The difference between merely owning a package and truly benefitting from it can be the difference between an anemic bottom line and robust yearly sales.
The biggest Web analytics decision you face is, who is going to be responsible for it? What set of eyes is assigned the task of following the data, week after week, sharing it with the proper personnel?
|Web Analytics: A User's Guide, Part 1|
|A primer on the all-important science of tracking your e-commerce site's user behavior to improve your bottom line.
JupiterMedia analyst Eric Peterson, an expert in Web analytics, notes that too many companiess say, "we'll just have the Web master, or some junior marketing person, 'watch it,' and hopefully that will work." However, hoping usually is not enough. "The reality is that almost always doesn't work."
While companies that use their analytics software on a casual basis may find interesting data, mining the data for maximum advantage requires assigning a full-time staff member to it. This employee must truly "own" the data — and the responsibility for disseminating it throughout your business.
A Little Bit Tech, a Little Bit Marketing
The employee you choose should have a specific set of skills. "It's usually not an IT person or a developer — those people are downstream consumers," Peterson said. "You want to have someone who understands the analytics application and your business's needs, so they can provide insight to your Web masters, marketers, merchandisers and IT staff."
This analytics staffer must be able to straddle two worlds, technology and marketing. The individual needs basic familiarity with HTML and Web development so they can explain why or how — technically — a page needs to be changed in response to the data. On the other hand, the analyst must have a strong business or marketing background. After all, business insight is really what deciphering analytics data is all about.
In addition to a healthy blend between technology and business, the analytics expert must have "a strong demonstrated history of working with cross-functional teams," Peterson notes. In other words, they need to be fluent in the language of both IT and business people.
"Being able to straddle the fence between business and IT — which are notoriously antagonistic — really helps," he said.
Also essential, this expert must have enough authority in the organization — or report to a senior person — so that hir or her insights carry weight. Ideally, this analyst would report to the vice president of marketing or operations to discuss which departments should get which data segments. "Anecdotally, we see this as being very successful," Peterson said.
If the analyst isn't empowered, "you've got an expensive application that isn't doing anything," he said. The greatest data in the world helps nobody, if only one person knows about it. Yet even marginal data, shared widely, helps the entire enterprise, "because at least it gets people thinking critically about the data, and how it helps you."
Ideally, this employee can be hired from within the company. In fact, this may be a necessity: Web analytics is still new enough so that there are few real experts, and most veteran analysts are already employed. "These people are very difficult to find," Peterson said.
Companies may balk at hiring an analyst, he notes, however, "if they're the right people, they're smart, and you empower them, the return on investment should be very strong."
Analyzing Analytics Vendors
Online merchants can benefit greatly by understanding the world of Web analytics vendors. Not only does this understanding help e-tailers choose the right program or upgrade to a better program, it may help them get the best price on a program.
Peterson forecasts that 2005 will be a year of stability and growth for Web analytics vendors. The early years of e-commerce saw wild turbulence in the analytics market, and certainly these vendors still face volatility, but it will start to subside.
Contributing to this move toward stability is the fact that many large merchants will stop hunting for a new analytics vendor. Among companies that spend in excess of $50,000 on analytics, 52 percent will "definitely not" buy a new package in 2005, and another 32 percent say "maybe," to buying new software next year, according to JupiterMedia.
Among companies that spend between 10-50k on analytics, 35 percent will "definitely not" buy a new package, and 31 percent say "maybe." In this same segment, 19 percent say "definitely yes" to buying a new package next year.
What this means for vendors, Peterson says, is that analytics firms that sell higher-priced packages "are going to be able to step back and not fight the fires of customer defection, and instead focus on improving the overall quality of their application."
And this process of regrouping appears to be necessary. Many online merchants are far from enamored with their current analytics packages. A common complaint among e-tailers who buy "hosted" analytics packages - Web-based programs that charge a monthly fee - is service outages. That is, the program is still collecting data but the user interface is down; or, there's some odd data change; or the viewer is slow. In some cases, vendors fail to alert users before a service problem, causing great user frustration.
Some executives at online businesses complain that today's analytics packages are not providing the level of accuracy they expect. The reasons for this inaccuracy are numerous: log files are full of requests from search engine spiders, shoppers use front and back buttons, interpreting proxy caching causes problems, and other technical snafus. Some analytic packages cannot handle a sudden flood of data, like those caused by a dramatic uptick in users following a major news or product event.
Analytics vendors respond to complaints about inaccuracy, Peterson notes, by saying "it's hard to be completely accurate because it's the Internet, which is an inherently inaccurate medium." These vendors instead encourage their customers to focus on trends rather than precise data. But, Peterson notes, "you wouldn't accept that from your sales force automation system or your bank - why accept it from your analytics system?"
In fact, when executives were asked what their most important criterion was for selecting an analytics package, 69 percent said "accuracy of information," according to Jupitermedia research. "I would be very surprised to talk to someone who says 'we've been using this application for a year and we've never felt that there are inaccuracy issues," Peterson says.
As thorny as these technical issues are, he forecasts that vendors will address and solve them. "We see the vendors - they've all expressed this - looking back at these issues of accuracy throughout 2005, with hopefully some resolution by early 2006."
Getting the Best Deal
Given the fierce competition between vendors, it remains a buyer's market, Peterson notes. That's good news for e-businesses looking for a deal. "And I cannot see far enough into the future to see when that will change."
Vendors realize that the competitive marketplace means they are "challenged with having to consistently delight customers with high quality support, innovative offerings, and training," Peterson says. "It's such a competitive market, if you're willing to ignore your customer, you're willing to lose that customer."
Because of the competition between vendors, if an e-tailer shops hard, he or she could find not only the best price on a given package, but even 10 to 15 percent less. One possible technique: "play that competitive card late in the deal," Peterson notes. "If you go back late in the game and say, 'I just got a proposal from your competitor that's 10 percent lower,' the vendor will probably lower the price." However, this strategy is not without risk, he notes, because some vendors will not go this low.
The smart shopping strategy required to get the best price on an analytics package is not dissimilar to the strategies used in car shopping. "You can buy late in the quarter, late in the year," Peterson says, explaining that analytics salespeople may be pressed to make quotas then. "If you really want to get the bargain basement price, you always can."
Some Leading Vendors
Surveying the many vendors in the analytics space, Peterson points to a variety of companies that are top contenders.
In particular, Omniture, [http://www.omniture.com] is an industry heavyweight. The company "has clearly established itself as a market leader in 2004," Peterson says, noting that this year they have won AOL's business.
Other market leaders are Coremetrics [http://www.coremetrics.com], which makes a package Peterson indicates is geared for retail, financial services and travel sites; and WebSideStory, [http://www.websidestory.com] whose name is included in most all requests for proposals, and which received a substantial cash infusion after going public this fall.
He points to WebTrends as "among the founding fathers" of analytics. [http://www.netiq.com/webtrends/default.asp] Peterson, who once worked for the company, recalls a time when the phone rang constantly with businesses calling to buy the application. While that has slowed down somewhat, he says, the WebTrends 7 application offers a good refresh on its interface. Most importantly, WebTrends, "has a strong organization and has made a commitment to supporting its customers."
Generally seen as one of the more intuitive Web analytics applications, ClickTracks 5.0 [http://www.clicktracks.com/] is available as both a hosted product and a server-based one. Significant in its newest version, which shipped last spring, is the What's Changed Report. When you log onto ClickTracks, it searches for what it considers unusual shifts in the numbers of visitors, search queries, referrers, campaigns, entry pages, visited pages, and countries of origin since your last visit.
An analytics vendor that is not as well-known in the industry but that offers a quality product is Visual Sciences. [http://www.visualsciences.com]. "They basically took the mold for the analytics package, threw it away and rebuilt it from scratch," Peterson says. "And they've done an amazing job - it's a very strong data analysis engine, with an innovative reporting interface."
James Maguire is a contributor to ECommerce-Guide.com, where this column originally appeared.