Tech Candy

At an estimated $292 million of an enormous software market estimated at $144 billion worldwide by Forrester Research Inc., the web analytics market in online retail circles is the mouse that roared. With the ability to diagnose customer experience problems on the site itself, quantify the success of campaigns launched into the web universe and do it in increasingly graphic formats, web analytics have grabbed marketers` attention and spawned a fiercely competitive field that now numbers an estimated 80 vendors.

All that competition has analytics vendors jockeying for position with varying product claims, pricing and service offerings. And that leaves many retailers with a challenge: how to make sense of competing--and tempting--offers.

Not much longevity

In a market that`s still young, there are few lengthy track records or 20-year relationships between customers and vendors. Though some web analytics vendors claim renewal rates of 90% and even 99%, the life expectancy of a web analytics vendor with a particular customer currently averages about 24 to 30 months, estimates Forrester Research analyst Bob Chatham, about the same customer tenure as in another hotly competitive market: credit cards.

In its relatively short experience with web analytics technology, the average company already has purchased an average 1.8 analytics packages and reviewed 2.6 vendors. For some retailers, that number is much higher--in its five years of operation, online jeweler Ice.com, for example, has used four different web analytics packages.

Whether looking to make a switch or get on board with analytics for the first time, industry consultants--and retailers who`ve already been down that road--find asking the right questions at the start of the process is key. They say knowing exactly what functionality they want, what functionality they will actually use, establishing with certainty whether it`s possible to get it from the current vendor, and being able to accurately establish from complex pricing models what the package will cost over time goes a long way toward making the difference between buying a solution that works--or just buying more aggravation.

Asking Why?

"When clients ask me for a recommendation on an analytics vendor, the first thing I ask them is why they are thinking about switching," says Jupiter Research analyst Eric Peterson. When Peterson finds the reason is that someone on the team has seen a new demo or read about a new product, that signals that the company might be overlooking capacity already available but going unused in the current analytics package. The web team may not be using the most current version of the software it already has, he adds--or it may have truly reached the upper limits of the capabilities of what it has.

A real need to graduate may constitute a reason to switch, but it`s still advisable to proceed with caution. For with some exceptions, "None of these applications is really that different from another. It`s about how you use them," Peterson says. "If you are using a vendor`s current version, and they haven`t lost your data or billed you incorrectly, and they return your phone calls and make an effort to work with you, the vendor you know is almost always better than the vendor you don`t know."

If a switch is still in the cards, the next step is to assemble an RFP. In an effort to anticipate what they may want in the future as well as the upgrades they`ll use immediately, some marketers have loaded up RFPs to what Peterson terms "ridiculous" proportions. A better solution is to come up with 15-30 measures the company uses or would use to run its business, then divide that list into must-haves and nice-to-haves. A third list holds every analytics capacity the retailer ever wanted. Prospective analytics providers see the needs list, the wants list and the dream list, and they get a chance to respond. "That`s not an extensive RFP; it`s a reasonable list of what the company actually will use," says Peterson.

Restaurant and consumer kitchen supply company AceMart.com followed a similar process when it decided to switch to Coremetrics Inc. last year. Under its previous vendor of a year and a half, key data were split between two repositories. That made it difficult to tie together for the view Ace Mart wanted. "We`d gotten into analytics with the idea of seeing what people were doing on the pages. That quickly evolved into needing to know exactly what was going on with our marketing campaigns," says Ryan Rodkey, Ace Mart`s webmaster. "But you couldn`t get a quick picture of your revenue and how it tied all the way back to visitors` interaction on the site."

Lost data

Ace Mart came up with a short list of the analytics capacities it wanted and would immediately use. Its first requirement was to track campaigns at the keyword level; its second was to look at session level data by visitors so it could find out who was visiting its site and direct its marketing efforts accordingly.

It got both, and more, by going with Coremetrics, but it also lost 18 months` worth of historical data in switching from one vendor to another. "It`s two different systems, so I guess we knew we`d lose it, but it wasn`t something we`d really thought about when choosing an analytics provider," Rodkey says. "Losing the data we`d had before wasn`t that bad for us, because we hadn`t been getting the answers we needed out of the previous tool anyway. But if we had to switch now, it would be different."

Forrester`s Chatham says the loss of data when switching analytics vendors isn`t uncommon. "I always ask companies how wedded they are to having historical results, because they may not be taking them with them," he says. "The comparability of data across vendors is zero. You are never going to get the same count out of one vendor that you will out of another. There is a whole recalibration process that has to take place."

Rodkey says having data that extends back through history is important: the more data, the better the ability to justify decisions on future direction. Even so, the gains from the new analytics package were more than a fair trade for the loss of the earlier data, he believes. "We had been spending so much on marketing campaigns that just getting that under control saved us money every month," he explains. Within five months of implementing Coremetrics, Ace Mart was saving about 25% on marketing campaigns by identifying and eliminating non-performing keywords and upping its spend on better-converting terms. At the end of five months, it had trimmed its 800-word keyword program by half.

A key driver

Functionality was a key driver in Ace Mart`s decision to switch analytics vendors, but it wasn`t the only one. Rodkey says the rest of what swung the deal was having an account analyst available at the new vendor to answer questions and make recommendations based on best practices. "You can get a ton of data in those reports and it can be hard to make a decision, I don`t care what vendor you are using," says Rodkey. "Having someone there to walk you through the process, show you how to spot a trend, and how to build reports pertinent to your business is invaluable."

IRCE 2008 Conference Multi-Media CD-ROM

Indeed, Forrester says, with many marketers wrestling with that same issue, "Education and account management are becoming a more prominent feature in vendor sales pitches." Representing only 10% to 20% of a typical analytics vendor`s revenue today, vendors may someday derive half their revenue from professional services, Forrester predicts.

Customer support is part of what keeps the three web sites of apparel maker Guess with its current analytics vendor, WebSideStory. "A lot of vendors will say, this differentiates us or that differentiates us. I have found the differentiating factor to be a great working relationship," says Mike Relich, senior vice president and CIO at Guess. Relich adds that not only is support available when he wants it, but his recommendations on product improvements are carefully reviewed by the vendor; he knows because he`s seen some of his recommendations implemented in new releases of the package.

Relich`s opinion is that most analytics vendors collect the same amount of data, so given that, his decision to switch to WebSideStory from another vendor eight months ago had more to do with two other factors that drive such decisions: how the tool presents the data, and pricing.

Pricing questions

When choosing a vendor, it`s important to consider who will be using the analytics data and how often. "People at the highest level just want a quick overview. But if you are a merchant using the data daily, the interface becomes very important," he says. In comparison to other vendors he reviewed, Relich finds WebSideStory`s interface easier to use. "I have a dashboard that I can pull up that shows every relevant statistic. It allows me to put specific goals in place and know where I am in relation to those goals," he says.

Relich`s view that data collection was roughly equal among the vendors he considered also pushed the pricing question higher in his considerations. It`s why Guess steered clear of another vendor whose interface it also liked, which Relich estimates would have cost Guess 25% to 30% more than the solution it chose. "I just didn`t think it justified the expense," he says. Today Guess spends about 10% more for analytics across the three web sites it now operates than it did with the former vendor, which supported only one site.

If retailers believe functionality is approximately equal, the appeal of the user interface is a matter of personal preference, and the importance of pricing falls somewhere in between depending on individual circumstances, what other criteria are they using in evaluating analytics vendors? With approximately 80 vendors now fighting for a piece of the market, shakeout and consolidation are likely, so it`s no surprise many are also weighing factors such as a vendor`s longevity and future prospects.

Longer-established market leaders may have the benefit of maturity that`s served to refine the collection and presentation of data beyond the offering of a smaller upstart. If they`re also better capitalized, they`re less likely to evaporate or become candidates for acquisition, which could raise questions around data integration, migration, or new technology. On the other hand, newer players may be in the marketplace precisely because they`ve developed a novel approach that differentiates them from the competition.

"It`s the $64,000 question for analysts. Do you guide clients away from technologies that are new and unproven because they may have less robust funding strategies than some of the established vendors? If you just go with the market leaders, then what about some of the innovative technology that`s emerging elsewhere?" says Peterson.

Taking a chance

Nine months ago, online jeweler Ice.com took the plunge to become an early customer of a then little-known analytics software vendor, Visual Sciences, its fourth analytics solution in five years. Compared with analytics vendors such as NetIQ Corp.`s WebTrends, Visual Sciences is a much newer player. Founded in 2000, it`s kept a lower profile than some, but what helped close the deal for Ice.com was that Visual Sciences offered two things Ice.com hadn`t found in other vendors it considered at the time. They were the ability to do A/B testing in simultaneous segments so it could push out winning promotions and messaging more quickly, and the ability to see campaign conversion and clickstream data in one visually compelling user interface.

Executive vice president of marketing and co-founder Pinny Gniwisch says Ice.com did consider Visual Science`s position as a relative newcomer and its prospects for the long haul, concerns he says were allayed by its knowledge of other retailers who were having a good experience with the company`s installed software product. "It was a question of feeling, belief and trust," he says. "The worst thing that could happen was, it wouldn`t work." To cover that contingency, Ice kept its other analytics solution in place for a time after it implemented Visual Science`s analytics.

Gniwisch says the bet has paid off in higher conversions, with the ability to do simultaneous segmented testing making a significant difference in decisions on marketing spending. "Let`s say we`re paying $25,000 for a front page on MSN. Every point of conversion is going to make a huge difference because of the number of clicks you get," he says. "On a campaign like that we have seen our conversion rate go higher based on the testing we have done."

As for Visual Sciences, it`s gathered enough converts and new customers to push its growth up to 100% last year, according to Forrester.

Features and functionality, pricing, the subjective appeal of the user interface and marketplace longevity and prospects are when retailers look at switching analytics vendors. Smart retailers will clear yet another set of questions before they even get to the point of considering alternative vendors, such as whether they can get what they want from their current package, whether they are using the most current release, how easy the implementation will be, and how difficult to hang onto historical data in any changeover. Layered over all of that is what may be the most important aspect for many retailers--the intangible of relationship.

"No piece of web analytics software is a silver bullet," says Peterson. "It`s simply a tool to let you learn more about the relationship you have with your site visitors. Web analytics isn`t easy: if it were, everybody would get into it. So assuming that`s the case, companies need to be able to rely on their vendors to help them do more with the software."

mary@verticalwebmedia.com

Click Here for the Guide to Web Analytics Products & Services