Eric T. Peterson is the acknowledged guru of Web analytics. Not only is he a senior analyst at JupiterResearch, but he’s also author of two books, “Web Analytics Demystified” and “Web Site Measurement Hacks: Tips and Tools to Help Optimize Your Online Business,” the latter released this month by O’Reilly & Associates. Peterson spoke on the state of Web analytics with DM News executive editor Mickey Alam Khan.
Search is drawing attention away from other online marketing activity. Is it affecting Web analytics?
Not at all. If anything, the attention being paid to search marketing lately highlights the need for Web analytics – see the purchase of analytics provider Urchin by Google and Overture’s purchase of Keylime Software years ago, now manifest as a measurement application available from Yahoo Search Marketing.
The essence of search marketing is that you have to know which words are converting and which are not – measurement that is central to all Web analytics applications.
Where is Web analytics today?
Growing by leaps and bounds despite market consolidation. On the upside, Omniture recently raised $40 million, WebSideStory went public last year and their stock price has doubled, WebTrends was recently purchased from NetIQ by Francisco Partners and repurposed as a privately held company.
On the downside, some of the smaller companies have been purchased by other application providers – e-mail, retail application management. This trend of consolidation will continue into 2006, and we project that by 2010 there will be fewer than 10 successful analytics offerings serving the low, mid- and high ends of the market.
We believe that in 2005 the Web analytics marketplace sizes around $400 million, growing to nearly $1 billion by the end of the decade. Most of the growth will be attributed to the entry of businesses historically under-invested in Web analytics. But as consolidation occurs, the bigger vendors will capitalize by upselling secondary applications – bid management, search, A/B testing – into their existing customer base.
What are some of the biggest issues under discussion?
Cookies and patent infringement are the hot topics. Recently, NetRatings Corp. has sued Coremetrics, Omniture, Visual Sciences, Sane Solutions and Sagemetrics.
What about consumer cookie rejection and deletion – how worrisome is that trend?
Quite worrisome. Marketers need to come to terms with the notion that the data they’re using to look at visitor and customer retention is not nearly as accurate as they might think. Estimates of monthly cookie deletion range from 25 percent to nearly 50 percent depending on the type of cookie and the type of site – numbers high enough to severely affect retention metrics and undervalue campaigns, especially campaigns with long lives.
Is it better to use first-party cookies over third party? What’s the difference between the two?
First-party cookies are those cookies set by the Web site proper. Third-party cookies are set by external Web sites/domains. For example, if I type www.dmnews.com into my browser, any cookies that your site sets are first party. If, however, you’re using DoubleClick for advertising, any cookies they set on your behalf from the “.doubleclick.com” domain are third party. Sites are advised to use first-party cookies whenever possible to reduce the risk of cookie blocking and deletion.
Is the cookie at the heart of Web analytics?
Yes, either the cookie or something like a cookie. The problem with the Internet is that it’s stateless. By design, every time you show up at a site you look “brand new” to the site unless they are able to somehow identify you as having visited previously.
Without cookies, Web analytics is simply an examination of what “just happened.” Cookies allow Web analytics and marketers to look at “what has happened in the past and how does it impact what just happened,” which is substantially more interesting.
Have you come across any advances, analytic tools or technology worth noting?
Visual Sciences has a pretty interesting approach to data analysis for the higher end of the market and is worth a look. I think the coming year will be interesting from a measurement perspective as analytics vendors roll out ancillary offerings like bid management, search, A/B testing, etc.
Whom would you consider as among the leading Web analytics firms?
In our October 2004 report on the analytics market, Omniture and Coremetrics held market leadership positions considering a combination of overall business value and market suitability.
WebSideStory is also very strong and has aggressively grown their business since their IPO. WebTrends is very strong from a market suitability position and is working to improve their overall business value. Visual Sciences is a dark horse that has great potential.
Can marketers do Web analytics in-house?
Sure, but running software takes extra skills and a strong relationship with IT. Hosted analytics is not free from the IT requirement but typically lessens the need for IT intervention. The most important thing marketers need to know about Web analytics is that they need to dedicate resources. Simply having someone from IT or marketing “mess around” with Web analytics will not get the job done.
At JupiterResearch we recommend that any company serious about analytics hire at least one dedicated Web data analyst, ramping up numbers as the number of internal data consumers or the overall investment in applications grow.
The prevailing technology seems so similar. How do vendors differentiate themselves?
Some have more aggressive sales tactics, some brag about the customers they already have, some brag about the strength of their company, some claim to have features nobody else has … but it’s really kind of a mess. At JupiterResearch, I spend much of my time advising clients on analytics vendors, helping clarify some of the marketing fluff and highlight the real value proposition for each vendor.
Which industry categories are using Web analytics to their advantage and which are lagging?
Retail and advertising-based business models are really starting to see the value of analytics, and those companies mostly focused on lead generation and customer support are a few years behind.
Anytime you can directly attribute revenue to an online event via analytics, it makes it that much easier to stomach the necessary investment. When it’s more difficult to quantify the value the information provides, it’s more difficult to make the investment.
Is such knowledge of consumer behavior online helping retailers format better campaigns and tailor offers?
Absolutely. The combination of Web analytic data and marketing optimization – i.e., A/B testing – platforms is helping some retailers do amazing things with their marketing. Look at case studies for any of the analytics vendors and from companies like Offermatica and Optimost.
What best practices would you recommend for marketers?
First, work with an established Web analytics vendor, one who has been around and is likely to be around for a while. The experience they bring is invaluable when it comes to getting things set up and using the data to drive decisions.
Second, dedicate at least one full-time resource to analyzing the data and making sure the rest of the organization “gets it.” Too often, companies deploy analytics applications and hope they’re a silver bullet. There are no silver bullets. By assigning someone to mine and analyze the data, proactively share the results and then make and manage suggested improvements, the true value proposition of Web analytics can be realized.
Third, consider deploying a marketing optimization – A/B testing – platform side by side with your analytics platform. Being able to make changes to marketing content, buttons, images, etc., without IT intervention and then measure the results of these changes is how Web analytics truly enables marketing optimization. Vendors like Offermatica, Optimost, Kefta, SiteSpect and TouchClarity all have platforms that can be plugged into your analytics framework without much difficulty.
Fourth, have your dedicated data analyst evangelize the data throughout the organization, not just in marketing. By using key performance indicators to transform Web data into a business-relevant context, larger numbers of employees are able to digest and understand the data.
Especially consider identifying two or three top-line key performance indicators for your senior executives to orient them toward the measurement you’re doing.
How should vendors better serve clients’ needs?
Probably through analytics education, something that market leaders like WebSideStory, Omniture and WebTrends have focused on in the past year. Web analytics is hard, and nobody really learns about it in college. Education is the key.
What’s the risk if you don’t have a Web analytics program?
Wasting your hard-earned marketing dollars on campaigns that don’t provide nearly as much value as you think, at least in the context of this interview. Click-through rate is a very poor indicator of campaign success. Web analytics allows marketers to look at conversion, revenue, retention and how acquired and converted visitors behave on the site.
It’s one thing for someone to respond to a campaign and make one purchase. It’s another to have that person spend 30 minutes on your site exploring other products or services and return frequently over the next week. Analytics enables that kind of measurement.
Why the need for Web analytics?
Nobody can afford to spend money without knowing how that money is being spent. Nobody.