RAMP Up Your Web Success
Online marketing's dirty little secret? Analytics is hard.
If you’ve worked in a marketing group for very long, you’re sure to have come across Web analytics reports of some kind produced by vendors like WebTrends, Omniture or Google Analytics. It doesn’t really matter where they come from or how they’re produced, the general gist is to somehow identify which of your online marketing efforts are working and which need help. But if you’re like most people, you might find these reports obtuse, full of archaic language and more confusing than enlightening.
Trust me here, you’re not alone.
The dirty little secret in online marketing — the one thing vendors don’t want you to know — is that Web analytics is hard. Despite assurances to the contrary by companies large and small, there is little that is actually “easy” or “intuitive” about transforming the data collected via Web analytics systems into practical and actionable advice. And no amount of repeating, “Web analytics is easy” and “Our application is easy” like a monk chanting a mantra will make Web analytics any less hard.
So what is the average marketer supposed to do?
It’s not like you simply can give up and refuse to measure, especially given the purported “infinite measurability” of the Internet. Senior management expects you to be making solid, data-driven business decisions about where to spend those hard-earned marketing dollars, and I suspect that pleas to be let off the hook because “visitors delete their cookies,” “these systems are complicated” or “there are no practically applied standard definitions” will fall on deaf ears.
But I have a solution for you. After nearly 10 years in the industry and working directly with some of the largest and most well-known brands in the world to make Web analytics “work,” I have a foolproof way to sell and execute Web measurement projects. All you need to do is RAMP up your use of Web analytics.
In this case, RAMP stands for:
● Resources: the technology and people you’re going to need;
● Analysis: the recommendations and insights your people produce;
● Multivariate Testing: the way to get the
rubber to meet the road; and
● Process: the road and road map to get you where you need to be.
The most common mistake companies make regarding Web analytics is to believe salespeople when they say, “Our technology is so easy to use everyone will get it!” It’s hard to fault them for wanting to make the sale, but … this more or less is an outright lie — unless, of course, everyone understands the difference between a page view and a visit, and the impact that AJAX has on both; a daily unique visitor and a lifetime returning visitor and how each would be used in the context of a multichannel campaign; and a session referrer and original (lifetime) referrer and when to reach them. Then everyone will be just fine.
But let’s assume the terms in my last sentence were Greek to you. In that case, you might want some help interpreting Web analytics data. In fact, some people (like really senior people) probably will want more than just the data; they’ll want analysis and recommendations.
Unfortunately, there is no Web measurement application available today that produces analysis or generates recommendations. These systems gather and report data and information — you need people to make them useful. Those reports you get are designed to be an input, not the final output, in the Web analytics process.
Your bright people need to be tasked with generating analysis and making actionable recommendations. This is the difference between giving someone a spreadsheet and giving him a spreadsheet with a cover letter that says, “Based on the information I’m providing, I recommend that we X and Y and Z.” Sure, making recommendations requires putting your neck on the line, but when your resources
have helped you gather good data, the risk is mitigated.
When you have your analysis and recommendations in hand, you have to put them to work. While you can start simply making wholesale changes, you’re far better off using some kind of controlled experimentation platform like Google’s Website Optimizer, Offermatica or Amadesa to allow you to randomly test changes in a statistically relevant way. This way, when you’re asked how good those recommendations were, you can confidently respond by saying, “I’m glad you asked! It turns out that one of the 16 combinations we tested improved our conversion rate by more than 30 percent with a confidence interval of 95 percent.”
I know, “Ewww, yuck … process!” But, really, process isn’t a four-letter word in Web analytics. Web analytics is hard, and you need to do everything you can to increase your likelihood of success. If you revisit how you use Web analytics in the context of your marketing efforts and initiatives, you might discover opportunities to gather better data as an input to your analysis and multivariate testing efforts.
Focusing on the process of “doing” Web analytics allows you to increase organizational awareness of measurement, reduce the likelihood of mistakes and begin to treat Web analytics like any other strategic business initiative. As Tom Davenport points out in “Competing on Analytics” (Harvard Business School Press, 2007), many of the best companies in the world treat measurement as a strategic priority. A process-oriented approach to Web analytics will let you “compete on Web analytics” and gain an edge over your competition in the complicated realm of online marketing.
Yes, Web analytics is hard. But you don’t want to make it any harder by continuing to believe the lies and pretending that all those reports make perfect sense. Start to invest in your own RAMP today. You’ll be glad you did.
Eric Peterson can be reached through his Web site, www.webanalyticsdemystified.com