So much has been said of web analytics “best practices” that one begins to imagine these success formulae are well known and implemented across the spectrum of analytics practitioners.
However, based on facts observed on the ground, in real customer situations, it is less likely that “best practices” are well understood than that they are completely missing in even some of the most advanced enterprises undertaking analytics today. And as in most cases where wisdom and experience are underutilized, the result is both time-consuming and costly to the enterprise.
As analytics vendors continue to market ever more complex and powerful versions of their products, one gets the sense of a candidate racing ahead of his/her electorate—in other words, losing touch with the basic needs of the customer. Can we track mobile usage? Yes. Can we perform multivariate testing? Yes! Can we integrate analytics data with third party systems? Yes we can!
But have we planned carefully enough to take advantage of even some of the most basic, most hoary features of the analytics package to which we devoted all that marketing spend?
In so many cases, the answer is a sad, not to say deflating “No.” Internal analytics expertise at large organizations remains relatively low and in some cases almost nonexistent. No amount of new tool features will remedy this. But I am pretty sure I know what can help the beleaguered analytics team more than any new technical feature, and here it is:
Planning.
In many companies, this is rightly referred to as “process”. It is applied to nearly every part of the organization’s endeavor—from washroom supplies to investor reports, everything goes through an established process designed to wring efficiency from the potential chaos that threatens the operation of every business today.
Is there a web analytics “process” that can be learned and implemented?
Yes, and we don’t have space in this article to detail that in its entirety. What we do have is some space to talk about the simple act of planning to do something before you do it, and how that alone will cut probably half of your analytics headache without making any other changes.
In our case, planning means:
DO find out from business stakeholders what they need to know in order to make business decisions (as opposed to asking “what reports do you want”—a question to which they probably do not have an answer).
DO NOT look at “what the tool can deliver” and generate hopefully relevant reports based on that.
DO locate enough expertise—inside or outside the company—to turn business needs into what I will call a “reporting design”. Granted, this requires thorough knowledge of what a tool can do, cannot do, and what types of information are “worth” getting based on the complexity of “getting”. But this skill set cannot be overlooked except at the peril of the entire analytics effort.
DO NOT settle for “basic reports” as defined by the tool vendor, as it will have little true relevance for anyone needing to make real decisions about marketing spend.
DO take your “reporting design” and develop a tagging strategy that will enable your tool to parse specific data that feeds the reports in your design. Again this may require a thorough understanding of what tags can do and where they should go; plus knowledge of the necessary javascript, html and architectural issues that will help you build the right tags for the right place.
DO NOT download basic tags and put them on every page, expecting this to generate meaningful reports.
DO force yourself and your team to take the time necessary to follow a results-oriented process.
DO NOT try to launch analytics against a campaign or new feature retroactive to the planning and launch of the trackable item or items themselves.
Marketers tend to be top-line oriented: get the campaign out the door today and get ready for the next. But with the bottom line always lurking, and with analytics providing a way to control cost and adjust spend as needed, it is more dangerous than not to attempt a serious approach to analytics without allowing either time or budget to the one thing that will give it the best chance at success: that’s right—the word is “planning”.
There are probably a dozen more do’s and don’ts that I have not had time to include here—but I think we can see a pattern. Planning plus expertise will win every time against rushing blind into an unknowable result.
