Real-Time Analytics

Must-have insights or hyperbolic boondoggle?

By Eric T. Peterson
Monday, March 2, 2009

By now most marketers have realized that there are both advantages and disadvantages associated with the online channel. One of the advantages frequently cited is the availability of great volumes of data to determine the efficacy of campaigns, granular to the level of placement, creative, or content. While this great aggregate of data—the so-called “infinitely measurable” Internet—is certainly valuable, the speed at which the data can be gathered has led to an interesting but dubiously valuable marketing input: Real-time analytics.

Briefly, “real-time analytics” is the term used to describe the availability of online marketing data through systems like Omniture, Coremetrics, and WebTrends mere moments after the banner ad has been clicked or the conversion event occurs. With the right systems you can literally see campaign response numbers increase before your very eyes, and if you’ve never seen real-time campaign data you should check it out, even if it’s only to put the Internet into context relative to your longstanding direct marketing and advertising efforts.

What Is the Real Value of Real-Time?

But within the digital analytics community, we have long debated the true value of real-time analytics and, suffice to say, this is one of those things we have collectively agreed to disagree about. On one side, we have the “data now” folks who are adamant that real-time analytics is both valuable and a fundamental use of digital measurement solutions. On the other side, we have the “real fast” folks—those analysts who are content to have the data in a reasonable amount of time, even if that means a few hours delay similar to the latency found in Google’s Google Analytics application.

Personally, I fall into the “real fast” group, mostly because of my experience over the last decade working with companies of all sizes and the observation that the availability of “real-time” data has no impact on the organization’s ability to actually use that data with any true immediacy. Unfortunately, as you may well imagine, having the data is meaningless unless you’re actually willing to use it to make decisions, and this is where real-time fails. Most companies who have paid for low data latency are more likely to panic when the data flow slows, yet take days, weeks, and even months to actually take any action based on the data.

There are exceptions: For example, the online media editor who uses real-time data to determine which articles stay above the fold, which go below, and which need to be cycled out in favor of more engaging content. Or the occasional search marketer who has so perfected his or her bid management skills to be able to squeeze every last penny from the daily budget while optimizing the flow of clicks from Google or Yahoo!. But these cases are very much the exception, not the rule, and so if you happen to be in the process of selecting a new analytics application here are a few things I would consider while you navigate the otherwise dense Web analytics landscape:

Determine how quickly you’re likely to actually make changes to your campaigns, site or content. Many companies, even when armed with fairly clear data, have a tendency to make changes very slowly, often because of IT roadblocks, the need to involve numerous stakeholders, or pure, unadulterated politics. If you’re slow to make changes, you probably don’t need real-time data.

Determine how likely you are to actually make changes based on analytics. Because marketing in the online channel is still relatively new to most people, it is common for decision makers to spend more time questioning the data, source and quality than they do actually responding to the insights delivered. If you’re predisposed to debate the nuances of the data, the opportunity will likely pass.

Determine whether the added cost will provide the benefit you expect. Google is currently making “real fast” data available to everyone at the best possible price—free—whereas alternative solutions can cost even the smallest sites tens of thousands of dollars annually. With this in mind you should ask yourself whether you are better served starting with “real fast” and migrating to “real-time” as your use of this data matures.

Few debate the opportunity available to marketers in the online channel to better understand customer behavior. But this opportunity needs to be based in reality and focused not on which bells-and-whistles you can purchase, but rather which features and functions you’ll actually use and benefit from. Before you spend more than is necessary on a digital analytics solution so that you can have data now, make sure you’ll actually benefit from that data in a way that provides benefit back to your company and, more importantly, your customers.

Eric T. Peterson is the founder of Web Analytics Demystified (www.webanalyticsdemystified.com) and has worked in Web analytics for more than 10 years as a practitioner, consultant and analyst. Mr. Peterson is currently building a strategic Web analytics consulting practice to provide guidance to companies working to maximize the return from their investment in Web analytics.