What Should You Measure? Part 1: E-Commerce Metrics
My last column raised the persistent question, “What should be measured for my Web business?” More data is at our disposal, but we have less time than ever to make sense of it all.
If you are wondering what to measure to evaluate your online business, the simple truth is it depends. The metrics that matter depend on your company's goals and strategies, your own role in the company, the time period under consideration, and more.
To illustrate the variations in which measurements are considered “critical,” we'll look at a few key industries and differences among businesses that would appear to have somewhat similar goals.
We'll start with online stores.
At first glance, this seems a narrow enough category. Revenue and profits come to mind immediately as critical to measure. So do new and repeat customers, order size, product information, and shopping cart abandon rates. But it would take all day, every day, to stay on top of every measurable action or attribute. What's most important? As I said above, it depends.
Steve Furst, president of Djangos.com, an online aggregator of used CDs and DVDs, provides an interesting example illustrating how what to measure varies by short- versus long-term goals and objectives.
“On a daily basis, we focus on the results of our individual marketing promotions… the standard set of statistics to measure reach, response rate, conversion rate, and revenue, all the way down to profit margin,” said Furst. “It's very important that we validate each promotion down to the margin level because our affiliates' profits are affected as well as our own,” he explained.
Furst also raised the longer-term, more strategic view of what to measure. Beyond the daily routine of running a business, a periodic, in-depth review of data can reveal strategy-altering revelations.
Four years ago, Djangos.com's strategy was to acquire independent music stores and ship their inventory to our Portland fulfillment center. But our data began to tell us we were doing it all wrong. We learned that we were importing inventory into Portland from say, Chicago, only to turn around and sell it to consumers back in the Chicago area. It's faster and cheaper all around to partner with stores and let them keep the inventory locally. It was a major strategy change, and I could not have sold it to the board without the geographic data.
For the tool-seekers among you, Djangos.com uses a customized version of WebTrends Reporting Center eBusiness Edition.
Jeff Antisdel is VP of marketing and business development director for FurnitureFind.com, a build-to-order furniture e-tailer with stores both online and off-. Jeff focuses on clickstream analysis and its relationship to sales.
“I spend about an hour each morning with my daily sales reports. I start with revenue, looking at strengths and weaknesses in daily sales,” said Antisdel. “Then, I work backwards to evaluate the clickstream behavior that relates to those sales: how the customer got to our site, interacted with our site, and ultimately concluded with a purchase.”
Antisdel has found studying the clickstream is the best way for him to identify what the customer cares most about, whether it's a focus on fabric, furniture style, or price range. This enables him to identify trends quickly, capitalize on emerging opportunities, and maximize marketing return on investment (ROI).
FurnitureFind uses ClickTracks Analyzer as a reporting interface for the company's own marketing information system.
For a different online store spin, I spoke with several people at KnowX. KnowX sells online access to public (courthouse-type) records, so its store sells data rather than a physical product. Jane Rafeedie, general manager, has narrowed her definition of critical stats over the years, based on actionability.
“We have an automated dashboard report that we all use as our temperature gauge,” Rafeedie explains. “It's generated almost in real time. It starts with revenue, but there's a lot of detail: visitors to the site, registrations, purchases, product information. It's segmented further by customer type, such as new customer versus repeat customer. If revenue looks normal, it's business as usual. But if we see any unexpected spikes up or down, the detail in the report helps pinpoint the cause — and the action to take.”
Chris Luttrell is charged with KnowX's customer retention efforts. She evaluates which data is most relevant on an ongoing basis.
“Today, I'm working on an e-mail campaign so I'm focusing on your typical e-mail stats: click-through rates, registration rates, and revenue. But even my e-mail data needs vary sometimes,” said Luttrell. “The clickstream to conduct an assets search is different from the clickstream to perform a physician background check search. So, I modify what I measure based on the focus of the campaign. Then next week, I'll be moving on to a customer satisfaction survey, so I'll be focused on relating satisfaction to customer demographics and purchase behavior.”
KnowX uses a home-grown reporting system.
The above scenarios are proof the value of any given Web metric changes depending on the perspective of the user. For an online store, revenue and profit may be common measuring sticks anyone in an organization can focus on.
These marketers use different data sets and methods to reach a common goal: increasing revenue. I was glad to hear a common theme: To be evaluated properly, marketing campaigns must be tracked all the way down to the revenue and profit level. Also worth noting is nearly all utilize daily data to immediately identify emerging opportunities.
In my next column we'll hear from online media folks. They surely share some common data needs such as CTR, but I suspect we'll find they, too, have a different perspective on what to measure.
Melaney Smith is a
freelance marketer with more than ten years of experience in marketing,
finance, and data mining. She evangelizes the need for businesses to mine
their data and make it easily accessible for decision makers. She
specializes in helping businesses utilize customer, transaction, and other
data to improve or create sales and marketing programs.
