What Should I Be Measuring?

One of the first questions people ask me upon learning my profession is: “Can you tell me what I should be measuring on my Web site?” I’m not referring to client meetings — I’m talking about social events, chance hallway meetings, and other random encounters.

It makes me wonder how doctors, attorneys, and even priests must feel when they divulge their professions only to see their new acquaintances’ eyes light up and their mouths open to say: “I have a quick question for you…” At least I know I can give a quick answer.

My standard answer is this: “There is no standard answer.” Well, actually I tell them that, though certainly a few key measurements are unique to the Web, what’s measured will ultimately vary by functional area and depend primarily on the company’s goals.

The question of what to measure on your Web site is not a quick question. Dare I refer readers, yet again, to the Yield Model series from last year? A yield model is a timeless method for determining what to measure for your entire business, and it’s a good place to start in answering the aforementioned question.

But let’s stop for a minute and think about why people might be so befuddled about what they should measure on their Web sites.

One reason is lack of familiarity. For someone getting involved with the online arm of a business for the first time, the Web can seem like a completely separate business entity. In reality, for many companies it’s just a new sales channel. If the company were starting a telemarketing effort for the first time, that, too, would come with new terminology and measuring sticks.

Abundance of data is also a factor — and a legitimate reason for treating the Web as if it requires a different set of measurements. Information about customers, marketing campaigns, and internal operations is measurable in ways only dreamed of in the offline world. So it’s easy to see why someone new to the online world might hope there’s a standard set of data she should look for.

Don’t forget the new terminology. New and changing terminology and buzzwords can stump even the savviest online veteran, but even the most basic terminology is confusing for the beginner. I can’t tell you how many times I’ve been asked, “What exactly is a ‘hit,’ and what does it mean to me?” So perhaps someone who is trying to figure out what he should measure is, in reality, simply trying to figure out the names of the different statistics he’s looking for.

There are certainly some standards where online measurement is concerned. Banner ads get clicked, so marketers measure CTR. E-mails too. Sure, there are guidelines to fall back on. But even banners and emails have exceptions. In a few cases banners are purchased simply for exposure, so the number of impressions or unique eyeballs might be more important than click-through. Some email is sent for informational purposes, and click-through is not nearly as important. So, just like in the offline world, it comes back to what is important to your business. To know that, you have to know your goals and your strategy for reaching them.

Out of curiosity, I asked Rob Sherrell, general manager of interactive marketing for Delta Airlines, how he would answer the question, “What numbers should I be measuring on my Web site?” His answer was similar to mine:

It depends on your goals and objectives, and the strategies to achieve those — and most businesses have multiples of each. It’s tough to pin down any universal truth, but at the end of the day it usually comes down to dollar bills.

Jeff Roberts, senior marketing manager for CareerBuilder.com, answered:

Even within a single company, individual departments need different measuring sticks, even though they’re working toward the same goal. Here’s an example: On a daily basis I focus on measurements that identify whether or not site users are pinpointing the information they need as quickly and efficiently as possible. But overall, what I’m really doing is helping shorten the sales cycle and improve customer satisfaction… which leads to yet another set of measurements.

So how do we tackle this issue of what to measure? How is one to get started in weeding through the myriad statistics and data sets, to hone in on what’s most important? In the coming weeks we’ll narrow our focus to a few case studies in key industries and see if we can shed some light on the subject.

Related reading

site search hp