The Web Numbers Game

Confused about what you’re supposed to be measuring on your e-commerce site? Probably. Unless, that is, you’ve thrown up your hands and decided not to measure anything at all.

Don’t laugh. A jaw-dropping 27% of 300 e-commerce site operators surveyed this summer by Portland, OR-based analytics firm WebTrends said they do not measure their demand-generation activities at all.

But take heart: For the most part, the metrics that matter for your e-commerce site are the same as or similar to metrics that were around long before “click to buy” became a familiar phrase. There are, though, a few twists of which you should be aware.

“Multichannel retailers who have grown up with direct marketing discipline are the retailers who get it the best,” says Jason Palmer, vice president of marketing for WebTrends. “They are the ones who are used to measuring their businesses in a way that is complementary to online marketing.”

For example, Palmer says, many of his more advanced clients are tracking the average order sizes of first-time buyers vs. those of repeat buyers: classic database marketing.

Metrics that matter online aren’t all DM 101, however. One useful metric you can use to help boost online sales is how many visits on average it takes before the visitor becomes a first-time buyer.

“For a nonconsidered purchase, if they’ve come back twice in a 30-day period, and for a considered purchase, if they’ve come back three times in a 30-day period, then that’s a really hot segment,” Palmer says, adding that a 30-day window usually offers a large enough population to make it worthwhile to take action on their behavior.

Another tactic that can pay off online is monitoring the products that people look at or spend time on after making a purchase. Products that get a lot of attention may be ripe for an after-the-fact cross-sell. “When I plan an outbound e-mail campaign to that target market, I have a really good idea of what they’re most interested in,” says Palmer, who refers to this as an “intent” metric.

A motorcycle accessories retailer can see, say, what products and categories people who bought a certain type of helmet spend most of their time looking at. Then the merchant can craft the next offer accordingly. “You can rank them by page views or by the amount of time spent,” Palmer says.

Intent metrics “are all about getting a better understanding of the audience that didn’t convert,” he continues. “Or of the audience that did convert, what was their intent toward other products?” Focusing on intent metrics can also help avoid having to rely on pricing promotions. “Improving the communication with your customer and being more targeted is more valuable than giving them 10% off,” Palmer says.

As it does in traditional direct marketing, recency of purchase plays a large role in e-commerce marketing. “We’ve found across all the retailers we work with that 10-13 days [after the initial purchase] is the right time to send out an offer for accessories,” Palmer says.

Tracking buyers by source is another offline metric that should be measured online as well. Otherwise you can’t determine which vehicles are delivering the most-valuable customers and how to allocate your marketing dollars most effectively. But Matt Voda, senior director of product management for Digital River, an e-commerce outsourcing firm based in Eden Prairie, MN, says that fewer than 10% of online marketers do so. Those who come out of traditional direct marketing, he adds, are more likely to than others.

One reason for the reluctance among Web merchants is that tracking by source online isn’t as easy as one might initially think. “People are having difficulty because they’re using different vendors and tool sets across their efforts, and there’s a challenge around getting all the data into one consolidated place,” says Voda.

Indeed, in a recent survey by New York-based Jupiter Research, 48% of companies with more than $50 million a year in revenue cited the lack of integrated data as the greatest marketing challenge they face in their electronic marketing efforts.

Visits and visitors

Many merchants believe that tracking the number of unique visitors their Website receives in a given time period shows them how many people they are reaching and, consequently, the size of their prospective audience. But others say that measuring the number of visits is in fact more meaningful.

Under the unique visitor metric, someone visiting twice in a month is counted only once. Measuring unique visitors is the only way to take advantage of intent metrics, Palmer says. But Matt Belkin, vice president of best practices for Orem, UT-based analytics firm Omniture, contends that visits are a more accurate measure of a site’s audience than unique visitors.

On his company’s blog at, Belkin argues that each visit should be counted as a separate event because every one represents an opportunity to convert a visitor to a customer. “If I visit a retail site four times in one week, and purchase twice, what is my conversion rate?” he wrote. “If you use weekly unique visitors, my conversion rate is 200%. If you use visits, my conversion rate is 50%. Which is a better representation of site effectiveness? Clearly, the 50% [number] is much more valuable in understanding where your site may or may not be performing optimally.”

But Palmer counters that gauging conversion rates based on visits doesn’t measure the actual conversion rate of a campaign, comparing it to measuring sales against the number of times someone opens a catalog. “It’s an interesting data point, but it’s a different metric,” he says.

Adds John Squire, vice president of product strategy for San Mateo, CA-based Web analytics firm Coremetrics: “We think [Belkin’s argument] is fundamentally flawed for multichannel merchants who think in terms of unique customers, share of wallet, and how to build loyalty. You don’t build loyalty around a ‘visit.’ You build loyalty around a person.”

Conversion-rate metrics have in fact undergone several permutations online. Conversion is traditionally defined as the percentage of targets of a pitch who take a certain predefined action, such as filling out a form or making a purchase.

But Websites offer the opportunity to measure and gain useful information from parts of the process, such as “visit conversion rates” and “basket conversion rates,” says Digital River’s Voda.

Voda defines the visitor conversion rate as the percentage of visitors who convert to purchase. He defines basket conversion rate as the percentage of people who complete a purchase after placing one or more items in the site’s shopping cart.

Breaking conversion rates into parts helps to see where the purchase process may be flawed. One of Digital River’s clients, for example, determined that an unnecessary “confirm order” page that popped up after the customer clicked the checkout button but before the actual purchase page was responsible for a quarter of its shopping-cart abandonment rate. “They took out the ‘confirm order’ page and increased their conversion rate by 30%,” Voda says.

Voda calls the shopping-cart abandonment rate “the inverse of your conversion rate. If you’re measuring your conversion rate, you’re also looking at your cart abandonment rate because they go hand in hand.”

Conversion rates within various stages of the shopping process are what Ali Behnam, senior consultant for San Diego-based analytics firm WebSideStory, calls micro-conversion rates, and he considers measuring them highly effective. One example is figuring out the percentage of those who make a purchase after reaching the product page.

“If you build a report, and it says your conversion rate has tanked, what does that mean? How actionable is that information?” Behnam says. “In addition to looking at your site’s overall conversion, look at conversion at different steps of the process. Of people who come into the site, how many stick around? Of those who stick around, how many make it to the product page?”

Coremetrics offers a similar philosophy it calls the five conversion points: getting the visit; getting the visitor to browse; getting the browser to start a process, such as filling out a form; getting the visitor to complete filling out the membership form or making a purchase, and getting the customer to come back and make a repeat purchase.

Clicks and calculations

Coremetrics’ Squire also recommends tracking people’s purchasing behavior across channels when possible. For example, by tying unique 1-800 numbers to its online efforts, bed merchant Select Comfort found that an astounding 50% of the sales that could be attributed to its search-marketing program were actually coming through its contact center. The company also determined that the call center had a 38% higher conversion-to-sale rate than did the Website, Squire says. Select Comfort subsequently changed its original mission, which was to save on marketing costs by trying to close sales online.

For his part, Voda advises monitoring where on your Web pages people are clicking. The insights that clicking behavior offers will be different from merchant to merchant, he says, but in general, it’s important to understand where people are clicking the most. If visitors aren’t clicking where you want them to, then it’s time to make some changes.

“It might mean changing the layout of the page, but it also might mean rejiggering the flow of the pages,” says Voda.

And WebSideStory’s Behnam recommends measuring revenue per visit and profit per visit. “Say you’re selling a software package for $200, and you get a conversion rate of 2%,” he explains. “Now say you drop the software to $100. Your conversion jumps to 3%, but in reality you’ve lost money because you’re generating less revenue.”

Calculating revenue per visit is easy, Behnam says. “Pretty much any analytics package will give it to you. Profit per visit is a little more complicated because you have to know what products were sold and what their margins are. You can get this from a handful of packages.”


  • Number of visits to convert a first-time buyer
  • What buyers look at after a purchase
  • Recency of purchase
  • Source of buyer
  • Number of visitors
  • Visit conversion rates
  • Basket conversion rates
  • Micro-conversion rates
  • Revenue per visit
  • Profit per visit