SEM by the Numbers

Target Marketing
Hallie Mummert, Editor in Chief
April 2006

What would you do if you knew for a fact that 50 percent of all prospects who spend 10 minutes or more learning about your product line and view at least one of your competitive comparison charts go on to buy one of your products? You would focus your marketing efforts to drive this kind of behavior, or at least make it easier for prospects to find this information, right? Well not only does Web analytics make this kind of insight possible, but it also helps you determine if such sales are profitable in the short term and the long term.

By analyzing search engine data that show what happened before people got to your site along with Web site data that tell you what these people then did on your site, you can better manage your keyword bidding, site optimization efforts, affiliate programs ... basically your overall SEM investment. What's even better is that analytics is not an all or nothing game. While Lisa Wehr, founder and CEO of Lake Leelanau, Mich.-based search engine marketing firm Oneupweb, says it's "shocking how many marketers still don't even track their SEM and online activity," she notes that even the most basic analysis can produce real results.

What Are You Tracking?
Just as you could in the offline world, you can gather data on just about any permutation of online activity. But the online space provides even more options for slicing and dicing. How sophisticated you get depends on what you need to know to make an impact on your SEM investment and the resources you have to track, analyze and act on this information.

According to Josh Stylman, managing partner of SEM services firm Reprise Media in New York, marketers should focus on those metrics tied to their program goals, such as sales, leads and ad impressions, and the costs associated with driving these actions. To do this, you will need to look at the "building blocks" to the metric. In most cases, this involves determining what percentage of people might be exposed to your search engine link; what percentage of this audience is likely to click on your link; what percentage of this group then is likely to progress past your homepage; etc. This exercise gives you a starting point against which to measure your campaign performance and test options to improve your return.

For online retailers, this process is fairly rigorous, explains Wehr. They must keep a tight rein on cost per acquisition (CPA) and ROI for their SEM programs. When it comes to B-to-C organizations that aren't selling products online-say a publishing firm with an ad-supported site-it's a little less critical to calculate these metrics to the penny. Instead, these firms tend to focus on branding, conversion rates and number of leads generated.

The measurement process also is a little different for B-to-B firms, says Wehr, as the longer sales cycles in this sector makes it difficult to track how the online component impacts future sales. In this case, she advises B-to-B firms to track what they can, particularly cost per lead, how many leads convert to sales, the clickpath for these online visitors, etc. For example, the clickpath can tell you which visitors took a needsassessment survey, qualifying behavior that might lead you to offer visitors different whitepapers based on their survey results.

At its most basic, SEM analysis boils down to measuring clickthrough and conversion rates, says Jeffrey K. Rohrs, president of Optiem, an interactive marketing agency in Cleveland, Ohio. Where it gets sticky is determining the value of behavior other than sales. For example, he asks, how do you value a catalog request, which can be a costly type of activity?

Some behavior you won't be tracking for leverage, but rather to eradicate it. "[Marketers] need to be monitoring click fraud and making sure traffic coming into the homepage is progressing past the homepage," says Wehr.

The When Factor
"The problem with pay-per-click (PPC) advertising," Wehr points out, "is you can be chugging right along and meeting your goals ... and then someone new can ride in ... throwing your numbers way out of line. That's why it's so critical to manage PPC campaigns with a magnifying glass. If companies are running PPC campaigns on auto-pilot, they're likely not getting as much as they can out of the return."

Rohrs agrees, adding that the spoiler can be a competitor that's optimized its search program to the hilt or that simply doesn't know any better about how to bid effectively. Either way, your results go out the window.

To effectively manage PPC programs, online retailers need to look at keyword performance, cost-per-click and costper-conversion numbers daily. Wehr stresses, "You have the potential to bum through a lot of cash quick [with PPC], so you need to be looking at [metrics] regularly."

On the search engine optimization side, program measurement is important to keeping up with adjustments that can push your search rankings higher. But since you can control your costs for this type of work more easily, it doesn't need the daily monitoring that paid search campaigns demand.

Making Sense of the Numbers
The point of measuring online activity from the search engine to your site is to determine where you could tweak your marketing efforts to drive more profitable online interactions.

The aspects of a search program that see the most refinement and testing are keyword terms and landing pages. The goal is to connect the search user with the landing page that best fulfills her information needs. If 20 percent of search traffic is bypassing your homepage and going to a particular product landing page, then it would suggest that this group used a very targeted keyword query. But, perhaps, not much of this 20 percent is converting to a sale, says Wehr. The analytics will help you determine on which pages visitors dropped off, so you can create hypotheses as to why and then set up tests to address these holes.

"You should be testing offer, product, design, position of the buttons ... all those things that can impact whether or not the conversion rate goes up. As you leam from that testing and figure out which landing pages perform better, you can further refine your [keyword] ad groups and start slicing and dicing those into smaller groups. Then you can design a more targeted landing page around those groups. So, eventually you're going to have more landing pages and more keyword categories, but they're going to be a lot more targeted," Wehr explains.

Keep in mind that as you refine your keyword groups, you will have more tail terms-e.g., "Tiffany Eisa Peretti bean necklace"- versus head terms-e.g., "necklaces"-says Stylman. Tail terms produce less data for analysis, and so your sample size will be insignificant. Refinement of a search program requires the creation of taxonomy for search terms that can be rolled up into slightly broader groups for more conclusive analysis.

Missed Opportunities
Besides not capturing online activity related to search, online marketers are missing a few other opportunities to leverage their SEM investments.

One of the biggest search problems Wehr sees is retailers not keeping up with inventory. "They are purchasing PPC ads for products they no longer have in stock or have discontinued. If you're managing hundreds of thousands of keywords and URLs, that's a hard task to keep up with. But for every click that ad's getting, it's dropping your ROI number," she says. For those companies that have SEM firms managing their programs, she says, they need to keep the firm in the loop on discontinued or out-of-stock inventory.

On the natural search side, Rohr finds that less mature online marketers contine to use legacy online systems that are not optimized for search engine spiders. So, they're missing out on a great deal of low-hanging fruit when it comes to driving search traffic.

Stylman would like to see marketers do a better job of determining lifetime value (LTV) for their online campaigns. There are so many ingredients in the search recipe, he notes, that it's hard to get all relevant factors in the LTV model for online marketing.

Wehr agrees: "Most [marketers] focus on CPA or cost per sale. They want to spend 20 cents to make $2, but they're not accounting for the relationship they've now built with the customer and how it impacts future orders. That number is important, because it allows them to stay competitive and increase their spending if it makes sense. Otherwise, they start sliding down the slippery slope ... they lose market share to a competitor or they become frustrated because it looks like their competitor is bidding more aggressively, so they may invest money in different channels because they didn't understand search and LTV in the first place."

Given that cross-channel shopping activity continues to rise, getting a handle on LTV will only get more complicated. A study conducted earlier this year by JupiterResearch for SEM firm iProspect found that 62 percent of shoppers who went online to research purchases this past holiday season used search engines to find what they wanted; 47 percent of this group of consumers then bought these online-researched products offline.

"The sweet spot is determining how much online activity drives offline sales," says Rohrs, who advises multichannel marketers to create online mechanisms to help them track this behavior-such as coupons good for the retail environment. And, he adds, be sure to put other types of conversion options on your homepage and other key landing pages linked to your search campaigns to allow visitors to convert in their channel of choice.

Copyright North American Publishing Company Apr 2006