|
Press Releases:
THUNDER FACTORY
Accelerates Growth
in Southern California
Fleetwood Rv, San José
State University Choose
Thunder Factory For
Integrated Marketing
Programs
Business Growth at
Thunder Factory Spurs
Headquarters Move,
Expansion into L.A.
and N.Y.
In the News:
Thunder Factory Wins
Fleetwood RV, San José
State University
Article List:
Confusing Spin
With Strategy
- Patrick Di Chiro
An Open Letter
To Al Ries, Ad Age
Columnist
Here's Where to Find
Integrated Marketing
- AdAge
Getting a Bead on 'Buzz'
- Virginia Postrel
Survey: Network TV
Does Worst Job of
Proving Advertising ROI
- Judann Pollack
Toughening Your Brand
- Lynn Upshaw
Coffee's For Closers
- Patrick Di Chiro
The Role of Key Opinion Leaders (KOLs)
in The Pharmaceutical Industry
- Joseph Gutman, MD
Playing the
Search-Engine Game
-Mylene Mangalindan, WSJ
At Last,
a Way to Measure Ads,
- Michael Totty, WSJ
Small Firms Can
Survive Sqeeze
By Revamping Marketing Efforts
- Jeff Bailey, WSJ
Study Says
Marketers Shifting
Toward Internet, Direct Mail
- Erin White, WSJ
|
|
At Last, a Way to Measure Ads
By Michael Totty
Staff Reporter of THE WALL STREET JOURNAL
Marketers have long quoted John Wanamaker, the Philadelphia retailer who famously
said, "Half my advertising is wasted. I just don't know
which half."
With the Web, online marketers say, they finally know which half.
That hasn't always been the case. Until recently, online businesses could figure
out how many visitors came to their Web site from a particular banner ad or e-mail
pitch. But they couldn't easily tell which of those visitors were spending money,
or how much they were buying.
But now a new set of tools, by tracking every click a visitor makes through a
site, can not only tally the number of visitors drawn by a particular ad but also
tell which campaigns are bringing in the most sales and the most profits.
These tools "marry data about who's buying what with traffic data -- who's
going where," says Matthew Berk, a senior analyst with Jupiter Research in
New York. "When you combine those things, it can be really powerful."
Just Looking, Thank You
The ability to measure ad performance is only one feature of "Web analytics"
software, which e-tailers increasingly are using to better understand how customers
interact with their sites. Forrester Research Inc., a Cambridge, Mass., technology
consulting firm, counts at least 27 vendors selling Web analytics tools, including
NetIQ Corp., San Jose, Calif.; SPSS Inc., Chicago; WebSideStory Inc., San Diego;
Omniture Inc., Orem, Utah; and Coremetrics Inc., Burlingame, Calif.
The tools make it easy to spot ads that bring in a lot of browsers who don't spend
a lot of money. For instance, Dartek.com, an online discount computer store, last
year launched a series of banner and pop-up ads.
The campaigns were expensive, accounting for about a third of Dartek's fourth-quarter
advertising budget, and they brought a lot of visitors to the site ( www.dartek.com
<http://www.dartek.com/> ) -- at the peak, the pop-up campaign brought in
nearly 1,900 visitors in one day.
But these visitors weren't buying enough to cover the cost of the ads. "Fortunately,
we were able to see that [the banner campaign] was not for us," says JoAnn
McNeely, Dartek's marketing projects manager in Naperville, Ill. Ms. McNeely uses
analytics services from WebSideStory to track the 30 or so advertising campaigns
she's running.
Paying by the Word
Mr. Berk, the Jupiter analyst, says retailers' growing reliance on search-engine
advertising is helping drive their use of analytics tools. Advertisers often buy
scores or hundreds of search-engine terms, and pay whenever someone clicks on
the ad. Costs can add up, so sites want to know which ones are delivering.
Restoration Hardware Inc., the Corte Madera, Calif., home-furnishings retailer,
used an analytics tool from Coremetrics to find out which ad terms would pay off
and which wouldn't. It tested about 120 keywords on Microsoft Corp.'s MSN search
portal and found that many of the most popular -- and expensive -- terms were
attracting visitors but not producing sales.
For instance, Restoration discovered that the term "mirrors" accounted
for 5% of its total "click" charges in March, but not many of those
shoppers actually bought anything. More specific terms -- "bathroom mirrors,"
"wall mirrors" and "decorative mirrors" -- cost less because
there was less demand from other advertisers for them, but they produced three
times the conversion rate, or the percentage of visitors who ended up buying something.
"We now can quickly and easily identify the individual terms that are driving
sales, or driving costs, taking the guesswork out of search marketing," says
Feather Hickox, Restoration's Web manager. "For a company our size, it's
critical."
Coremetrics, like WebSideStory and many other Web analytics firms, is an "application
service provider," selling its analytics tool as a subscription-based service
that its customers access using a Web browser. Costs for the services vary: Coremetrics'
charges start at $6,000 a month for its total analytics package, and $4,000 a
month for the marketing module alone, depending on the volume of site traffic
or total number of marketing campaigns that are being tracked. WebSideStory's
service starts at $15,000 a year.
Inevitably, all this data collection raises concerns about privacy. Many analytics
companies use "cookies" to identify each visitor to their customers'
sites; the cookies let trackers tell when a customer returns to a site, so they
can give credit to a particular marketing campaign even if the shopper doesn't
purchase until the second or third visit. The analytics companies say they don't
keep any personal information, such as names or credit-card numbers, that can
be used to identifyindividual shoppers.
Still Some Mysteries
Despite their success at unraveling the mysteries of online marketing, more data
can't always explain everything.
CompUSA Inc., a Dallas-based computer retailer, began using Coremetrics last fall
to test the effectiveness of different e-mail marketing messages. Its marketers
sent to a group of prospective customers nearly identical messages with different
subject lines. Some subject lines touted specific prices and products, while others
offered more-generic low-price messages or highlighted new technologies, such
as Intel Corp.'s new Centrino wireless chip.
But they found that no single message consistently brought in more sales. The
message with the price information did well during the post-holiday period in
January and early February, while the Centrino message outperformed the others
around the time Intel launched the new chip in March.
The test is continuing, but CompUSA's marketers are still trying to make sense
of the results. "We haven't found the key," says Al Hurlebaus, CompUSA's
director of e-commerce. "Sometimes I have to remind my team members, and
remind myself, that we've been in this business for years and we're very successful.
It's not necessary that we forget everything we've learned and use the tool to
make decisions for us."
Mr. Totty is a news editor for The Wall Street Journal Reports in San Francisco.
Write to Michael Totty at [email protected]
|
|