5/11/98 Fortune 170+
1998 WL 2501267
Fortune Magazine
Copyright 1998
Monday, May 11, 1998
Issue: May 11, 1998 Vol. 137 No. 9
Above The Crowd
Ginsu Comes To The Web Internet advertisers are trying to figure out how to
make this embryonic medium pay. Maybe they should model themselves after the
sellers of Ginsu knives.
J. William Gurley
Internet ads, like magazine and newspaper ads, have always been
priced according to a metric known as CPM--the cost an advertiser
pays per 1,000 people who supposedly see an ad. The Internet,
however, creates a market for more precise measures of success. With
the performance-based advertising the Net allows, an advertiser will
pay only when contact with the customer leads to an actual sale. In
other words, advertisers will pay for results, not ephemera.
To get an idea of how well the Net lends itself to pay-for-
performance, go to GetSmart (www.getsmart.com). There you can
comparison-shop for mortgages, credit cards, mutual funds, and
student loans. The GetSmart site, created by a small startup, is
paid only when a customer actually signs up for one of those
products. Another form of success-based advertising is so-called
affiliate programs. Amazon.com has no fewer than 35,000 affiliate
sites, each of which gets a 15% cut of book sales made by surfers it
directs to Amazon.
Nevertheless, most advertising on the Internet still takes the
form of banner ads paid for on a CPM basis. DoubleClick, a hot
Internet ad service that recently went public, is trying to push some
clients toward pay-for-performance. But most of the new-media
community shies away from this market for a very old-media reason--
pay-for-performance smells like direct marketing. For media types,
advertising is a creative venture, while direct marketing is crass
salesmanship. Asking them to consider pay-for-performance is like
asking them to move their vacation from Paris to Las Vegas.
In spite of the snobs, there are four reasons success-based
advertising will flourish on the Web.
Advertisers want it. Paying only for advertising that succeeds
shifts the risk from the advertiser to the content company running
the ad. If an ad generates no responses, the advertiser loses
nothing while the content company has used up inventory. So it is up
to the content company to determine which ads will produce results
and then to make sure that those ads get in front of the most
appropriate eyeballs. Lest you think content companies would never
swallow such a shift, remember the age-old adage: The customer is
always right.
Success-based ads took off during the last media revolution--cable
television. Cable channels have excess inventory--some have 24 hours
of airtime but only, say, three hours of decent programming. Direct-
response advertising has proved a great way to fill slots in those
less desirable shows. Want a Topsy Tail? Need a Ginsu knife or a
citrus slicer? Want to add Superhits of the Seventies to your music
collection? Just dial up the number on any of those 1-800 ads. In
most cases the cable channel gets a cut of whatever you spend.
More and more advertising space on the Net is going unsold. In
its most recent quarter, Internet bellwether Yahoo reported record
revenues and earnings--but it also reported that the number of pages
viewed on its site grew faster than revenues. The implication: On
average, Yahoo is collecting less and less money per page view, a
measure also known as effective CPM. The same is true across the
Web. While Internet advertising is growing at a healthy rate, it's
not growing as fast as overall traffic. The resulting inventory glut
could lead to price erosion. Performance-based advertising seems an
obvious solution. What better way to bring economic equilibrium to
the market than to fill marginal page views with performance-based
ads?
Success-based advertising offers Internet retailers the
opportunity to pay a rational price for marketing. How much should
you pay for a customer who is referred to your Website? Seems
simple. Treat the bounty you pay for referrals as an outsourced
marketing expense, and pay the same percentage of sales that you
currently pay for marketing. If you normally expect to spend 12% of
sales on marketing, pay a referring site 12% of any guaranteed sale.
Estimating the cost of acquisition is easy, but getting it just
right takes serious quantitative analysis. In markets like telecom,
cable, and credit cards, where customers sign up for subscription
services and churn is measurable, many companies compare customer-
acquisition costs not with any immediate sale of goods or services,
but rather with the lifetime value of the customer--future cash flows
for the entire time they expect to do business with the customer,
discounted by a reasonable rate.
This is a straightforward calculation for businesses that have a
long track record. The problem for Internet retailers is that their
track record is currently measured in months, not years. That makes
it hard to figure out, say, how many customers per month you lose, or
how many customers moving from a referring site to your site are
repeat visitors. Knowing those variables is key to making an
accurate estimate of what you should pay for referrals.
For instance, I have heard that CD retailers pay as much as $40
for a customer lead. Let's assume that at a given retailer the
average consumer spends $100 a year, there is a modest 10% churn in
the customer base, and the sales produce a 10% cash-flow margin. By
the standard calculation, the average customer would be worth about
$50. But what if 25% of the leads are repeat buyers? Then the
average value falls to $37.50, and the site is losing money on the
typical visitor.
There are scenarios in which overpaying might be acceptable. If a
Net retailer attracts enough customers, it can start earning
significant additional fees by referring them to other sites.
Consider Amazon.com. With 2.5 million users, Amazon is in a good
position to produce leads for others. Looking for a book on surfing?
Amazon could send you to a travel site that has a great deal on a
Hawaii vacation package. Now let's say Amazon gets 5% of the price
of each vacation sold. At that point, the company might be willing
to treat bookselling as a loss leader, making up the difference on
referrals, which have 100% gross margins. Bad, bad news for other
booksellers.
There is no question in my mind that success-based models will
eventually rule on the Web. Manufacturers selling products, doctors
or lawyers selling services--all will eventually use success-based
programs. This is not to say that impression ads will go away, but
this remarkably efficient advertising will serve a larger and larger
portion of this powerful new medium. The result will be interesting.
Extrapolate this vision, and you see an Internet that looks more like
a flea market than like Hollywood.
J. WILLIAM GURLEY is a partner with Hummer Winblad, a venture
capital firm. To receive an expanded version of Above the Crowd,
visit www.news.com; to subscribe to the E-mail distribution list,
send E-mail to listserv@dispatch.cnet.com with the following in the
message body: subscribe atc-dispatch. If you have feedback, please
send it to atc@humwin.com
Quote: AMAZON COULD TREAT BOOKS AS A LOSS LEADER AND RELY ON
REFERRALS, WITH 100% GROSS MARGINS.
---- INDEX REFERENCES ----
|
