NONE: Re:"industry average" in click-through
Re:"industry average" in click-through
Michael McCarthy (mac_at_wpi.com)
Fri, 21 Jun 1996 18:15:27 -0700
The question of "average" clickthroughs is fraught with peril, making it
an excellent stick for advertisers to beat defenseless publishers.
How readers respond to your banner ad clearly depends on many factors, not
least of which is the banner itself. A given advertiser can gauge the
effectiveness of a site in generating clickthroughs compared to other
sites using that advertiser's same banner. But comparing across
advertisers, or based on Net-wide "averages," is a fool's game.
One of our Web-only magazines, SunWorld Online
(www.sun.com/sunworldonline), is a technical magazine aimed at Unix
professionals. Our advertisers sell expensive hardware and software; one
sells $15,000 Unix laptops, another sells $40,000 servers, for example. We
sell, not single banners, but packages of banners scattered throughout the
site. Our clickthrough rate *appears* at first glance to be 1% to 2% -
which doesn't sound right if this is Unix gadgets aimed at Unix readers,
and industry-wide it's supposed to be higher.
But our banner package makes the numbers misleading. A single reader, who
on average visits four articles (according to our surveys), will see a
given ad banner three or four times. He's only going to click on it once
in any given visit. The better measure for *our* advertisers is how many
visitors click an ad in a given month out of all the visits made to our
site that month.
That calculation turns out to yield 4% to 10% as a percentage of our
readers who click on some ad in any given visit. This makes more intuitive
sense.
Of course, clickthroughs are as useless as the old bingo-card measure of
reader responsiveness. The true measure of the value of an ad vehicle is
how much it drives sales. For the most part, advertisers online are as
reluctant to do the work necessary to follow the value chain from ad to
sale as they are in the world of print. When they do, they get surprises.
SUch as one advertiser of an attractive but expensive gadget who got 3,000
clicks in two months - which was good. Forced by their CFO to measure the
effectiveness of their ad budget, they found that out of these 3,000
visitors, they got only 25 qualified buyers worth pursuing - which was
bad. Of those 25 buyers, three had signed purchase orders - also bad.
Those three buyers had bought $330,000 of equipment. !!
It gets even better: These sales had taken two months from response; their
typical sales cycle from print ads is nine months.
Their clickthrough rate was 1.5%, measured conventionally.
So what?
So there are so many variables, the idea of an average clickthrough for
the industry - which combines surfer sites, entertainment sites, consumer
sites, sports sites, scientific sites, technical sites, research sites,
search engine sites, the entire range of publishing -- is ludicrous. How
can you average Publishing?
Michael McCarthy
President
Web Publishing Inc.
SunWorld/JavaWorld/NetscapeWorld
mac_at_wpi.com
