NONE: RE: ONLINE-ADS>> Is targeting a bad thing when everyone does it??
RE: ONLINE-ADS>> Is targeting a bad thing when everyone does it??
Jim Meskauskas (Jim_at_HAWKMEDIA.com)
Thu, 28 May 1998 11:35:17 -0700
Richard has brought up some very interesting points about
'targeted' impression selection. When I spent most of my
days as a buyer, I bought a lot of technologically targeted
impressions, everything from impressions served to
Netscape-only browser users to SIC codes to only those
accessing the web from California. I've seen them all work
with varying degrees of success. Though adding targeting
criteria to a buy comes at a premium, the theory is that
both exposure to the more specific audience is worth more
and that response rates will be superior because your ad is
finding the more appropriate audience for its message. But
Richard brings up an important issue by asking what is the
overall effect on inventory, the campaign, and even the
property as an ad vehicle for other advertisers' campaigns.
Let's break this down into questions:
>If everyone is picking only the "targeted" impressions,
what is the leftover stuff worth? Does it have any value at
I say that the leftover stuff is worth whatever the vendor
can sell it for. Granted, the remaining inventory is worth
less to the advertiser wanting the targeted impressions.
But, sort of like with airlines, once that plane takes off,
you can never sell that empty seat again. Better to get
something for it than nothing. That's why we have Flycasts
and AdAuctions out there. This is provided that the site
isn't already popular enough with buyers to sell it's
"non-targeted" inventory to other advertisers without any
adjustments to rates or delivery. Otherwise, the inventory
can either be used to bonus existing advertisers, giving
them an impetus to continue running with the site, as well
as lending extra frequency that may allow for some amount of
awareness building. Or perhaps the site can run that
remnant inventory against cost-per-response buys. I've
negotiated some very substantial cost-per-visit buys with
some online properties you would never think willing to
entertain them, but they've gone for it and thrown that
remnant inventory against this sort of buy. You think a
property like Yahoo! sells out of their ROS inventory? It's
got to go somewhere...
>Does this "dilute" the inventory?
I would say 'no.' Though it does have an effect on the
remaining inventory by lifting out the targeted impressions
from the rest of the inventory, just like any other buy, it
merely restricts the inventory. Given the example of
"Win95" targeted impressions, however, if the site doesn't
have a very big audience, and you've purchased what is in
essence ALL of the "Win95" impressions, it's possible that
the remaining inventory wouldn't be very valuable and
wouldn't generate a lot of exposure for other advertisers.
But this is rare. Most online properties experience
"growth" inventory, which would allow other advertisers'
banners to be served on some of those targeted impressions.
For example, I buy out the inventory on the keyword "rose"
for 4 weeks in February based on impression levels generated
in the month of January. Let's say it's 100K. But I forgot
that Valentine's Day is coming, and there are 3 times as
many searches on the word as there were in January,
generating 300K impressions on "rose." Well, I only get the
100K (unless I've built in some kind of Right of First
Refusal on growth inventory), allowing other advertisers to
run on the keyword. This principle would hold with any kind
of targeted buy.
>Moreover, how does my cherry picking affect other people's
banner performance? Will banners running concurrently have
depressed >performance levels because that bastard Richard
Hoy is getting first pick of the impressions he wants?
This is probably the most interesting point. It's difficult
to quantify the effects on other advertisers' banner
performance. Are the impressions "that bastard Richard Hoy"
is getting going to those more likely to click on banners in
general? Are "Win95" targeted users going to click on his
banners in lieu of another advertiser's? My instinct is
that response rates for other advertisers' banners would be
to scale of the weight levels they are running. Depressed
performance levels would be more a function of competitive
advertising or less targeted inventory. By getting first
crack at the impressions he wants, Richard is taking them
out of circulation, restricting the inventory and perhaps
forcing a come-lately advertiser into less desirable
inventory, which could in turn affect response. But it
probably wouldn't have significant ramifications on BEHAVIOR
or the flow experience, which is what really lies at the
heart of response. If my ad taps a viewer's latency, that
viewer is just as likely to respond to my ad as if the
"Win95" impressions weren't taken out of circulation. In
fact, if all the "Win95" platform impressions are gone,
leaving only ad impressions to be served to older platforms
or Macs, let's say, those ads may get better response rates,
given the likely "newbie" psychographic of an older platform
user. Chances are non-Win95 users would be more susceptible
to the 'novelty' aspect of banners and would "Click Here" if
prompted to do so. But then there are a lot of 'newbies'
who've just gotten the latest hardware and have a Win95
platform. It's hard to say.
Just like general market media, online inventory is subject
to supply and demand. The deeper pockets get the bigger
buys, buy further out, and can potentially lock out other
advertisers after the same inventory (this is the case with
many of the e-commerce advertisers, e.g. Amazon and CDNow).
This could have a number of ramifications. Higher CPMs on
those properties where demand is high will probably be the
first result. Though, as the online universe continues to
grow, so does the inventory. Another could be advertisers
using more sublime strategies, going after inventory on less
popular but still desirably targeted sites. Often times
these properties yield better response rates at a lower
731 Sansome St., 5th Fl
San Francisco, CA 94111
This week's Online Advertising Discussion List sponsors:
Internet.com and ICONOCAST
50% of Internet.com's high tech audience does not frequent CNET or ZDNET
Reach the Internet audience you're looking for! For more info email
sl_at_internet.com or go to http://www.internet.com/mediakit/index2.html
"ICONOCAST is something I look forward to getting and continuously
share with my sales staff." Jed Savage, VP/Sales, CBS SportsLine
To subscribe, visit http://www.iconocast.com
Online Advertising Discussion List To Unsubscribe send UNSUBSCRIBE
http://www.o-a.com/ to online-ads-request_at_o-a.com