NONE: ONLINE-ADS>> Re: HealthGate Guarantees click-rate
ONLINE-ADS>> Re: HealthGate Guarantees click-rate
Bob Wyman (bobwyman_at_healthgate.com)
Fri, 27 Sep 1996 12:34:50 -0400
Varun Arora wrote:
>While it can be said that the covers of magazines,
> newspapers, etc., are most effective in terms of eyeballs,
> click-throughs are a different story altogether.
> The effectiveness of a click-through depends ENTIRELY
> on the promise of the product, it's relevance to the
> audience, and the design of the ad itself. "Spot's
>yield" makes no sense.
This is mixing apples and oranges. You are comparing impressions
(number of eyeballs) to response to impressions (click-throughs). One
is necessary for the other, however, they can't be directly compared.
"Spot's yield" does make sense as can be clearly seen by analyzing
the ad logs of any well run site. For instance, at HealthGate, we
currently offer four ad "spots". Two (HealthyEating and HealthyWoman)
are in areas which offer consumer oriented wellness information.
Another spot, "search_form" is on a search form used when searching
our 16 million document online health information databases. The
final spot, "search_results" is on the page that shows the results of
a search in the database.
Given identical ads, each of these spots generates a different
average click-through rate. The rates are: (this week...)
HealthyEating/HealthyWoman: 5% click-through average
search_form: 3.4% click-through average
search_results: 2% click-through average
A bit of analysis (and questioning of users) shows that these
numbers make sense. When people are reading general the
"magazine-like" information in our HealthyLiving sections, they are
informally browsing and usually don't have specific data they are
looking for. These people are easily distracted by an effective ad.
On the other hand, readers who are on the "search_form" page are
there because they have some specific information they are looking
for. They are more focussed and are thinking about how to form their
query, not looking at the scenery. These people are harder to
distract. Now, the "search_results" form gets the lowest average
click-through rate for similarly clear reasons. When people are
presented with the results of their search, their minds are
completely focussed on evaluating the results. It is *very* hard to
distract someone who has just been presented with an answer to a
question they asked -- until after they have finished studying the
answer.
Given that we can see different response rates for identical ads in
different spots I think you are very wrong when you say that
measuring a spot's historical yield "makes no sense" or is
"ridiculous." We have clear statistical evidence that it does make
sense. Just as back-covers on magazines, SuperBowl Ads, and
classified ads all have different response rates. The difference is
that, unlike traditional media, we can measure and accumulate clear
evidence concerning the varying response in our spots.
Varun Arora wrote:
>A spot effective for one ad may be totally ineffective
>for another.
This is, of course, a very true statement and advertisers should be
very careful to ensure that they pick their spots well. However, the
evidence I present above should show you that some spots are more or
less effective as click generators no matter what ad is shown in
them. What we can see from this is that advertisers should be very
careful about the spots they pick to display their ads. They should
also refuse paying a flat-fee for impressions. The CPM should be tied
to the historical yield of the spot in which the impressions are to
be generated.
Consider this: At a CPM of $20 an ad of "average effectiveness" in
HealthGate's HealthyWoman magazine would generate a click-through
rate of 5%, thus, there would be 50 clicks per thousand impressions.
This gives a "cost per click" of $0.40. On the other hand, the same
"average" ad would generate only 20 clicks if shown on HealthGate's
"search_results" spot. These latter clicks would cost $1.00 each if
the CPM was the same for both sections. Given this,... which spot do
you want your ad to display in?
Of course, picking spots involves more than simply looking at the
spot's historical click rate. In the example above, of HealthGate's
site, the HealthyLiving spots are seen more by consumers while the
"search_form" and "search_results" sites get more traffic from health
professionals, physicians, etc. Thus, there is a clear difference in
audience even within this one site. If you are targetting the
professionals, you should probably focus on the "searching" spots
even though they have a lower historical click rate. Similarly, our
pricing for the spot should reflect the demographics of each spot's
users. Thus, you won't see a direct mathematical correlation between
CPM and click-through rate for each spot on a single site if the
audience of the different spots is different. An advertiser should,
however, expect such a correlation between historical clickrate and
CPM for spots that have a similar audience (either within the same
site or on different sites).
Varun Arora wrote:
> [Filtering based on geography] isn't quite as simple as you
> might think. Even non-US domains can have a simple .com
> / .org / .net, wiithout the need for a country identifier.
Please... I've been using the Internet for over 15 years... I know
it's not simple. We base our filtering on more than just the root
domain name of the node. We also analyse who the IP block was
assigned to and who it's admin or technical contacts are. For
instance, your "indiawatch.org" address (for whom you are the
administrative contact) is clearly registered to an Indian company
(Pritish Nandy Communications in Bombay) even though your IP numbers
come from an IP address block assigned to MCI. If this was "simple,"
I would offer filtering today not "soon."
bob wyman
-----
HealthGate Data Corp.
<URL:http://www.healthgate.com/>
-------------------------------------------------------------------------
This week's Online Advertising Discussion List sponsors:
"The Ad From Hell" is just one of 11 songs that poke fun at marketing.
Check out http://www.bizblues.com for an ideal gift for anyone in the biz.
It's great blues music.
-----------------------------------------------------------
>>>>>>>>Announcing the 1996 CyberAtlas Web Advertising Roadmap<<<<<<<<
Need quick answers? Get a copy of the Roadmap and find your way around
click-through, banner design, and much more: http://www.cyberatlas.com
-------------------------------------------------------------------------
