Visual effectiveness metrics

From: Greg Edwards <gstuff_at_eyetools.com>
Date: Thu 12 Jul 2001 12:03:39 -0500

Hello all,

My company, Eyetools, Inc., is developing an _online
advertisement visual effectiveness_ module that reports
where people look on web pages and advertisements in
order to enable advertising companies and advertisers
to evaluate the visual effectiveness of various ad types
and creative. We do this by providing quantitative
feedback on where people in a study look on web pages
and where (and when) they look at advertising. WE NEED
YOUR FEEDBACK to determine whether advertisers would
be interested in knowing WHERE PEOPLE LOOK and what
they see (as opposed to only knowing where people do
and don_t click)? Additionally, should we continue to
offer this only as a full service, or would advertisers
be interested in being able to INTEGRATE THIS INTO THEIR
OWN STUDIES to do this themselves?

While watching the discussion streams about advertising
value and about how to make on-line ads more effective,
it occurred to me that this controversy lacks hard
metrics about what people actually SEE in addition to
what they click on. It seems that this is a missing
link in establishing the viability of on-line brand
building and the most effective ad formats and placements
to achieve this.

Eyetools is about to launch a software solution that
works in conjunction with eyetracking hardware that
enables companies to assess where participants in a study
look (yes, with their eyes) when they visit a website
(even a website that scrolls), and whether or not
visitors SEE the advertising on those websites. To date,
Eyetools has been providing this as a full service, but
we_re excited about releasing a software solution that
will empower companies to run for themselves the kinds of
studies we_ve conducted for the clients we_ve served.

Since the application automates much of the setup, data
acquisition, scroll-correction, image-capture, computation
and analysis that used to make eyetracking cumbersome to
use in the past, I_d like your thoughts on what additional
features you_d find most helpful.

My questions are:

1) If we were to offer such a _do it yourself_ solution,
   what metrics should we include?

 Statistics on individuals and groups already incorporated are:
 * time viewed on each ad/ad-type,
 * percentage of time spent on each ad-type,
 * amount of time before each person looked at each ad-type,
 * relative comparisons between ad formats,
 * and the ability to slice-and-dice by demographics.

2) What types of ad formats would be most interesting to
   compare/contrast?
 
 Formats already included are:
 * banner ads,
 * tile ads,
 * pop-up ads,
 * sky scraper ads
 * and the new large-format flash ads.

3) The tool is currently optimized for assessing the impact
   of visual design. What other factors do we need to consider
   in applying the tool when measuring advertising impact?

4) If we stage a limited release, a) who might be most
   interested in using the tool first, and b) who would be
   willing to share their findings with the rest of the ad
   community?

5) Would you want to collect this kind of data (or are we way
   off base)?

Any thoughts would be most appreciated. It's been fun following these
discussions on this list, and I look forward to hearing your ideas.

Best,

Greg Edwards
Chief Technology Officer & Founder
Eyetools, Inc.
gedwards_at_eyetools.com
phone: (510) 440-1600
fax: (510) 405-2022



Received on Thu Jul 12 2001 - 12:03:39 CDT