One of the questions that regularly comes up in my discussions with friends like Kaiser Kuo and Bill Bishop,
is the lack of reliable 3rd party data for web traffic in
China. Their beef is that data from sources like Alexa and iResearch
is unreliable and open to abuse.
My answer is usually quite simple: It really doesn’t matter that much for advertisers.
An effectively implemented web analytics program will provide all the
data that is needed for a comprehensive online advertising optimization effort.
- Your media plan will tell you how much you spend on each media and each banner.
- Campaign tagging will tell you how many visits came form each banner and each media. It enables you to calculate cost per landing page visit (CPV).
- Goal tracking and conversion tracking will tell you how many
visit from what banner actually ended in the desired end action on your
site. This enables you to calculate the cost per end action (CPA).
These data sets enable effective optimization of media
spending, without relying on any data from the publisher or the ad
tracking vendor. All data is owned by the advertiser.
-
For campaigns that emphasize brand awareness, CPV is a good key
performance indicator (KPI). It tells you how many times your ad has
been seen by a visitor to the publishers site that found it interesting
enough to click on and interested enough in your content to actually wait until your landing page was loaded. - I find optimizing for CPV more effective than optimizing for CPM (Cost per 1K impressions) or CPC (cost per click).
- CPM is sub optimal, because an ad impression just means the banner has
been loaded from the ad server (or the tracking pixel has been loaded
from the tracking server to be exact). It does not mean that the ad has actually
been seen or that, if it was seen, had any impact on the observer. - CPC is
sub optimal because in China many clicks on banners seem to be
accidental and users will abort the loading process once they notice
that are actually leaving the publisher’s site -
For campaigns with a specific end action (or a set of specific end actions) as campaign goal, CPA is the most
relevant metric, since it allows you to directly link your ad spend
with the desired return.
There are two drawbacks to this approach
- It required continuous testing and regular optimization. (which you should do anyway)
- It does not help investors or VC’s, since their main interest in
tracking publishers traffic is gauging revenue potential by projecting
its revenue potential. Well, maybe it helps to talk to advertiser who
actually spend money on the site. The advertisers ROI could be a metric to understand if a publisher has the potential to grow its ad revenue.

Hi,florian, I’ve already translate this article in Chinese in my blog. See: http://www.chinawebanalytics.cn/?p=85 . Thanks.
That’s all useful to see if your ad WAS effective but it doesn’t help much when it comes to guessing if your ad WILL BE effective.
The third party online metrics controversy – part 2
Reader Tai Te, posted an interesting comment on the “The third party online metrics controversy” post, that I used to argue that 3rd party metrics are unnecessary to plan online marketing campaigns and and to evaluate campaigns success. This post generated a lot of interest due to Kaiser Kuo’s mention his the Digitalwatch blog. Tai Te writes:”That’s all useful to see if your ad WAS effective but it doesn’t help much when it comes to guessing if your ad WILL BE effective.”I highlight this comment because it echos common concerns about using web analytics metrics for campaigns planning and evaluation.I …