I want to do nothing short of revolutionising how digital advertising is measured.
Currently advertising success is measured based on attribution models that are determined beforehand to estimate how much of the sales are attributable to each channel.
I want to move towards a control vs experimental condition design instead. The details of how it would work are as follows:
I need a way to create a tracking pixel that can store a few variables of information in form of a cookie in a browser.
That pixel would be attached to the ad in the ad server (I know how to do this in the Doubleclick products) and would fire once the ad is served.
As the ad is served the pixel creates a Unique user ID in a cookie on the browser of the user as well was at least 1 piece of information (for now): Was the ad in view (i.e. actually able to be seen) or not.
Those users that were targeted and served the ad but the ad was not in view would be excluded from further advertising and serve as the control group.
The second part of this system would be to place a pixel on the sales page of the advertising client which reads the cookies of the user.
This way we can measure a proper baseline of users who were not exposed to the ad (due to semi random selection) and those that were exposed to the ad and if there is any real uplift in likelihood of conversion.
This gets rid of the need to preemptively guess the mechanisms of conversion (did people click on the ad, when in the journey did they click on the ad or see the ad) and would simply measure true uplift between two groups that are identical besides the exposure to the ad.
This is as close as I can think to a proper experimental setup without incurring the prohibitive costs of serving placebo ads to part of the user group.
I have experience with Advertising in my day to day role and understand the DoubleClick tech stack well.
The coder must be able to set up a server from scratch and more importantly design the two pixels that can be placed in the ad and on the landing page.
Due to industry connections I can trial the product swiftly.
Please only apply if you have all of the below:
You have experience working with DoubleClick products (specifically DCM and DBM)
You know how pixels and cookies work and would be comfortable creating the pixel generator and the required server structure that records fires of the conversion pixel vs and which group they came from.