Question:
We have created quite a unique model in that we use a blended targeting process which need to use 5 different targeting models. They are behavioral, geographic, demographic, preferential, and contextual. We have a process that runs verification algorithms that quantifies the user targeting data which then enables us to distribute highly targeted ads. We are in the process of moving this over to OpenX and need some consultancy and help getting OpenX to be able to support this ability. The other big function is that we are very video ad based. Our primary model is based on 1-30 second video ads which also must be able to support delivering a companion ad(s) with the video ad.
Answer:
Let me understood what I understand from your requirements.
You have gathered or you have process which will gather user’s data and using this data, you want OpenX to deliver best matched banners (or ads).
OpenX will have five type of targeting modules.
1) Behavioral :: According to users’ online history. Suppose, he clicked on “Auto” type banner so next time, OpenX should give some more priority
to same category’s banners.
2) Geographic :: Display ads by his location. (It can be found by IP or from user profile data.)
3) demographic :: Display ads using user’s profile data.
4) Preferential :: Display ads using user’s preference industry or categories.
5) Contextual :: Display ads by sites’ text or contents where banner is going to display.
Also, you want to display video ad on page. It means advertiser can upload 1-30 seconds video as banner and it will display in site page same as normal banner.
Implementing Geographic , demographic and preferential model in OpenX will take 3 days.
Implementation of functionality which will allow advertisers to upload flv banner and OpenX will display it in publisher site as it displays normal banner
will take 1 days.
But we need more details/clarification on how you want to implement contextual model and behavioral model.
So, We need more details on above targeting models and after your detailed requirements, we can give exact time-estimation of this functionality.
