Category Archives: data mining

Social Network, To SAAS or To Build Inhouse?

Websites like ning.com and onesite.com provide the ability for people to build their own social networks. If the network is no different than a web 2.0 face lift to an already existing mailing list or an egroup, then this is probably fine. However, if you are considering building the next startup that needs a social network using a SAAS offering, that may not be the right thing.

Only yesterday Facebook has released Facebook Ads, a new advertising platform. I think this is going to be lot more successful than the 0.04% click-through rate with which Facebook had been criticized a few months back.

Last few years, Google has ruled the online advertising landscape with “contextual ads”. The early social networking sites including LinkedIn and Ning (which I have used), while pioneered new concepts in social networking, hardly make use of the “intrinsic knowledge” of the members within the networks to make money. Instead, they just rely on Google’s AdSense (atleast partially), which makes use of the context, but not the network! But Brand is lot more to do with the network than with the context. Hence, to differentiate and make a good revenue out of custom advertising offerings, a Social Network should make use of the data it has about it’s network. Has to use advanced data mining techniques on both structured (such as well defined profile properties) as well as unstructured (comments, discussions, widgets being used) data to provide highly targeted branding advertisements. This is possible only if the data related to the network is available inhouse so that it’s possible to build those advanced mining techniques and experiments, and ultimately serve the ad impressions using the knowledge within the website as opposed to content on the website that is available to Google or any other contextual advertising platform.

Advertisements

1 Comment

Filed under data mining, Online Advertising, SAAS, social networking, targeted advertising