Using the Social Graph for Targeting Ads
By James Wanless
President & COO of Talkster
A lot is being written on the implications for advertising and marketing within the context of social media. I read an article written by Joe Marchese of socialvibe. We both share the view that social media can be leveraged for targeted advertising, but there is a difference between knowing something about you, the owner of the profile, and the people who visit your profile. I would say that this distinction goes further. The knowledge of you and what marketers should target to you is not only a function of your profile but of your entire social graph which includes your relationship to your friends.
In this example, knowing that Joe and his girlfriend are getting married, visitors to her profile should get wedding gift ads. This can be further refined based if something is known about the visitor to the page.
In short, this highlights not only how to use information in the individual but the combination of knowledge about two individuals and their relative positions in each others’ social graphs. Take it a step further. Knowing that other friends on the same proximity on the social graph had already clicked through on certain gift related ads, this could also be relevant to visitors to her profile.
The most complex part of the equation for marketers is how to use this information judiciously. If you know just enough, then the ads become relevant and of interest. If you know too much then they become “scarily relevant” and will turn your targeted consumer against you as they perceive that you are invading their privacy. A whole new industry will evolve around marketers looking to strategically use this wealth of information that previously they could only have dreamed about. Done right, it will positively change the face of marketing forever. Let’s hope that a few don’t abuse the privilege of accessing the information and negatively skew public perception.
5 Responses to “Using the Social Graph for Targeting Ads”
Leave a Reply
You must be logged in to post a comment.
July 7th, 2008 at 7:33 am
The example given is why context based adds will never work: Why exactly would the visitors get interested in the marriage adds? Will s/he marry because of his/her friends marrying? No!
Those visitors should be getting adds based on their own interests and activities. The holders of that profile, in this case Joe and his girlfriend, should be getting marriage adds though.
This scenario changes with many events where friends might either become interested or not. Or in other words where it either makes sense to display contextual adds to the holder, announcing his/her interest in something and maybe looking for additional opportunities / information – or to the visitor who will now face the content and then eventually become interested in the topic.
July 7th, 2008 at 12:02 pm
Thanks for your comment. I think you have misunderstood my example. What I was trying to say was that you, the friend of the person getting married, would not get ads related to YOU getting married, but rather wedding registry ads which are relevant if you planned to give a gift at your friend’s wedding. The question of context and relevance is a very broad and far reaching one. My belief is that the more you know about a person, including their friends (related to the social graph), the more relevant you can make offers. Balancing privacy with relevance is the topic of the moment.
Thanks for reading our blog.
July 7th, 2008 at 5:47 pm
I have concerns about faceted identity. Are you serving up ads derived from my bondage channel to my wedding guests? Are you still serving up wedding ads after my divorce?
Relevance is the grail, but it’s not an easy serve, even with a social graph boost.
July 7th, 2008 at 6:41 pm
Thanks Phil. Faceted identities are very much a part of the social web i.e. we are different things to different people. However, I wish that the article I originally referred to had had a better example than the wedding gift. It’s a poor one as it is something that does not play out over a long period of time. However, to take the best element of the example, this time period of relevance can be useful. The wedding gift is only going to be relevant up until a period of time before the wedding. How close to the date is going to determine whether your ad comes from a bricks and mortar or online store. If the wedding is cancelled, then ads for online marketplaces might be good for you to offload your friends gift that you already bought and now aren’t going to need. If it’s after the wedding date, then the wedding “relevance” is no longer used as it has lost it’s relevance. So, I too share your concern for how to manage the relevance extracted or interpreted from the social graph. To try and pinpoint the relevance is going to be difficult and of potential concern to the targeted consumer. Personally, if I use Gmail, I have gotten used to the fact that the ads are “scarily” relevant to the content of my emails. In most cases, from a marketing perspective, it doesn’t cause me to click more when it is so targeted. Knowing that my friend and I both like a certain type of music, knowing that I have clicked through on similar offers in the past, and presenting me with music downloads is more likely to get me to click and then buy. I think that relevance can also be derived from a much more simple, old school, source. Engaging people in dialogue based on a question, quiz or opinion and then offering a recommendation for a product based on that dialogue can yield much better results than the more surreptitious use of the social graph.
November 30th, 2008 at 6:34 am
[...] target advertising more effectively. James has commented on this aspect in his recent post “Using the Social Graph for Targeting Ads“. Rather than being “just another call back service”, Talkster expects to [...]