ADOTAS – Whenever you ask a marketing executive about his or her company’s investment in social media, we hear this statement a lot: “Our social media investments just don’t quite work like online ad investments such as Google PPC, and hence it is very difficult to measure ROI in social media.” We examine the reasons and recommend some solutions here.
Social Media defined
For the purpose of this article, we assume all major social networks like Facebook, Google+, LinkedIn, Pinterest, Twitter and YouTube to be a large part of social media.
Why does Google PPC work better than all other social media?
Whenever a user is searching for a product on search engines like Bing and Google, there is an implicit intent, whether to purchase, to get more information or to complain about a product or service. It is because of this implicit intent that users are self-selected and are more prone to clicking on the ads.
On the other hand, Facebook is about connecting people, and hence its focus is the social graph. Therefore, an ad trying to sell a product directly on Facebook is equivalent to someone selling network marketing services at a social gathering. What do you do when someone at a social gathering tries to recruit you for their network marketing service? Ignore them! This is why most ads don’t work as well on Facebook, unless they are playful and/or just trying to reinforce the brands. There are a whole host of interest -based social media sites like LinkedIn, Pinterest and Twitter, to name a few. Unlike Facebook, these networks are built on some common interests that bind people together, whether it is job search or high-tech news. Therefore, discussing these common interests, whether one-on-one or in the form of ads, may be more acceptable in those forums than on Facebook. However, return on investment in these networks is still hard to gauge, as you really don’t know the intent of the users.
Finding the Intent
If we could easily discover users who express some intent in social media, whether it is intent to buy, intent to sell or customer service for a particular product or service, it will make the marketer’s job easier. One could try to do Twitter search manually and might find people with intent to buy certain products. However, considering that there are 230 million tweets per day, it is not humanly possible to do this manually.
Big Data meets NLP
This is a problem of big data, wherein technologies like Hadoop are necessary to be able to process this large data feed. When processed, Natural Language Processing (NLP) techniques need to be applied to discover various intents, including intent to purchase, intent to sell and customer service.
At my company, SocialNuggets, we have done the hard part and are now mining tweets and other social media content to discover intent to purchase. In analyzing a fairly large sample of 1 million tweets for few days, we found that about 1 to 2 percent of the users express an intent to purchase some product or service. When these users were offered a coupon or deal, a CTR (click-through rate) of 11 percent was observed.
If you are interested in increasing your social media return on investment, then why not try to find people that have expressed an intent in buying products you are selling?