Sunday, January 25, 2015

Power of Data Mining in the Digital Marketing Age for Consumers Yuan Xu (Danny)


The Target store knows your family member’s pregnancy progress better than you do. Can you believe it?

For many people, this might be hard to even imagine, but it can be true if you shop there often. Just three years ago, New York Times published an article that raised a lot of tension on the topic of data mining. The story started with an angry father stormed into a Target store in Minneapolis demanding to see the manager. The father assail the store’s abrupt move of sending her teenage daughter coupons for baby clothes and cribs, accusing the store for encouraging teenagers to get pregnant. The store manager was so confused by the strange event, he gently apologized and called back a few days later to console the family. To his surprise, the father had a conversation with his daughter and apologized to him instead and said: “It turns out there’s been some activities in my house I haven’t been completely aware of. She’s due in August. I owe you an apology.”

All these might have been shocking for many, but definitely not statisticians such as Andrew Pole who was hired by Target to do data mining. Just a couple of years back, the terms Big Data, Data Mining were still vague to the general public but a small group of people.  At that time, people had already been utilizing computers to improve work efficiency in various organizations, yet not so many have imagined how Information Technology could revolutionize the way we actually conduct businesses, especially marketing.  Target is one of the many companies who are utilizing the amazing power of these methods to improve their business conduct, well, “improve” might be hard to justify given the unpleasant experience of the father. Nevertheless, these methods do open doors for future development in how marketing and sales are conducted. As much evidence would suggest, this is just the commencement of a greater trend. But how do the consumers view it?  

Every technology advancement we made could be a double edge sword, Data Mining is no exception. If not played safely, there is no guarantee that the power would not backfire. For businesses who rely heavily on long term customer loyalty, more often than not this could lead to a massive publicity crisis where thousands if not millions of customers would be affected. For example, when Facebook just started to put ads in the event streams on people’s Facebook page, there were a lot of push backs from users who view this as an insult to their privacy. However, these ads did prove to be more effective than billboards on the streets. A lot of the firepower of these targeted online ads actually come from data-mine information that people voluntarily post on their webpage. Assuming most of these information is true, Facebook actually knows you so well that they probably know more about you than any one of your friends or even your parents. This offers the online marketers great potential to present suitable products to people who have the highest probability to be converted into customers, not to mention the speed this matching process has.

One anecdote I heard from before is about a jeans retailer who had a lot of trouble selling off the remaining stock of its jeans. When they advertised on Facebook targeting people based on their posts, they had a much higher conversion rate since people who were targeted were inherently more likely to purchase giving effective data mining methods.  The retailer sold out all their stock, and the newly acquired customers found their jeans at a cheaper and more efficient way, who wouldn’t say it is a win-win situation? Although not every story turned out to have a happy ending, people do take advantage of the benefit of data mining results when they are presented appropriately.

So what might the future be for data mining and statistical-model driven online marketing? No one knows, but it do seem like customer analytics and purchase predictions are growing, which could potentially help us find our best-match product in a more efficient way. That been said, it is hard to judge how fast the society would accept such a way of marketing and sales.


In closing, let me quote Joanne’s comment on the NYT article: Target, just because you can, doesn't mean you should.

No comments: