Using Data Mining and Big Data to Predict Customer Behavior

The future depends on what you do today, said Mahatma Gandhi. Especially if you’re a salesperson or a marketer. Predicting customer behavior is one of the essential elements of marketing research, but we can’t deny that it’s still difficult to know what our customers will actually do, when and how much they will buy, or how long we’ll be able to keep them. In spite of all these uncertainties, there are various sophisticated tools that can help us to, in a way, read our customers’ minds, and predict their next move. Data mining and big data analytics surely make all this easier, but some reports say that companies use only small part of the data they collect.

 Image Attribution

Larger Than Life Data

Predictive marketing analytics and big data are extremely powerful these days. People who aren’t into marketing and sales are usually very surprised when Facebook all of a sudden starts displaying them ads for products or services they Googled a minute ago. Companies use similar algorithms fed by big data in order to track and analyze their customers’ purchasing habits and behaviors and gain a competitive advantage. A few years ago, an unsuspecting teenage girl was a victim of big data analytics, when a retailer sent her coupons for baby products. Her shocked father denied that she was pregnant, but it turned out that the company was right. A seemingly simple analysis of the girls’ several last purchases showed that the items she had bought there were frequently found in shopping carts of newly-pregnant women, so the company put two and two together and decided to thank the customer for her loyalty. This true story illustrates that big data knows almost everything (even more than members of our immediate family), and that, despite this unpleasant example, salespeople should definitely put it to a good use.

Image Attribution

All Roads Lead to Your Landing Page

One of the most basic ways of identifying future customer behavior is starting with Google AdWords. This little tool can provide you with valuable insight into how purchasers ended up on your landing page. Basically, it starts from the beginning by identifying the keywords they searched, and establishing which ones, in particular, brought them to your site, and which ones resulted in a purchase. This kind of analytics will tell you what your buyers actually want and you can use it to additionally tweak your settings and remove all possible obstacles on their way to the shopping cart. Make sure to reduce the number of clicks necessary to complete that action.

Image Attribution

 

What’s in the Shopping Cart?

Affinity analysis is a data mining technique used to perform market basket analysis. It’s great for identifying associations and connections with certain products, and salespeople can benefit from this kind of analytics by offering their customers complementary products. Yes, we’re talking, among other things, about cross-selling and upselling, which means that this analysis is what feeds recommendation engines and it’s also essential for targeted marketing. There are various algorithms, but you can start by establishing whether your product line falls under the narrow but deep, or broad but shallow category. In the former case, your analysis should start from the item level by determining the best-selling product. In the latter case, you should analyze what categories are usually purchased. You’ll also have to decide whether you’ll observe the obtained data on the order or user level. The order level can help you improve your average order value by means of already mentioned cross-selling and upselling. The user level, as its name suggests, gives you some useful information about the customer life cycle so that you can develop nurturing strategies and maintain customer loyalty.

Image Attribution