Event data could be the fastest growing type of data, and chances are your company has a lot of it. Though the term is new for many, we all generate plenty of event data every day whether we’re at work, home, or even hiking on the trail. What you probably don’t know, is that event data is a resource that is proving to be absolutely essential for businesses to better understand the behavior of their users.

By definition, event data is any identifiable occurrence that has significance for system hardware or software. Translation please? Event data is a continuous stream of actions that reveals the patterns of actions people, products or machines make over time. It is common user-generated events such as keystrokes and mouse clicks, phone swipes and taps, and even sensor data, Internet of Things (IoT) activity and other actions. When you play a Spotify song, swipe your debit card, scan a retail item, text a friend or turn on your television, you produce event data.

Event data has three basic components: a timestamp, one or more entities, and attributes. Any data point that has those three characteristics qualifies as event data. Here’s what each means:

  • Timestamp: Tells us what time an action took place.
  • Entity: Who took the action. This could be a machine or sensor or even a person(s).
  • Attributes: These basic characteristics tell us what happened during the event, like hitting send or clicking skip. The more attributes captured, the richer the data. Event data typically has many attributes, which tells us more about the user.

The challenge of event data lies in drawing actionable insights from all these actions rapidly taking place in enormous waves. And it’s exactly the massive size of event data that distinguishes it from other types of data. Just think about the number of times you thumbs up or down on Pandora, how many texts you send per day or how long you spend channel surfing. Any one person could generate hundreds, even thousands of events per day. Multiply that by the 6 billion people on this earth. Boy, that’s a lot of event data.

Event data is the language of behavioral analytics. Let’s consider a simple chain of events:

  • Event 1: Shopper clicks e-commerce link from advertisement on search engine.
  • Event 2: Shopper views the product at that link only for a few moments, before bouncing to a related item.
  • Event 3: Shopper views the second product for a few minutes.
  • Event 4: Shopper decides to purchase product and clicks ‘buy’ button.
  • Event 5: Shopper is prompted to login or register to the e-commerce website.
  • Event 6: Shopper closes window without purchasing.

In this case, the business would want to know what led the shopper to select the company’s website initially, why the shopper was drawn to the second item, and why the shopper ultimately did not buy anything. With enough event data from similar interactions, patterns emerge and answers appear.

In the example above, event data can help explain why the business was successful in getting someone on the website and why they failed to convert that person into a sale. Ultimately, the purpose of event data is to better understand user behavior in order to improve a company’s core product and boost its bottom line.  Since event data comes in huge quantities, fluctuates rapidly and contains highly variable attributes, it can be difficult to properly understand.

In my next post, we’ll discuss in greater detail how this data can be analyzed to yield accurate insights.