In the third article of our behavioral analytics for products series, you’ll learn what’s a user from a product analytics perspective. A user can be either a person or a “non-person,” such as an automated process or a machine. For most commercial products this distinction is irrelevant as users are uniquely people. However, many products used internally by companies often implicate some level of automation or intervention by a machine. In this case, the distinction between human and non-human users is made using the properties associated with each entity, which are themselves captured through events.
As product managers, user experience designers and analytics strategists, we hope to link behavior to a unique user — it’s what makes our jobs easier and our goals, clearer. In analytics, it’s not as easy as you might think to determine who or what is the unique person or entity driving the events on your domain, but let’s try!
For practical purposes, we’ll only consider users to be people, not machines or processes. When a person interacts with a digital product, they do so on a device like a phone, a tablet or a computer. A person might have several devices. As a result, the data we collect on a user’s interactions is only as accurate as our ability to link a set of devices to a single person.
Distinguishing users and visitors
Every user’s behavior begins with an unknown device. If a user has never visited your product ecosystem before, then from your perspective, the only way you are able to categorize her is through her device. At this point, a unique ID, also known as a cookie, is assigned to the new device by your analytics system in addition to an anonymous user ID, in order to distinguish the person from the device. These users are referred to as visitors.
Because of a lack of information, the best we can do to categorize this visitor is to link all the events created from this device to an anonymous user or visitor. In effect, the visitor has no identifying characteristics other than the behavior (i.e. the set of events) that she has generated through her device.
If this same person accesses your product using a different device, to which you haven’t assigned a device ID, a new device ID or cookie is assigned to that device and another unique anonymous ID is assigned to the user of the device. From your perspective, you perceive this visitor’s 2 devices as 2 unique people.
Identifying visitors as users
The easiest way to know who is using your product is by relying on a visitor to identify herself. Most applications do this by requiring a visitor to create an account. Other applications do this by asking for identifying information in other parts of a product such as during a payment process.
When a visitor creates an account, a new unique user ID is assigned to that person. The visitor becomes a user. The device ID associated with the device used to create the account and the corresponding anonymous user ID are then attached to the new unique user ID. All the past behavior of the anonymous user is now linked to the identifiable user.
When the identifiable user logs into her account from another device, that device ID and its corresponding behavior also become associated with the unique user ID. As such, we are now able to link the person’s pre-account creation behavior with any post-account creation behavior. This is best illustrated in the table below.
Editor’s note: This article is the 3rd of a 4-part series about behavioral analytics for products. You can access the other posts using the links below.
- Part 1: Intro: The components of behavioral product analytics
- Part 2: The role of events in product analytics
- Part 4: How to define a session
To get notified when our next article is out, please sign up for our newsletter.