The intention of behavioural targeting is to track users over time and build profiles of their interests, characteristics, such as age and gender, and shopping activities. Online advertisements use behavioural targeting to display advertisements that reflect users’ interests.
Developing an understanding of behaviour is not a new idea, and it already has uses in a variety of related contexts. For example, behavioural monitoring is a long-standing technique in the context of Intrusion Detection Systems (IDS), law enforcement agents use profiling all the time (and when based on racial profiling interferes with the system of criminal justice), and so-called profilers, specially trained agents, look at a crime scene, autopsy data, victims, and likely pre-crime and post-crime behaviours of a serial killer to make psychological assessments (that are considered by some to be ‘worse than useless’).
Based on mining data, including inference and linkage attacks in which data is associated from one source with data and online behaviour from other sources, the profiling practice is being used in surveillance systems, to tailor our digital experiences for us, by for example marketers to try to get us to consume more, and by politicians to influence elections.
Behavioural analytics requires and includes:
Available data sets, possibly bought from data brokers or on a black market.
Real-time capture of vast volumes of raw event data across all relevant digital devices and applications used during on-line sessions.
Automatic aggregation of raw event data into relevant data sets for rapid access, filtering and analysis.
Ability to query data in a number of ways, enabling users of the analytics system to ask questions.
A library of built-in analysis functions such as cohort, path and funnel analysis.
A visualization component.
Common analysis functions
Cohort analysis breaks all users into related groups (sharing common characteristics or experiences within a defined time-span) for analysis in order to understand if a group is doing what it’s supposed to be doing (or what a company/organisation/agency/government wants those individuals to do) on a regular basis.
For example, in an IoT application, are a group of machines – say for example edge devices such as smart thermostats – streaming the necessary data? In an e-commerce site like Amazon, are customers in a certain demographic completing a purchase?
What are the desired actions or results? And what are the paths that can get the actor (user) to take those?
An individual may take any number of steps before reaching a desired state. A path analysis analyses all the points and actions that individuals take at each point. Streamlined paths to a desired state can be identified and points in a path where a barrier exists that keeps individuals from moving forward. Path analysis gives insight into why people are doing what they are doing, and at what point they are doing it.
Funnel analytics are used to identify users’ progress through defined steps towards a specific goal. It shows the narrowing of participants as they move along a sequence to an end state. For example, Amazon offers a portfolio of products, then a specific product page, online reviews about the product, a shopping cart button, fields for shipping options and credit-card entry, and finally a purchase button.
With funnel analysis the rate at which individuals follow the steps of a sequence to produce an end state can be analysed. In other words, how many people dropped out of the funnel versus those that actually bought the product? Who were they?
Funnel analysis can be combined with cohort analysis: Did a specific group of people or machines drop out at one stage of the sequence?