In a world where imitation is increasingly achievable, the margins for an organisation to gain competitive advantage are decreasing. As such, organisations are using modern technologies and analytical tools to find opportunities for reducing costs and creating value. Some of the most popular methods to do so are data and process mining.
Have you ever noticed that some offers pop up at just the right time? Or that the next show suggested by your streaming service is just the sort of thing you fancy? These are not just coincidences but are partially a result of data mining. This is what Capital One used to start tailoring the timing of their offers to individual customers, which grossly increased their rate of acceptance and return on marketing spend. It’s a similar tool that Netflix uses to track what you’ve watched, searched, or even hovered on for a while, to make suggestions for your next TV binge. These data mining tools take vast amounts of raw data and by tracing patterns, transform them not just into useful information but applicable and actionable knowledge.
Process mining is the less glamorous, yet just as vital, cousin of data mining that involves mining the data from processes to enable companies to identify and eliminate bottlenecks and create efficiencies. It may not be as well-known, but it’s the tool that could be part of making sure your deliveries get to you as soon as possible from the moment you order. Before looking in depth at process mining, we will focus on its basis in data.
Data mining is built on the foundations of classical statistical methods in combination with big data and algorithms. Big data has three defining properties – the 3 V’s – velocity, variety, and volume. Failing to appreciate any of these would limit the ability of a business to truly understand their data and to future proof it.
Velocity relates to the speed of data processing. Data can become outdated very quickly and it is therefore vital to always be working with the most relevant data. Accessing data quickly will be a priority for businesses going forward.
Volume is related to the amount of data being processed. With the advent of the Internet of Things and an increasing number of everyday interactions taking place online, the ability to record data has grown almost exponentially. The difficulty – and main objective – of data mining is to harness this data and turn seemingly irrelevant data points into quantifiable solutions.
Variety of data is as it sounds, the different types of data being recorded. For example, twenty years ago, the only way a retailer could track its customers would be purchases at the till and perhaps a postal subscription to a catalogue. Whereas now, a retailer can follow a customer through in-store shopping, online, an app and interaction on social media. These different data types can be combined and manipulated to build a more accurate image of who a customer is and, with forecasting or predictive analytics, what a customer wants.