For those of us who can’t see into the future – so seven billion of us or so – making predictions has traditionally relied upon experience, judgement and instinct. Humans are geared up to do a lot of things adequately, constantly processing information to make day-to-day decisions like managing finances, interacting with other's and navigating the complexities of modern life. Of course, we often fail at tasks like these on a regular basis. Some tasks are beyond the capabilities of even the brightest individual. Highly specialised tasks, like crunching data, are best left to systems that are specifically designed for that purpose.
You can view the differences as 'spreading resources across a lot of tasks' or 'focusing them on doing one task at a very high level'. Take Tesco’s stark realisation that computer systems are far better than people at predicting what products are bought in combination, for example.
That’s just one example of how predictive analytics in business – using data to predict future events – can be used in business, though. While relatively few companies are aware of predictive analytics and how accessible it may be for them, fewer still understand just how broadly it can be used to answer crucial business questions. Here are a few examples.
How should I target my customers?
You likely have all sorts of data about your clients, like age, gender and location, what they’ve bought, when and how often, how they use your website, how they prefer to be contacted and how they found you in the first place.
But could you have told us that a 64-year-old man in Scotland is about to buy a new set of golf clubs due to nearing retirement? And being in the country that has the most golf courses per capita in the world, that this man's browsing history shows he’s got a soft-spot for TaylorMade clubs? And that he’s most likely to buy on the last Saturday of the month?
Unless you're able to join, process and model a lot of data and apply predictive analytics, it would be almost impossible.
How can the company RUN more efficiently?
Humans are notoriously inefficient. We come into work tired, preoccupied with last night’s TV and (ahem) sometimes hungover. Fortunately, our inefficiency is entirely predictable. You don’t need a computer to tell you that Friday afternoons aren’t the most productive of periods. Using predictive analytics to pool productivity data with information about suppliers, orders and logistics, however, might well show you when is best to receive shipments, to package products and to send deliveries to improve your overall efficiency and boost your bottom line.
Which job applicants should we hire?
We can all pick out the best job applicants on paper. Great education, plenty of experience, lots of interests outside of work. But how can you be sure they’re the best fit for your company? By combining CV information from applications with data about your existing workforce – that's how. You may find that graduates from mid-level universities churn at a lower rate than those from Oxbridge. Or that people who have worked at certain competitors are particularly valuable in a certain area. This allows you to form a view on which candidates are likely to go on to be great employees based on data.
Once again, we can see how predictive analytics might be used to deliver very straightforward and actionable insights but can do so in a way that simply wouldn’t be possible for us mere mortals.