Weather isn’t just a British obsession – it’s becoming a global one. But producing an accurate forecast is difficult. For example, a weather service may be able to predict tomorrow’s conditions in Cambridge with 57% accuracy, but look two days ahead and this drops to just 48%. Four days out, and the changes of getting it right are less than one in three
This might be a minor inconvenience if you leave the house dressed for a summer’s day and get caught in a downpour. But for DIY, grocery and fashion retailers, event organisers, agriculturalists, breakdown services, energy companies, healthcare providers and insurers whose activities, risks, decisions and resourcing are often weather-dependent, long-range forecast accuracy can spell the difference between a good year and disaster.
The weather may not be within our control, but many variables in business can be anticipated and managed. Any organisation can benefit from being alerted to risks around the corner or foreseeing customer needs before their competitors do.
Traditional business intelligence is what I like to call “descriptive analytics” – using historical data to explain what has already happened. Predictive analytics have also used historical data to identify what might happen in future, based on existing patterns and relationships and, as such, isn’t new. It’s hasn’t set the world alight because isolated transactional data, such as credit card purchases or other types of data that change over time, contain very little insight into the behaviour of the individual who generated the transaction. But when you can combine multiple data sources and analyse them in real time, predictive analytics starts to get a whole lot more compelling.
The advent of real-time data virtualisation, aggregation and processing has enabled automated, actionable insights through predictive modelling, decision analysis and optimisation, and transaction profiling. This is already leading to the creation of advanced, almost “neural” systems which can learn complex patterns amid large data sets to predict the probability that an entity will exhibit behaviours that are of interest to the business. And it’s not confined to structured data – embedded social insight is allowing enterprises to embrace a streaming, crowdsourcing architecture to influence their strategy.
So next time you’re at the supermarket, your checkout operator may not be able to speculate about the chances of your weekend barbeque getting rained off. But when he or she presents you with a money off voucher for that coffee your spouse likes (which you forgot to pick up), that’s the power of predictive analytics at work.
For some useful tips on applying analytics to your challenges, check out these Top Tips.