Analytics: everyone knows they should be doing it, but far fewer know what that actually means. And when it comes to navigating your way between descriptive, predictive and prescriptive analytics, things can get really hairy. So how do you make sense of it all?
And when it comes to sales automation, how do you turn your analytics into positive results?
I’m going to let you into a secret. All those complex-sounding analytical models? It’s essentially the plot of Minority Report.
Let me explain.
In Stephen Spielberg’s movie adaptation of Minority Report, Tom Cruise’s character, police chief Anderton, has found a new way to fight crime. In the past, he lacked a way to accurately predict when a crime would take place – as do all police forces. When a crime is committed, the best they can do is take down as much information as they can, catch the perpetrators and prevent them, personally, from doing more damage.
At best, with enough data and enough care, they might be able to identify patterns that tell them more about the kind of people that commit the crimes, when and how. But it’s all reactionary: it doesn’t change the fact that things have already gone wrong and we’re all suffering the consequences.
In other words, crimefighting relies on descriptive analytics.
This is the simplest form of analytics. In the business world, where big data sets are abundantly available, it accounts for around 80% of the analytics that companies run and rely on. But, just as in policing, its use it limited: it’s great to know where problems cropped up and why, and to knock them on the head before they do any more damage. But if you can’t stop them from happening again, you’re still at risk of disaster.
Okay, back to Minority Report. Frustrated by the limitations of his descriptive-analytics model, Chief Anderton has adopted a new approach: he’s harnessed some brains that, apparently, predict the future.
Predictive analytics is extraordinarily complex. It uses a number of techniques (often in combination) related to data mining and modelling, statistics and machine learning to analyse data, recent and historical, to help analysts predict what is going to happen in the future.
The trouble is, once you start to rely too much on predictive analytics, you’re opening yourself up to some serious errors of judgment. Why? Because the data is intrinsically ambiguous. It requires specialist analysis to interpret it, contextualise it, and draw out reasonable conclusions – and even then, there might be multiple ways to read the results.
In fact, as Dr. Michael Wu of Lithium Technologies explains on his blog: “The purpose of predictive analytics is NOT to tell you what will happen in the future. It cannot do that. In fact, no analytics can do that. Predictive analytics can only forecast what might happen in the future, because all predictive analytics are probabilistic in nature.”
In Minority Report, this is where the whole project comes crashing down. No one really understands how the minds of the mysterious “pre-cogs” that predict the future actually work, just as plenty of companies that get excited about the potential of predictive analytics don’t fully understand how the technology operates or how conclusions are reached.
What they care about is getting predictions they can act on with confidence to stop disasters before they have a chance to happen. They’re seeking a simple solution that doesn’t actually exist.
As Chief Anderton discovers in the movie, there is more than one possible version of the future: in fact, the system he’s using regularly throws up alternatives (the “minority reports”), which are buried as a design flaw to prevent messing up the results.
Much of this is to with the adaptive nature of choice. Once we know how things might turn out, we have the potential to act differently, and so completely skew the predictions.
This is no less the case in business than it is, as Minority Report suggests, in the thorny issue of free will.
So where does that leave us? Thankfully, in a better place than Anderton. That’s because, instead of declaring efforts to predict the future fundamentally flawed and scrapping the whole idea, we have a third option: prescriptive analytics.
Prescriptive analytics is, at its core, predictive analytics with more realistic expectations. Rather than giving you one, absolute version of the future, it shows you multiple ones. It gives you a range of possibilities based on the data available, recommends one or more courses of action, and tells you the likely outcomes for each one.
In other words, it shows you the minority reports and asks you how you want to proceed.
What’s more, it constitutes at its core an ongoing feedback system, tracking the results of each decision you make and recalculating the likely future(s).
When it comes to sales and marketing automation, it’s easy to see how prescriptive analytics, in particular, present ongoing benefits.
Knowing what’s worked or not worked in the past (descriptive analytics) is great, but you need to know what actions to take in the future. Introducing a system that automates tasks based on an analysis of historical data (predictive analytics) might save you a ton of time, but you run the risk of making mistakes that only a human businessperson would be able to identify or contextualise.
Automated systems can collect and process data – but you don’t want a robot making your customer-focussed decisions without your input: they lack the soft skills and the subtleties to make it a success.
By linking up automation tools to your CRM with a prescriptive analytics system, however, you combine everything that’s great about automation with everything that’s great about having a smart, real person ultimately calling the shots.
Your prescriptive analytics equip you with the tools you need to make far better, well-informed decisions – but they don’t make them for you.
You get to make great use of the automatically collated data that gives you top quality descriptive analytics by turning into a useful roadmap for the future. Plus, you don’t have to bet the farm on one particular course of action. Instead, you can compare and contrast the choices that are open to you, with the help of in-depth predictions and insights to guide you on your way. You don’t have to make a basic, absolute, limiting decision between accepting or rejecting the solution that’s been presented to you.
In fact, by automatically feeding your choices and the outcomes of your choices back into the system, you can assess whether you’re on the right track and adapt your plans as the technology recalibrates the probable result. And it does this while all the time reminding you where it is you’re trying to get to and suggesting the most effective way to get there.
Less crystal ball, more SatNav. And far more empowering for you, in the driving seat.
Used right, the business benefits of combining prescriptive analytics and sales analytics are game-changing, guiding you well away from Minority Report-style pitfalls and comfortably towards actionable plans that take in all the possibilities.
Or, going back to Tom Cruise’s filmography, away from “Can you see?” and towards “Show me the money!”