Prescriptive Analytics

Analysing data to gather insights is just the first step. . .

Businesses are obsessed with data, and with good reason. The more data that organisations can collect, the more they can build strategies that consider previous, current, and future scenarios. Unlike descriptive and predictive analytics, which are concerned with what has happened and what could happen, prescriptive analytics is an umbrella term that binds together all forms of data discovery to work out what must happen to achieve a certain goal. Forrester defines prescriptive analytics as ‘any combination of analytics, maths, experiments, simulation, and/or Artificial Intelligence used to improve the effectiveness of decisions’. The field of data analytics is expanding, but what does this mean for businesses, and the decision making process itself?

Information, insights and intention

The aim of prescriptive analytics is to inform decision logic so that the best course of action can be taken in any situation. It’s plain to see why this has captured the attention of businesses, given the growing complexity of process operations integrated with disruptive technology. As well as making prescriptive analytics possible, smart data discovery capabilities, machine learning and the automation of the analytics workflow have motivated adoption. At the same time as drawing together vast amounts of relevant, real time data, prescriptive analytics methods can also explore notoriously difficult dark data. Companies and non corporate bodies now have the ability to use the software to pool ideas, and choose whichever is most likely achieve the desired end result.

Ongoing technological development, teamed with a willingness to understand data, means that prescriptive analytics will play a major role in forming the policies of businesses across the scale. Take healthcare, for example. The industry relies on an influx of data to function efficiently. Information is gathered from patient records, medicine administration, hospitals, pharmaceutical companies and even the social climate to offer better care at a lower cost. The better data analysis becomes, the more effective healthcare will be. And if that data can be collated, in real time, into actionable insights, many of the traditional roadblocks within the industry could be avoided. For example, if a UK pharma company knows from hospital admissions that a certain illness is on the rise, prescriptive analytics might suggest supercharging the production of the drug needed to treat it. What prescriptive analytics also does is explain the consequences of that action – for example, can the company afford to redirect resources, and how will it affect the production of other products? This complex interplay between choices can be applied to essentially any organisation in any industry sector.

Friction in prescription

Prescriptive analytics is not infallible, and faces the same pitfalls as any data driven tool. Data, unfortunately, is not always accurate – and neither are the systems that sort and store it. The quality of data analytics relies on the coherence of the data itself. One misleading entry or anomaly could confuse the algorithms, leading to ultimate decision choices that are far from the most useful or applicable. So, while handling dark data is certainly one of the strengths of prescriptive analytics, it could easily become a weakness. It’s up to the organisations that use data analysis, especially when applying prescriptive suggestions, to fully consider each possible outcome. Ironically, even this could compromise accurate, actionable analytics. The final choice is down to humans, who are certainly capable of getting things wrong. As well as digesting dodgy data or making miscalculated choices, another barrier to the wider adoption of prescriptive analytics is the sheer amount of data needed to come up with tangible conclusions. Only a handful of organisations currently have the ability to access the complex datasets needed to plausibly influence business strategy. Data analysis is fast becoming a ubiquitous business tool, but not all data is made equal.

Disrupted decisions

Despite the various obstacles that have and could continue to hinder prescriptive analytics, the complex tool will help data (in all of its many forms) to become more useful. In turn, this is likely to encourage organisations to be more diligent when it comes to collecting, storing and securing information in order to get the best results. Businesses will gradually abandon ‘unactionable’ analytics in favour of tangible decision suggestions, and, as a result, change the nature of corporate management. Prescriptive analytics means that decisions will theoretically be better informed, but it will also take away much of the pressure when it comes to choosing specific routes. In other words, if something goes wrong, then the software can be blamed. Distinguishing between human and machine error will be important when negating responsibility. The uptake of prescriptive analytics is also evidence for the slow but certain rise of AI – arguably the ultimate prescriptive technology – within business infrastructures.

Alongside the imminent enforcement of GDPR, prescriptive analytics is part of the relentless data revolution that has fundamentally changed how businesses operate. Instead of a daunting jumble of information, data is a tangible tool that organisations, corporate or otherwise, can take advantage of. This has largely been down to the development of advanced data analysis methods. Any business in any given sector could benefit considerably from the in depth decision insights of prescriptive analytics, from manufacturers to marketers. However, it should never be fully relied upon. Instead, organisations should integrate the software gradually and apply common sense to any and all decisions. Predictive analytics represents the future of Big Data and is something that all organisations, provided they have reliable information, can use to their advantage.

Should small organisations apply prescriptive analytics techniques to limited datasets? Can businesses still benefit from ‘unactionable’ analytics? Will GDPR compliance enable more organisations to use prescriptive analytics? Share your thoughts.