How prescriptive analytics could be useful for your organisation
Prescriptive analytics is described by Gartner as “a form of advanced analytics which examines data or content to answer the question What should be done? or What can we do to make x happen?” Often used to help organisations boost efficiency and optimise outcomes, prescriptive analytics is all about having a certain goal in mind and working out what you need to do to get there. It uses existing data and considers different variables that may affect progress to then prescribe the best course of action to get the best result.
Adapting the journey with the endpoint in mind
The Climbing Analogy
A great way to understand the concept of prescriptive analytics is by thinking about rock climbing. The ultimate goal for a rock climber is to arrive at the top of the rock face safely and quickly. You could apply predictive analytics to work out the success rate of the available foot holes at each step by analysing various factors. For example, the climber’s weight, current position, skill level, and difficulty of the next hole. A predictive model would be able to select the hole with the highest predicted success rate at each step. However, the predictive model wouldn’t consider the entire path to the top.
A better strategy would be to take the overall goal into account and optimise the sequence of decisions to achieve that goal. Rather than looking at the decisions independently. This way, you can improve the route through continuous learning. A prescriptive analytics model would consider external factors such as wind and rain, nesting birds etc. A predictive model would recognise the climber’s goal is to reach the top of the rockface safely and quickly, then work out the best route to reach this goal.
The GPS Analogy
Another useful comparison is a GPS. You enter your desired destination, and it tells you the quickest or most efficient way of getting there. As you travel, the application takes more data in and recalculates your route to find the best way to get to the destination. If there is a delay on part of your route, it will recalculate.
How can prescriptive analytics help my business?
In a business context, prescriptive analytics uses artificial intelligence (AI) and machine learning (ML) to recommend actions to optimise processes and outcomes. It can help to maximise gain, reduce risk, minimise unnecessary costs, and boost value and productivity. Just to name a few benefits.
One example is content suggestions or recommendations. Suppose your objective is to boost engagement rates on your website. In that case, you could use an algorithm that, based on the browser’s history, suggests similar content they may find useful or interesting, encouraging them to engage with your content for longer.
Or, with sales targets to hit for a new product launch, you may use an algorithm to analyse the performance and customer feedback from similar previous products and to determine the design or specification most likely to make the new launch a success and boost sales.
A real-life example of prescriptive analytics
We recently created a ‘next best product’ model for one of our charity clients, which is a great example of prescriptive analytics in action. The charity’s goal was to boost sales online, so we built a model that uses historical and current data to suggest the next product(s) a supporter is likely to be interested in.
Getting started with prescriptive analytics
Although prescriptive analytics isn’t easy to implement, it is well worth exploring. It can drive a number of efficiencies and performance improvements for your organisation. However, as is the case with all analytics, the success and effectiveness of prescriptive analytics relies on having good quality data. As well as a defined end goal in mind. It also depends on the right systems, processes, and technology.
Data analysis software, like Apteco FastStats, can be invaluable in helping you implement and manage your prescriptive analytics strategy.
Azure Machine Learning Studio is another useful tool. With out-of-the-box building blocks, it lets you build prescriptive models without the use of code. Alternatively, you can use it to build bespoke models, although this requires a level of technical expertise. Azure also has the ability to monitor ‘data drift’. This helps you can keep track of how the model is performing over time as more data is being collected. If prescriptions change dramatically, it can intervene.
One last piece of advice. If you are using AI for predictive or prescriptive analytics remember to watch out for any AI biases. It is up to you to make sure your data is always being used ethically and responsibly. (More on this coming soon.)
How to analyse your customer data for better marketing Getting your data processed, stored, and managed correctly…
Predictive analytics is a powerful analytical method which uses historical data to predict what will happen next….
It’s one thing correctly storing, processing and managing your data. But how do you translate that to…