According to advisory firm Quantzig, prescriptive analytics is one of the hottest data technology trends in 2018. Digital firms are churning out software to make using prescriptive analytics easier. We in analytics must also gear our minds toward this trend.
From Descriptive to Prescriptive:
If you’re a Texas resident, you know how moody the weather is. The forecast may tell you that there’s an 18% chance of rain. In a descriptive analytics state of mind, you would take in that information and connect it to a fact – “Yesterday, there was a 19% chance of rain and it poured.” This information becomes relevant to you as you’re stepping out again. You instantly switch to a predictive analytics mind frame, extracting a prediction from past experience. As a result, you grab your umbrella. This way, predictive analytics guides you through trial and error. It allows you to learn from mistakes and, in the process, gain useful insights. With the umbrella now in your hand, you’re overly prepared for something that might not even happen. Or it might – like yesterday, because “anything can happen”; read the black swan theory. But what if the umbrella that always unhooks, just doesn’t open this time? Did you think of that? No – because why would anyone think of that?
When trusted with clients’ money, we think of the unlikely. Predictive analytics are so often taken for granted, their own loopholes encourage us towards prescriptive analytics. The easiest way to understand this concept is by thinking of the process doctors go through when writing prescriptions. As doctors investigate symptoms, they’re traveling pathways on a mental map. They take one route, and then another. They assign higher probabilities to some routes and prepare themselves in the direction of that treatment plan. In doing this, they don’t rule out the infection that affects only 15% of the population. They will know to change the course of treatment mid-way or perhaps even end the treatment and start another one. If treatment A doesn’t cure, a doctor would draw an idea as to why and take that descriptive insight to suggest if treatment B is worth it. Think of the end-result as your wellness; the doctor is steering the wheel in multiple directions to get to that result.
For analysis, doctors use components such as experience, historical research, opinions of specialists and the real-time condition of the patient. Likewise, digital analytics uses algorithms, big data and data feeds to model possible courses of action, keeping in mind the agile environment where change must be implemented quickly.
The idea is to be prepared for our everyday analytics to ensure that we are keeping up within this wave of analytics. Here are some basic practices to adopt now:
1) Optimize: This is number one on the list, as the first rule of prescriptive analytics is to stay on your feet. Regularly adapting to changing results would warrant the flexibility to reverse a decision or action or take a different course. When a campaign starts to slide downhill, stopping it at that early stage will prevent a greater loss later. It will also allow for fair exposure to successful campaigns.
However, be sure to watch out for the context of the action. Sometimes a course of action may not materialize results until a given period of time. A medicine may require two weeks to show results. The doctor requires the patient to comply with this rule before recommending discontinuation of treatment.
2) Maintain Inventory: Responsiveness is crucial in keeping up with the agile environment. To be responsive, one needs to know the options. If Campaign A fails, what would be the next-best investment? Think of the inventory as a decision tree with various branches. Say you have the flexibility to take another risk. You would then head in the direction of experiments on your tree. On the other hand, if you were in need of a successful campaign at this time, you may play safe by opting for the option that provided success 75% of the time in the past. Keep planting branches on your decision tree – inventory goes wry if not updated. Make it a practice to renovate your inventory of options, courses and decisions.
3) Stay integrated: It’s easy to lose touch of the ground values when you’re keeping up with a changing digital space. It’s integral to know not only the ground rules but how elastic those rules are. How much of a leap is the client willing to take? What is the overall strategy for this year, and is your decision in line with it? Knowing the limits and bigger picture often saves us from redundant work. However, also be aware that circumstances change as the digital space is ever-evolving. Clients and firms today are more open to consider unusual ideas, especially if there is a path to implement those ideas.
Digital analytics is the rent we pay to stay in the digital space. As the technology that supports prescriptive analytics advances, our roles as analysts change. Our competitors know that and are already adapting to the changing technology. After all, everyone wants a larger share in the digital space – which is highly global in nature. This makes it even more challenging. Take a look at SVP Shannon Sullivan’s blog to learn more about the marketing challenges in today’s globalized world.