1. What should we do next?
Welcome. In this video, we'll cover the questions and solutions under the most advanced type of analytics, prescriptive analytics.
2. Understanding what should be done
Prescriptive analytics is similar to predictive analytics but it goes a step further. Instead of simply predicting what will happen, it considers a number of defined variables to achieve the best possible outcome, and then prescribes the course of action. It answers questions such as “What should be done?” or “What can we do to make something happen?”. We will now go through a few applications of prescriptive analytics in different industries.
3. E-commerce industry example
Let’s see an example that applies to the e-commerce industry. Whenever you go to an online shop, the site recommends a number of products to you. These recommendations are based not only on your previous shopping history but also on what you have searched for online, what other people who have shopped for the same things have purchased, as well as other factors including demographic information.
The analytical question, in this case, would be:
How can we develop an algorithm to recommend relevant products to customers, based on relevant variables?
The end goal is to find products that the customer has a higher chance of buying. A product recommendation algorithm would be the prescriptive analytics solution which would automate this process end to end.
4. Banking industry example
Another algorithmic use of prescriptive analytics is the detection of bank fraud. With the large volume of data stored in a bank’s system, it would be nearly impossible for a person to manually detect any suspicious activity in a single account.
The analytical question here is: How can we identify fraudulent behavior in our data, and automatically prescribe the best course of action?
An algorithm could analyze patterns in your transactional data such as location and amount, then alert the bank, and provide a recommended course of action. In this case, the course of action may be to cancel the credit card, as it could have been stolen.
5. Marketing example
Prescriptive analysis can also help reduce customer churn by applying marketing strategies on behalf of the marketing team.
So while predictive analytics say Customer A is 80% likely to churn, prescriptive analytics identify the specific actions marketers can proactively take to reduce that percentage and make it less likely Customer A will churn.
For example, a recommendation algorithm may suggest delivering an email, with a particular subject line on a particular date to Customer A to reduce the likelihood that they churn.
6. Airline scenario (1/2)
Now let’s go through a scenario in more detail. The pricing team in an airline wants to optimize airline ticket prices in real-time to maximize sales. A question that would point us towards a descriptive analytics solution would be What ticket prices in the past resulted in higher revenue? On the other hand, a question with a diagnostic solution would be: Why have the ticket prices last week result in sub-optimal profits?
Another way is to utilize predictive analytics solutions and ask questions such as: What is the expected customer demand in the coming weeks? These questions would provide important insights which would help the pricing team to adjust the prices accordingly. However, something we can observe from these questions is that these will result in a reactive approach and a manual process. A question that would lead to a prescriptive solution would be: What is the optimal price for a specific flight at a specific time to maximize sales?
7. Airline scenario (2/2)
To answer this question, a price recommendation system would provide a fully automated solution which would focus on adjusting the prices based on the different variables of customer demand, fuel prices and weather.
8. Let's practice!
Now that you have learned about the prescriptive questions and solutions, let's put that knowledge into practice!