Predict the future using Predictive Analytics

Predictive Analytics can be used to make reliable predictions based on new data. Every organization has to deal with this and anyone who cares about their business – and data – wants to work with this. Why? Because this powerful concept can simultaneously optimize processes, improve the customer experience, and greatly reduce costs. If you want to successfully apply predictive analytics in your organization, sooner or later you’ll have to answer the following questions:

  • How do you find the business case and areas where predictive analytics can make a difference in your organization?
  • Which disciplines in your organization do you need to involve in the development of predictive analytics applications?
  • How do you build expertise within the team, and which skills do you need to develop?
  • What role does Big Data play in the development and implementation of predictive models?
  • How should you deal with the necessary changes in the division of labor and company processes?
  • What Predictive Analytics software is available, and which tools are the best fit in your situation?

On top of that you have to consider awareness in your employees and management, because using predictive analytics can cause a sea change in your organization.

Caring about the customer

It all comes down to caring about the customer and wanting to make structural improvements to the business. A credit card company will use Predictive Analytics to detect credit card fraud, thus minimizing the damage to themselves and the consumer. A hospital will use it to detect diseases in an early stage, minimizing the damage to the patient and thus also saving taxpayer money. Predictive Analytics will always be successful when you combine these two aspects: improving core processes and caring about the customer.

Example I: Secret agents and the Rosetta Stone

Before there were computers, but there were secret agents, Predictive Analytics were used to decipher coded messages. The first step, building a model, consists of counting the words in random texts concerning the (most likely) topic of the message. By comparing the amount of combinations of letters and numbers, certain terms can already be predicted. That makes it possible to filter out articles and direct verbs. If you know what a small amount of words are in code, you can use that information to decipher the rest. This technique, among other things, is what allowed Jean-François  Champollion to use the Rosetta Stone to decipher the ancient Egyptian hieroglyphics.

Example II: predicting customer behavior

Predictive Analytics are also used by stores (including online) to predict customer behavior. Using these predictions you can, for example, manage the stock. If there’s a period of warm weather coming up in the summer, people are likely to break out the barbecue. For stores, this means greater demand for thin meats, meat on sticks, condiments, salad, and baguettes. When spaghetti sauce is on sale, pasta sales are likely to increase as well. On the other hand, if Coffee Brand A lowers its prices, other brands of coffee are likely to sell less. Understanding these inter-dependent relations is important when managing a store.

Predicting future behavior is not trivial

Simply making connections between certain types of data isn’t difficult, like the warm weather/barbecue relation mentioned above. Using these relations to predict future behavior, however, is not trivial. Important questions to consider:

1. Has the new information been gathered under similar circumstances and over a similar time frame as the the information the model is based on?

2. How do you incorporate a change of the circumstances in the model?

3. What is the initial reliability of the model’s predictions?

4. Is it possible to estimate the model’s long-term reliability?

The first question can be answered using the information and the way in which it has been gathered. The answer to the second question doesn’t just depend on the way the information is selected, but also the way in which this information is used in building the model.

6 common methods of Predictive Analytics analysis

The reliability of the model used, initially and in the long term, depends greatly on the method used to analyze the information. Examples of oft-used analysis methods for Predictive Analytics are:

Which method is usable for your problem, and which isn’t, depends largely on the questions you’re asking and the desired model you want to use to solve your problem. You can apply all these models in practice to fit your situation in our R training course.

Tools to develop predictive models

In the Netherlands alone, there are over twenty tools available for making Predictive Analytics models. The most advanced tools, like SAS Business Analytics and BOARD, are also capable of processing and visualizing large amounts of data, and they contain ready-made models. Click here for an overview of all BI Tools containing predictive models.

Unique selling points of Passionned Group

✓ We make organizations more intelligent.
✓ We are 100% vendor-independent. We serve your interests alone.
✓ Our data scientists are experienced and battle-scarred when it comes to Predictive Analytics.

Want to get started with Predictive Analytics?

Do you want to experience the possibilities of Predictive Analytics? Contact us for an appointment with Passionned Group’s data specialists.

contact us

A selection of our customers

Become a customer with us now

Do you also want to become a customer of ours? We are happy to help you with predictive analytics (predict using data) or other things that will make you smarter.

Daan van Beek, Managing Director


Managing Director

contact me directly

Fact sheet

Number of organizations serviced
Number of training courses
Number of participants trained
Overall customer rating
Number of consultants & teachers
Number of offices
Number of years active