Predictive Analytics – when to get started?

When is the right time to schedule maintenance for a machine? How much stock should you order for a particular product? Will a project stay within budget? These are examples of questions that can be addressed using predictive analytics. But when is the right moment to start? Since predictive analytics offers many possibilities and often involves uncharted territory, answering this question can be challenging. In this article, we provide guidance on determining when predictive analytics can be valuable.

Preconditions

1. The required data are available

You need sufficient data, and it must be relevant. For example, if you want to predict how many ice creams you’ll sell on a given day, you need data from multiple days, such as the number of ice creams sold (observations) and the temperature (a relevant variable). Without sufficient data and the right variables, reliable predictions cannot be made.

2. The organisation is ready

The organization must be willing to use predictions in decision-making. This can be a significant shift, especially if decisions are currently based primarily on experience and intuition. It’s not necessary for everyone to fully embrace predictive analytics, but it helps if a few people see the benefits and are willing to experiment. Decisions don’t have to rely entirely on predictions right away. Predictions can be used as advice, where human expertise and data combine to make better-informed choices.

Business case

Once the prerequisites are met, you can create a business case to determine whether using predictive analytics is a worthwhile investment. You estimate the costs and benefits based on assumptions.

The benefits primarily include more efficient processes and error prevention. Costs may include implementation expenses and software licenses.

For example: Suppose 3% of your inventory is lost annually due to expiration or obsolescence. By using predictive analytics, you could more accurately predict demand and reduce the loss to 1.5%. With an inventory value of €10 million, this would save €150,000 per year.

Pilot

You don’t have to start with a large-scale project. A pilot is a great first step and has several advantages:

• You can more accurately estimate costs and benefits.
• You test whether you meet the prerequisites.
• For a relatively small investment, the organization gets hands-on experience with predictive analytics.

During the pilot, you collect data and perform an initial exploration. This helps determine if the expected relationships in the data are present. For example, an ice cream truck sells more ice creams at 25°C than at 15°C. If this correlation isn’t visible, it’s essential to investigate why. There may be too few observations or missing key variables.

Conclusion

Predictive analytics offers opportunities for virtually every organization. Before starting, it’s important to check the prerequisites and create a business case. Then you can begin with a pilot project. This allows you to explore the potential of predictive analytics in a low-risk way. Curious about the possibilities? Feel free to contact us for a no-obligation consultation.ciency and informed decision-making. Let this be the first step toward unlocking the full potential of JD Edwards.