Get started with predictive analytics

1. Conduct a high-level review of your business performance and target one area to work on.

2. Write down a single goal that you want to achieve with predictive analytics.

For example, making your advertising more effective.

3. Gather data sets from your website analytics, CRM platform, or any other tools with information about your customers and leads.

For example, if you want to know what types of products people are purchasing through your email marketing, you might gather data from Mailchimp. Include data with both positive and negative outcomes, so that you can make more accurate predictions. Use external and internal sources to ensure you have a large enough selection of data.

4. Export your data in an Excel or CSV file to integrate with market vendors more easily.

Select a platform such as RapidMiner, which allows you to add data without the need to code easily.

5. Use an outside expert vendor to help you quickly create prediction models that are easy to understand.

Creating prediction models on your own, using code, takes a long time and requires someone with a great deal of experience. Alternatively, there are applications coming on the market that don’t need quite as much expertise.

6. Pick predictive modelling applications that integrate with the third-party platforms you use.

For example, if you’re analyzing email marketing, pick an application that will integrate with Mailchimp.

7. Test your prediction model on a small scale to see if the results match up.

For example, if you analyzed email marketing sales, use the modelling application to send emails at times predicted to be most effective, then analyze the results.

8. If your test results line up with your predictions, increase the size of your experiment and keep feeding results back into your prediction model.