119: How To Do Predictive Modelling

Predictive modelling uses mathematical and logic to make a prediction about the future. So, as your transformation aims to move to a new, unknown future state, how are you going to make sensible, intelligent decisions. As opposed to guessing in the dark. Join Oliver Banks as he lays out the 6 steps to guide you to successfully build predictive models to forecast the future.

So, listen in to this episode of the Retail Transformation Show podcast to hear:

• Why predictive modelling is so important for transformation.
• How to get started when creating a brand new model.
• The 6 steps you need to build successful predictive models.

 

 

Introducing predictive models

A predictive model uses data to forecast the future and allow you to take intelligent decisions.

Predictive models can be used for many different aspects in a retail transformation. It could be to understand operational processes, estimating future volumes, mapping customer journeys or customer experiences. Also, models could be to forecast workload, or labour and resoucing. Productivity models are another popular use for predictive models.

However, there is an essential aspect to remember:

All models are wrong, but some models may be useful
George Box, statistician

So, whilst you were hoping for an exact forecast of the future, your model will not be that (dang!). However, you can still develop models that are accurate enough to make them useful. But to maximise the chances of that, the episode features a 6 step framework for creating useful predictive models.

The 6 steps for effective predictive models in retail

During the episode, Oliver laid out 6 steps for effective predictive modelling. They were:

1. Define – what’s the purpose of the model.
2. Gather – collect and clean the data.
3. Prototype – construct an initial section of the model.
4. Build – develop and complete the model.
5. Validate – check that the model matches reality.
6. Use – take action and adjust the future.

 

Have questions? Connect with Oliver

Oliver has worked with retail clients to create a number of predictive models. From process productivity models, volume forecasts, organisational models and customer journey mapping. So if you’d like to talk to Oliver about your models, get in contact and perhaps there is an opportunity to work together.

Reach out to connect with Oliver Banks on LinkedIn.

Or send an email – oliver.banks@obandco.uk.

Further podcast episodes to listen to

During this episode, there were two recommended episodes to listen to next.

Firstly, check out episode 111: Defining Your Perfect KPIs.

Also, scroll back through the achives and take a listen to episode 8: Transform Your Data Into Real Results.