Optimising product purchasing through short-baseline predictions
Who for? Cold logistics operator (milk and dairy delivery)
What we did:
When you are selling goods that have a used-by date, the holy grail is to buy enough stock for your customers with no excess. We can analyse patterns in sales and supply to predict the required number of inventory items at various times.
But what if you don’t have years and years of data? Even in this case, you can still use statistics and business sense to generate a meaningful safety margin. That was the case with this client, who was developing a new business stream and did not yet have years of data to rely on.
Our approach was to take the data they did have, and use business rules to simulate a longer data set. We effectively added some statistical ‘wiggle’ to the observations, based on theory, to add a quantifiable, realistic and transparent safety margin. This forecast gave the client a baseline indication, which can be improved by adding real world data as time goes on.