Mobius Predict uses prior activity to predict future requirements.
In the case of product deliveries or pick-ups, stops can be scheduled based on either a ‘fullness’ prediction, a visit frequency in working days or a specified next visit date.
This model can be used to predict either the ‘fullness’ or the ‘emptiness’ of a customer’s storage facility whether it be a warehouse, a gasoline storage tank or a grain silo.
The model considers each customer individually. For each data point belonging to a customer, the model calculates the change in volume between data points and the number of working days between them (where an individual customer record can be set to include / exclude Saturdays and Sundays). It then calculates the average change in volume per working day between each data point – we call this the delta. A forecast of the current delta is then created based on these historical data points and using weights for each data point in a moving average calculation.
Based on the predicted delta, the last known remaining volume in the tank and the number of working days since the last data point, the ‘fullness’ of the storage facility is predicted for the ‘next but one’ working day. If this exceeds the customer’s threshold percentage, a stop is created. The quantity of the stop is the predicted volume on the next working day.
The system needs at least five data points for the full delta calculation. If the model does not have enough information to make the delta calculation, it automatically creates a stop for the customer and assumes the worst case scenario – the customer’s facility needs immediate serving.
The visit frequency is calculated using the number of working days between the last visit and the next but one working day. If this is equal to or exceeds the frequency, a stop is created (otherwise the frequency wouldn’t be met). Similarly, if no last visit data point is available, a stop is created. The stop quantity uses the predicted volume to be picked-up.
Next Visit Date:
The next visit date can be used to force the system to visit a customer immediately or earlier than planned. It can also be used to schedule one-off visits.