CartonCloud is able to 'learn' from the order in which Consignments are delivered to automatically order in the same way in future. Internally, CartonCloud uses a machine learning technique to retrospectively evaluate the order in which consignments were delivered, and then automatically order consignments in the future based on this.

It does this by assigning Addresses an order score based on the position they appear on Run Sheets. 

By using the data from multiple Run sheets the system can then average the score to estimate where an address should appear on newly created Run Sheets.

These orders scores can also be manually viewed and updated using the article Bulk Import Addresses (Set Address Delivery Order). It is important to note that if you have set up manual sort order values using the process in the linked article then enabling this feature will override the previous sort orders.

After an initial calibration period Addresses should then appear roughly in the desired order when a Run Sheet is created. 

If Allocators then make further adjustments to the order of a Run sheet for example in the Bulk Allocation screen then those changes will still remain as this setting only affects the initial order upon creation of the Run Sheet.

Changing the Setting

This setting can be turned off and on for the entire organisation, meaning all Address records will be given an ordering score.

Navigate to Organisation Settings by selecting More in the main menu, followed by Organisation Settings in the drop-down menu.

Then select the Features and Options tab.

Scroll down until you see the option 'Optimise address sorting order weekly' which should be enabled to begin delivery order optimisation. 

The number of Run Sheets used for sorting optimization is set to 5 by default, which means that the previous 5 Run Sheets are used to determine the ordering of the consignment in future.

We recommend keeping this number between 5 and 10. Using 10 previous Run Sheets will result in higher accuracy, but it will decrease how quickly changes to the order are recognised, for example, a store may move from the last drop to the first drop and this will take several days to start appearing in the first position by default.