Based on the same data we went one step further. In the previous investigation we kept the carton assortment as provided by our customer. When we analyzed the results, we found that a relatively large amount of air was still being shipped. As a result of this we challenged whether the available carton assortment is also the optimal carton assortment.

During the previous investigation we had a very close look at the data, removed incorrect data, identified unwanted behavior (packing of bulging cartons) and created the optimal carton selection for every order individually. As a result, we now have a clean and well understood reference set of data. This data set was extremely useful to investigate further if we are using the optimal carton assortment.

For this investigation we used a data set - in this case around 15.000 different cartons for one customer. We had a close look and we searched for cartons which did not filled well and made some guestimates as to additional changes to the carton assortment that would improve (decrease) the amount of air shipped. Based on that, we recalculated with two adjusted carton assortments. In the first case we replaced a carton with a smaller version and in the second case we added a smaller carton to the original assortment.

This trial ended with quite nice results. Although the number of cartons stayed the same (so no influence on transportation costs as we used a flat fee of 4 euros per shipping carton) the amount of air shipped dropped considerably. So, cost savings are the result of fewer consumables used (carton and plastic air bags) and reduced labor costs (labor to fill the empty space with air bags). In total the savings range from 17 to 40 cents per carton.

The good news is of course that these savings are on top of the savings originally achieved in our earlier test case as stated in the previous blog.

Back to our initial question: is your carton assortment chosen wisely?