FROM START TO FINISH—Here are a few examples of


The manual process for truck routing was prohibiting growth.

Dispatchers relied on a paper ticketing system to pick-up donations, which made it impossible to do anything efficiently. So we helped them by transforming their manual paper system into a multi-platform web solution, enabling them to triple their annual number of donation pick-ups.


We began by interviewing The Salvation Army officers and dispatchers to understand how donation pick-ups happen. We outlined the most common workflows, most of which followed the same general sequence of events: someone wants to donate clothes or furniture to The Salvation Army, they call their local center, the call center operator schedules their pick-up and creates a paper ticket, and then the paper tickets are sorted into piles by dates. Next, those tickets are distributed evenly to drivers based on general location, and then drivers would make their pick-ups and deliver the donations to a sorting warehouse.


We discovered that this tedious manual process not only limited the number of daily pick-ups, it also kept The Salvation Army from being able to monitor driver progress and to gain insight to analytics for improvements.


In order to help the dispatchers come up with more efficient routes for their daily pickups, we needed that process to be automated.  So we created an algorithm to solve what is essentially a unique version of the Traveling Salesman’s Problem (TSP). By incorporating some additional factors (such as the physical size of each pickup, and the time it takes to carry the items from a building and load onto the truck), we were able to create a fairly reliable prediction for how long each pickup, and each route would take. Over the course of the next few months of beta implementation, we gathered data on each pickup, and were then able to adjust the variables to make each prediction more and more accurate. Very quickly, it became clear that by implementing this “smart routing” algorithm, 3-4 times as many more pickups could be completed by each driver, each day.


The next obstacles were to give the dispatchers the ability to make manual adjustments to the automatic routes created by the algorithm and to get drivers the information they needed to make their deliveries. Learn more about the dispatch application we created here and the iPad app for the drivers here.

Here’s how we did it:

Devnext, Inc. © 2016  |  PO Box 5447  Canton GA, 30114