A guide to auto-dispatch. Part 0: Why should I care?
By this article, we open a series aiming to explain in detail dispatching, its impact on the overall last-mile efficiency, and what one should consider while automating dispatching.
Dispatch is the process of choosing a driver (or rider, or mobile robot) who will complete specific delivery, assigning, sending them off and tracking the progress. In modern high-pace logistics, dispatching becomes one of the most critical things because its quality defines both the level of service and delivery costs.
To avoid errors common to repetitive processes and master operational efficiency, companies automate their dispatching. However, “finding the right driver” is not a simple problem one could solve using straightforward approaches.
What is dispatching?
Dispatching means sending something or someone to complete some work somewhere.
In logistics, dispatching is a complex process that requires not only an assignment of a driver to specific delivery but also:
to find and choose the best driver
to confirm delivery with them
to track the delivery and resolve potential issues
So dispatching includes everything that needs to be managed since the order has been accepted. Depending on the nature of the business, it might also include managing order picking, meal preparation, etc. But steps associated with the delivery generally remain the same.
Dispatching could be manual, semi-automated (with different supporting tools like route planning and fleet tracking & management systems), or completely automated.
Depending on the level of visibility and simplicity of communication with drivers and customers, a dispatcher-a-human could manage 30-100 deliveries per day. While for a dispatcher-a-software, there are practically no such limits.
Good vs Poor dispatch
Good automated dispatching allows a company to manage deliveries efficiently and pack the fleet tight while ensuring on-time deliveries. On the other hand, poor dispatching leads to many late deliveries and an underused fleet.
What defines good dispatching? Simply put, it’s a fast, adequate and optimal reaction to what’s going on (such as a new incoming order, cancellation or a vehicle break).
But it’s easier to say than do. People get tired from repetitive tasks and easily make mistakes (and we shouldn't blame them!)
Assigning an order to a driver who is far from the pickup location
Assigning an order to the closest driver, but their work shift ends in 15 mins
Sending off all your drivers to complete far deliveries just before the peak-hours and as a result, having a lot more delays, etc
But we shouldn’t blame them! At early days of Optimate, I’ve took an operator job for a few days with our first customer - it was freaking exhausting and hard! I made so many mistakes. At the same time, straightforward automation suffers from edge-cases and often leads to "stupid decisions" human operators will never make.
As an example of the latter, I witnessed the following situation in rapid grocery delivery several times. The customer placed an order, forgot to add something to their cart, then put a second order with a single item (thanks to the free delivery option). And the dispatching algorithm decided to... send off two(!) different riders from the same dark store to the same delivery location.
And if you think there are not so many corner cases like this one, I have bad news - there are plenty of them. Dispatching automation is not an easy task.
Urgency in dispatch automation
Companies have been investing in partial automation of the dispatching all the time.
But recently, the demand for automation has doubled. Because, it became an essential part of delivery management.
In modern logistics, especially in the same- and next-day delivery, there’s a little company can do in advance. In a dynamic last-mile world, you can’t schedule everything and follow it then.
Companies are forced to react to what’s happening now - new orders, transport breaks, traffic and weather conditions. So dispatching becomes the cornerstone of delivery management. And the quality of dispatching defines the overall last-mile operational efficiency.
Poor dispatch to both delays and low fleet utilization. And as logistics scales in volume and/or speed - things become worse. Unsatisfied customers vote with their feet. Under-paid drivers also leave. “Thanks” to the labor shortage - drivers are scarce as hell, and there’s no problem finding a new job.
The only way to overcome that is to automate your dispatch intelligently so the automation can leverage the scale and complexity in your favor.
By dispatching automation you can:
Assess whether incoming order is deliverable within required time and SLA. By giving your customers a realistic expectation upon confirmation, you build a trust. Even if you committed to 15-min delivery, but overbooked right now, promising your customer 25-min ETA and fulfilling it is the way better than promise 15-min ETA and get late by 10 min.
Better utilize your fleet. By automatically assessing different drivers, routes and options, you could find the best one who will complete the delivery faster and/or cheaper. The tighter you pack your fleet, the lower cost-per-delivery you have. Some advance methods like dynamic order batching and predictive fleet repositioning allow you to achieve the lowest possible costs. We’ll discuss them in our blog later.
But it’s practically impossible to do such things manually.
Automatic dispatching is not a silver bullet; it won’t help your company upgrade “unprofitable by design” last-mile to a high margin one.
But, it could help you to cut-off 20-30% of delivery costs by loading your fleet tight while increasing your on-time KPIs to the “normal” rate of 95-97%.
In most cases, it would be enough and let your business thrive.
In the following posts, I’ll go through the main steps of the dispatching process, explaining approaches to its automation, common pitfalls and lessons we learned during our journey in building Optimate.