A few weeks ago, a press release from PostcodeAnywhere caught my attention. The company introduced a new routing solution for private fleets that uses real-world data from tens of thousands of customers driving with TomTom devices in the UK. This is the first routing solution that I am aware of that uses “real-world” average speed data.
Standard routing solutions input data—such as the number of trucks in a fleet, truck characteristics (e.g., truck height, weight capacity, cube-out capacity), and customer delivery points—into complex optimization algorithms to determine the “best” route plan for the fleet in terms of minimizing the miles traveled or the time it will take to complete the deliveries. Existing routing solutions are excellent at minimizing distance, but they are less effective at minimizing delivery time because they don’t contain “real-world” data—i.e., they lack data on the true average speed per hour traveled on particular sections of road, at different times of the day and week.
Some routing solutions allow users to modify route times. For example, if you know that delivering to a particular location, via a particular route, at a particular time of day will take longer than usual (such as rush hour on Tuesday mornings), you can modify the time parameters in the system. But this is a manual and time-consuming process. In contrast, PostcodeAnywhere’s software-as-a-service routing solution bases its fleet plans on road statistics, which include average speeds across five minute intervals, on any day of the week. The data on average speed on a section of road by time of day is updated quarterly.
I spoke to the company’s managing director (MD) and the lead developer of this product and learned that this solution is extremely affordable. The MD mentioned a payback period of months. I believe it based on the company’s pricing, which is a bit difficult to explain. The managing director also mentioned that there is good TomTom data available for North America. PostcodeAnywhere is looking for reseller partners in North America.
So, what is next in route planning? The PostcodeAnywhere average speed data is updated quarterly. Ideally, we will see this data updated more often. Further, this data is not just available from TomTom and other GPS device providers, but also from phone companies that are increasingly selling GPS-enabled cell phones. Therefore, I don’t expect PostcodeAnywhere to be the sole provider of this type of solution for very long.
There are also solution providers like Descartes Systems Group (an ARC client) that offer real-time route planning. If a truck gets stuck in traffic, a user can rerun the whole fleet plan for the day taking into account what is happening on the ground. Even better, however, will be the day when these solutions use real-time congestion data to re-plan routes and prevent trucks from getting stuck in traffic in the first place.
I absolutely agree about the importance of improving “time modelling” in vehicle routing software. And the more accurate the travel speeds, the better.
But if you analyse how vehicles operating in a multi-drop environment actually spend their time you will find many of them are over 50% engaged in loading or unloading the vehicles. And while most people have a reasonable idea of road speeds and can get a reasonable approximation to that, even if not as accurate as the example you give, they can be way off in their estimates of loading or unloading times.
The goods you have loaded on the vehicle, the packaging and containerisation plus the unloading equipment (roll-cages, pallets, pallet pump trucks, tail-lifts, loading docks) and just simple things like the walk distance to make a delivery (for example to small shops) all make a very big difference to how fast you can offload.
If you deliver exactly the same quantities each time to the same places, then you can simply record how long it took last time. But if you deliver different quantities and different combinations of products you need a “model” to work out how it is going to take. And it should look at those factors.
This has long been a “blind spot” in the use of vehicle routing – mainly because of the difficult of obtaining the raw data. But now so many companies are shifting to Proof-of-delivery systems and other telematics the data is becoming available.
So “real world data” is a lot more than just improved road speeds. Having your system better calculate how long you spend at customers is as great if not a greater prize. But to do so you need more than just a weight and a cube – you need products, containers, vehicle and site capabilities – the real world of your delivery operation.
Tim Pigden
Optrak.com Vehicle Routing Software