From theory to practice: the restraints on the use of route optimization
Last-mile logistics is known to be the most complex and expensive in the whole supply chain. In recent years, the boom in e-commerce and the proliferation of points of sale in urban centers have reinforced the trend, increasing the volumes of packages to be delivered. In addition, the proliferation of delivery services makes consumers (and professionals) more and more demanding and less and less inclined to pay for delivery.
Customers tend to react negatively to unpredictable or large time slots. They feel even more frustrated when those slots are not respected, No matter the context : receiving a package, a delivery, welcoming a technician or a health professional.
These underlying trends make the last kilometer more and more complex to manage for organizations which now need to be able to plan it very carefully.
Route optimization is the relevant answer ! Yes, but…
Faced with this problem, optimizing routes is a relevant response : by allocating the right resource in the right place and at the right time, it allows us to better respect the commitments made to customers (especially in terms of time slots), reduce costs (fewer vehicles used, fewer kilometers traveled) and keep environmental impact under control.
Nevertheless, in practice few actors are equipped with such a solution. It is estimated that only 20% are equipped among the actors with this need. Among those equipped, it is also very common to observe an absence of use of the solution or a partial use, supplemented by a large number of manual corrections and replanning.
Where does this paradox come from and what are the obstacles to using route optimization solutions?
The inevitable question of data quality
Like many technological solutions for supply chain optimization, route optimization is mainly based on data. Addresses of points to be visited, size and weight of packages, time constraints of the sites to be delivered: the data define the problem and their quality depends on the relevance of the solution that will be delivered.
In some organizations, there is also information known to operational staff but which does not exist in any database: this is sometimes the case for the skills of technicians for companies requiring the planning of technical interventions, provider / client incompatibilities or still the tacit agreements which are the fruit of common habits.
Consequence: many players who could optimize their routes do not do so because they are not able to provide the solution with sufficiently thorough and reliable data as input.
Logistics never stops and information continuously arrives in the system until the last minute. The prospect of having to restart an optimization after each change (or manage them manually) can discourage potential users.
One solution to this problem is the digitalization of logistics organizations and “data projects” are the main preoccupation. But it is also possible to equip yourself with intelligent technologies, capable of operating on imperfect and / or incomplete data and to take into account corrections on the fly without calling into question the entire optimization process already underway.
The latest intelligent geocoding solutions use technologies capable of working on unstructured data via the use of Deep Learning, significantly improving the quality of the results provided to route optimization solutions.
Taking into account specific constraints, these little details that change everything
Many activities must respond to very specific constraints including these 4 examples :
- furniture delivery requires 2 delivery people when the delivery exceeds a certain weight,
certain interventions require the passage of technicians in a PUDO center (Pick-Up Drop Off) in order to take material,
- the time spent by a technician at a customer can range from simple to quadruple depending on the technical complexity of the intervention,
- in bulk transport, some collection sites can only accommodate one vehicle at a time
If these constraints are not taken into account when planning the routes, the solutions provided will neither be suitable nor realistic, forcing the user to complete the planning obtained with a large number of manual adjustments, making use an even more complex and time consuming tool.
The ability to take into account very specific constraints must therefore be a fundamental criterion when choosing a route optimization solution.
Real time widens the gap between theory and practice
Traffic jam, absent customer, wrong address, difficulty in parking: the last mile is full of pitfalls. While a large majority of route optimization solutions work in static (route calculations then cut-off), the last mile is dynamic and requires absolute reactivity!
How do you make sure that a traffic jam at the start of the day does not impact a slightly tight time slot planned later? How to rebalance the tours of technicians in real time if the schedule of one of them has been lightened?
The significant differences that companies observe between the performance of planned routes and those actually carried out generate disappointment regarding route optimization tools. In the context of urban logistics in particular, it is becoming increasingly futile to calculate static routes.
To mitigate risks, the route optimization solution must be able to adapt instantly to changes when they occur in order to provide operational staff with the best possible routes, whatever happens in the field.
Redefining your organization requires change management
This is particularly the case if the solution adds constraints instead of eliminating them or if it does not respond well to the reality in the field (for lack of flexibility or because of a poor definition of the need upstream).
The right approach is to identify and involve all stakeholders from the start of the project. Site manager, planners, field teams, all of them have business knowledge that has high added value during the stages of building the solution with the publisher.
The implementation of a route optimization solution is a major project that will bring real change and require a significant investment within the company. It is therefore crucial to go prepared, with sufficient mastery of your data and the will to drive this change with the teams. However, these projects are also new and it is important to be accompanied by an editor who is also an expert in logistical issues, who will be able to anticipate and solve the technical and organizational problems that are sure to arise and which offers recent technology. , in line with the real time needs of the market.