Planning Optimization

Transforming a key operation with applied maths and a new UX design

Context

Our client is a world-wide freight-forwarder, arranging every day all kind of freight transportation for multiple clients, to hundreds of destinations in the world. In freight transportation, dealing with more shipments leads to greater economies of scale, thanks to easier space bookings with air and ocean carriers, and the possibility to negociate lower prices. However sources of economies could also be found in increased operational efficencies. For our client, this efficiency is obtained through consolidation, the practice of combining shipments from multiple consignees into one single shipment to the same destination. Consolidation permits a mix of dense and light shipments together, and thus reduce the cost of air freight, partly based on the density of shipment but also based on the number of operations the air carrier will have to perform to arrange shipments in the aircraft (less handling operations means less cost). Much more than just assembling Lego, or heterogenous blocks like in Tetris, consolidation is quite a complex task for a human operator because of the multiple dimensions to take into account: weight, dimensions, density, nature of goods, rotatability, time constraints, amount of shipments and possible destinations to handle, etc. This challenge appealed for a decision making tool, that could help operators to take the right decisions to optimize the cost of operations. Thanks to our deep expertise in data science and system engineering, our client decided to call us to help them solve this problem.

Shipment Grouping in Air Cargo

What did we do ?

To perform smart and cost efficient consolidation, the idea is to develop an application providing recommendations to hub agents on shipments consolidation, ie. shipments that must be sent together on the same flight. This application would use an algorithm to automatically find the best shipments mix possible and assign a flight for each shipment.

Solution

Transforming a decade-long process thanks to a new tool

The decision making tool we would built should help our client to make a considerable shift in its process.

Before our intervention, shipments were assigned to a flight, and consolidation of shipments could only be made within the same flight, on the same day. It took some time to be accepted, but by having a nice tool that allow quick decision taking, we could completely change the logic of freight forwarding. By taking into account all shipments that must shipped to a destination within a week, rather than in a day, we allowed our client to have much more shipments to consolidate to find the best suitable consolidation, optimize both the shipping and handling costs and have more efficient operations.

Hub agents are being augmented with a tool, and are now in charge of deciding when to send shipment, on which flight, taking into account other shipments for the coming days.

Form push to pull Form push to pull We tackled this problem with a double approach, relying on our UX and Applied Mathematics expertises. The first one would be needed to build a tool that fits smoothly within the agent workflow, the latter would serve to develop the core feature behind the wall, the intelligence of consolidation.

User Experience Approach

The motto of User Experience is to focus on users’ needs. This project was special in the sense that users were really enthusiastic about the value proposition, but they did not know what should be their expectation in terms of how the tool would fit in their daily work. Indeed, we would bring a great innovation into their current job, but they did not know to what extent it would change their work.

By building this new application, we were also building a new way of working (i.e new process) for all operators on the air freight logistic. Adopting a disciplined UX approach was mandatory to ensure the good fit between what our client and us imagined, and users’ needs and habits. In other words we deployed a user-centred strategy to ensure the success of our decision making tool and its quick adoption in the current operational environment.

After an initial draw of a user flow and of the application structure, we built the application wireframes in successive sprints. We designed interactive tests with users in order to both test the usability of our application and collect the users’ feedback on the application and the project itself.

Mathematical Approach

The first step was to mathematically formalize our client’s problem. Our objective is to find shipments’ allocation on booked flight capacities in order to minimize the cost while respecting various constraints: physical constraints, operational constraints, customer constraints… Uncertainty on shipments arriving the following days must also be taken into consideration.

To do so, we had to deeply understand our client’s business to list and prioritize the constraints to be considered, and suggest a formalization. Then, using various mathematical tricks, we transformed this problem into a specific Optimization problem that we were able to solve using various Operational Reseach tools.

This mixed approach combining UX and Maths was a truly innovative process and key for the project success. Indeed, building UX for a complex and smart decision-making tool requires a deep understanding of the recommendation and optimization possibilities, and a critical thinking on what we could imagine and realize. Similarly, feedback from users interview and testing was key to validate the relevance of our business understanding and our modelization of the problem.

Results

Thanks to our expertise in both Applied Mathematics and User Experience, we were able to develop a web-application integrated to our client’s operational chain enabling optimal shipments assignments to air carrier capacities. This project leads directly to major benefits:

  • A tool enabling a deep and sustainable transformation of operations at scale
  • Increased productivity with empowered operators
  • Financial savings thanks to cost optimisations