As the busiest period of the year for seasonal shopping begins, freight transport and logistics professionals must be fully prepared to anticipate and adapt to sudden spikes in demand, even amid challenging market conditions. Ontruck, Europe’s leading digital road freight platform, has developed a predictive tool that leverages Artificial Intelligence and machine learning to predict business behaviour with an average margin of error of 10% in demand and 16% in load capacity by vehicle type.
“These reliable and consistent predictions are the result of a machine learning system we have designed based on a predictive model that learns from constant iterations of data” explains Javier Escribano, Ontruck’s Chief Product Officer. “The flow of information that we receive is a combination of unique data that the predictive model is provided with and data that the model also collects automatically across company information systems. We have even programmed the model to learn from its own prediction deviations.”
Intrinsic to the success of these forecasts is the ability of the model to account for a complex range of real-time variables. These variables include the history of business loads, customers who placed orders, their sectors of activity, the recipients, origin and geographical destinations of the loads, the type of vehicle, and even the exact pallets used. With this data, the model can effectively analyse seasonality and the impact of key shopping holidays such as Christmas and Black Friday and how they affect specific locations. The Ontruck team also provides the system with official economic data sourced from the Office of National Statistics, such as percentages of businesses that open or close in a city or country, or extraordinary macroeconomic events, (limitations on mobility, a strike, a border closure), which are entered manually.
On top of seasonal peaks, transporters face the added pressure of working within the dynamic conditions created by COVID-19, which has showcased the increasing importance for flexibility, agility and organisational planning in resilient supply chains. The fast-changing developments of the pandemic are factored into the prediction tool analysis, meaning that Ontruck can evaluate and best prepare for any potential impacts of the COVID-19 outbreak on business. “When China closed its borders in February, we made simulations at Ontruck to try understand how a similar situation in Europe would affect our business over a course of two months” explains Escribano, “the predictions of the model were accurate, with only a 20% margin of error, so with this tool we can always stay one step ahead”.
Ontruck’s predictive model is split into two phases: the first offers a forecast report of up to two months in advance. Following this, a fortnightly review is conducted to take into account any updates that may have arisen. Using this closely audited approach, it is possible to make adjustments to account for any unexpected situations which may arise, such as a new confinement order or border closure, to allow for the most accurate predictions possible.
“This model can provide us with invaluable information and visibility regarding our business growth, equipping us with critical market data that can power our day-to-day decision making” explains Escribano. “We know the exact number and type of vehicles that are needed in each city, how many orders we are going to receive, and the potential revenues these can generate for the business. Understanding this load estimation is game-changing for planning our resources and provides us with a full overview of any extended capacity to increase sales volume ”, says the Ontruck CPO.
