Clash of titans: Failed deliveries vs suboptimal routes!
Sustainability has been a topic discussed for a while now, institutionalized actions such as the ones taken by the United Nations in 2015 in the form of the UN Sustainable Development Goals help as guidelines for companies and governments to achieve its 17 goals for 2030.
E-commerce has been evolving for many years, and swiftly increased its share from the COVID-19 pandemic. From urbanization effects in the last years, e-commerce has it is a larger effect within the last mile.
Delivery costs and environmental effects, which increase along with restrictions (e.g., narrow time window deliveries) in its Vehicle Routing Problem (VRP TW) solutions, as well as with a large number of failed deliveries, lead to analyze its inherent trade-offs. This was made with the support of a database with household coordinates (X, Y) of Cancun City in Mexico, an Opensource VRP Spreadsheet Solver (Erdogan, 2017), and quantitative analysis developing 90 scenarios with combinations of database sets with a different number of customers, “narrow” time window width (NTW) and percentage of customers served by these NTWs. Scores computed for compensating values of failed deliveries resulted in 1425 cases and support for a sensitivity analysis applied to the main scenarios.
The focus relayed on assessing, how economical and sustainable KPIs (CO2 and PM2.5) behaved in a Vehicle Routing Problem to reduce the failed deliveries through the usage of time windows. Findings showed that even significant reductions in failed delivery costs did not compensate for the increase in the cost of guiding time window deliveries. Furthermore, there is not a clear advantage of trying to lead time window deliveries for avoiding delivery failures by using a Spreadsheet Solver for VRP in the last mile in an e-commerce context.
Lastly, sustainability KPIs were differently impacted by each scenario. CO2 represented 8% and PM2.5 emissions 4% of the total costs in monetary values. The direct costs (fixed and variable) associated with the distance traveled were the main driver adding 88% of the total costs.
