Food Delivery Cost Optimization Using Machine Learning
Customer: The Food Company
Location: Us-West1,San Francisco, CA
Partner: Pluto7 Consulting Inc.
Customers Specific Challenges pre-Google Solution:
The company had challenges finding routes for delivery efficiently. This was previously done manually which is very time consuming. The company was unsure if their current system led to optimal routes. The company wanted to find the optimal route in terms of “cost per box.”
Pluto7 educated the customer on GCP capabilities, ML concepts, ML feature engineering and incorporating key metrics associated per delivery. Pluto7 used a recommendation engine which suggests the route that can deliver new orders based on their prediction call to the ML model. The cost optimizer module outputs the cheapest of all the recommended routes for the customer.
GCP Components Used:
App Engine, Compute Engine, Google Cloud Storage, BigQuery, Cloud ML Engine.
Customer Acknowledgement of the Work Done:
Leaders from Food Company were pleased with the results of the proof of concept and will use the ML model everyday to make delivery decisions. The company sees a lot of potential in the route recommendation model. They understand the efforts required to do feature engineering and will work towards that. Data captured will be their key focus so that the model will be effective as they evolve to leverage machine learning as per the recommendation documentation that Pluto7 has shared with the customer.
Benefits Customer Has Seen so Far:
The model demonstrates recommends routes that are up to 100 percent cheaper than the routes historically used by the business for some postal codes. The ML model development process was instrumental in identifying strategies that could be utilized to save costs for the company and make the landscape more competitive. The power of BigQuery was leveraged by Pluto7 to derive the updated costing sheet. This was crucial to identify the cost levied on the business by its vendors. The CEO of the company and his team are aware of that possibilities of GCP and ML as they transformed their business.
Click here for our use cases!