Papa John's uses geodata to develop its pizza chain and reduce delivery times
We share with you the experience of our projects and respect the NDA
Objective:
Reduce the ROI period when opening new points and reduce order delivery times
Context:
The chain has reached the natural growth limit in a metropolitan area and has also minimized delivery times through properly aligned logistics and automated order-taking. New tools are needed to improve results.
Key indicators:
Over 50 points. ROI period reduction: for 98% of new points ROI proceeds as planned Delivery time reduction: all points have reached the target delivery time of 15 min.
Solution:
The project is implemented in a European metropolitan area. The local pizzeria chain, which at the pilot project stage included several districts of the city, had set two tasks for themselves and the GML team. The first was completing an in-depth data analysis to manage the chain in a way that improves its overall efficiency. The second involved taking the delivery time factor into account when solving the first task. At no point in the chain should the delivery time exceed 15 minutes, including various modes of transport: bicycle, kick scooter, car, or walking.
To solve the first problem, we considered the kind of data we needed. We used customer data from which personal information was removed: where they live, what they order, how often, how satisfied they are with the service, etc. We also took the figures for each point: when it opened, payback period, ROI period, data on sales volume, product types, order frequency, delivery time, and user ratings. The remaining data is geodata, socio-demographic information, and data from telecom operators.
We combined all the information about residents, neighbourhoods, traffic, property values, competitors, transportation, advertising opportunities, and infrastructure with company data and used AI and our geo services to identify the most important factors and their impact on business success and displayed them on heat maps.
Working with a chain is a long process. After assessing all the factors and evaluating the chain points, together with the customer we reviewed the category that each point falls into: a successful point in a prime location - 1, a potentially successful point in a not-so-great location - 2, an unsuccessful point in a good location - 3, and an unsuccessful point in a bad location - 4.
We immediately discarded number 4 because there is no point in wasting time on it. Number 1 helps number 3 by sharing their experience, and for number 2 we use recommendations based on data analysis to look for suitable locations and select the best location. All actions are subject to a limit of 15 minutes to reach customers, and points should not cannibalize each other within the chain. This algorithm is essentially the basis of the project that we have carried out. It has a number of fascinating details, for example, which types of data unexpectedly turn out to be important and which do not, what features can be obtained by combining customer profiles with geodata, etc., although this is a different case format.
Result:
The project has proven to be very successful and has been expanding geographically. Almost all of the new locations that opened using AI recommendations based on data analysis returned the investment on time. The rate of 98% is virtually a 100% success rate.
At the moment, all the chain points that participated in the project deliver orders in under 15 minutes. Geoanalysis was used to move and relocate some of them, to close some of them, and to open new ones. Now it is an optimal chain that completes the tasks at hand.
© 2019 - 2024 Global ML
