Abstract:
We develop an optimization model for dynamic pricing in a coffe shop that seeks to maximize total contribution. Based on the random utility maximization theory, the preference-based demand function is derived from choice data using sequential processes of estimating individual utilities and simulating and aggregating individual choices. Individual utilities are estimated using the hierarchical Bayes method, while individual choices are estimated using the randomized first choice simulation. We implemented the approach in a coffee shop in Jakarta, Indonesia, and considered only one product i.e. milk coffee, which constituted two-third of total sales. To avoid complexity in the implementation, we considered two fare-classes and developed three time-based pricing scenarios. The solution to the mixed-integer nonlinear programming problem was obtained using enumeration. We came up with optimal prices for a cup of milk coffee of Rp15,000 in the morning, and Rp23,000 in other time of day. Using this pricing policy, the monthly total contribution was estimated to increase by 11%, from Rp71.9 millions to Rp79.8 millions.