Using revenue forecasting to know when to introduce new products and promotions
Combining historical data with statistical models, such as Linear Regression and Holt-Winters, to help operators to identify the ideal moment to release something new.
26th Aug 2016
Changing the past is only possible in science fiction movies. But what if we can forecast the future and make a decision right now to make it better? With Passenger, transport providers will soon be able to explore key performance indicators such as revenue and predict how they are going to evolve in the near future.
Combining historical data with statistical models, such as Linear Regression and Holt-Winters, we can create an estimation for an upcoming time period. The amount and quality of data is crucial to building reliable models. That is especially relevant in this type of time series where seasonality plays an important role (e.g. Easter and Christmas).
One of the main opportunities of using revenue forecasting in tools like Passenger is to identify the best moment to launch a promotion or modify an existing ticket offering. In the following picture we can see a simulation of weekly m-ticketing revenue for a bus operator. The forecasting on this demo dataset was indicating a considerable revenue decrease. Having this insight, an operator can decide to launch a promotion with discounted tickets and a good marketing campaign. This strategic combination of pricing and increased level of service to the customer (e.g. by adding new mobile app features) has resulted in an exponential sales growth as shown in the plot.
Revenue forecasting is great, but it’s not a magic wand since there are many other side variables that cannot be controlled (e.g. competitor response and unexpected events). It can help to make more informed decisions, but an effective marketing strategy is also key to achieve success.
To find out more about the work we’re doing using mobile apps to give insight to transport authorities and mobility providers, please get in touch.