Paper Plane Updates - Shelf Price

Assumptions

One of the major changes that has been done is the addition of clarification points between the two existing models. They both use data, and the misalignments in the way they were used before has been removed. Now they are using data in exactly the same way, so you define one date for both models, and then you define how it works.

For example, one of the assumptions that are used now is that the models work on a weekly forecast, and Monday is set to be the first day of the week. This delimitation was not there before and that was and aspect that could lead to certain misalignment between the two. Now it is well-aligned helping you avoid mistakes.

Data sets

Errors that would occur between the different data sets have also been addressed and fixed in the Paper Plane update. For example, we have the transaction data and the stock data, which are typically from different sources and may be updated at different intervals.

The same goes for competition prices, which may be updated at varying frequencies. In the past, these dates and intervals were not always aligned, resulting in error messages. However, in this release we improved the error handling and alignment between the data sets.

In the new release, the models communicate with each other to determine the number of periods you can predict. For example, if you select another model that only fetches data for the next two weeks, you will not be able to make a mistake because it directly fetches the relevant data for that model.

Optimization engine

One of the main changes is that we have improved the way we prioritize optimization for revenue or profits.

The optimization engine now works in two steps.

The first step is to simulate what would happen if you don't change anything but maintain the same assumptions. This gives us a clean starting point to work with.

We position the objectives based on the simulation results and then run the optimization. It's a two-step process that happens behind the scenes, but it helps us balance revenue and profits better. There haven't been many changes to the parameters we define, so we can still set price increase/decrease thresholds, boundaries for competitor prices, and stock coverage targets based on the type of product. The channel reference and average difference also remained the same.

Dashboard 

We still have the same impact dashboard with charts, but we have added additional metrics to make it easier for users to review and check the outputs.

Now, the release includes the threshold that you define, as well as the recommended retail price, the difference between the discount and the share price, and supplier cost.