Price elastic planning for Retail

I have a retail customer they are interested in seeting "Price elasticity" based forecasting calculation for their products, we did calculate a correlation coefficient to identify the products which are price elastic, as a next step they want to use the model to predict a possible price-sales function and use it in their forecasting. Have we done this use case elsewhere, any pointers will be helpful.

Comments

  • @ArunManickam 

    Nice question and a tough one at that!

    Price elasticity, the actual coefficients, are generally calculated outside of Anaplan using some very sophisticated statistical models. Be careful though, retail is finicky because the price change often  doesn't impact just the item you're evaluating. It can impact similarly offered items and there's this concept of a "halo" effect, not to mention what happens if you stockout. There are some who believe the average number of dollars per transaction (or average value per transaction) is relatively fixed. Meaning, by lowering the price all you are doing is encouraging customers to spend on markdowns thereby giving away margin. Yes, the sales increase on the item whose price you lowered, but the impact on other items is that their sales dropped. Once you have the coefficients, though, they are very easy to implement in Anaplan. Here is a typical calculation I use: [(1+change in price)^Elasticity]-1

     

    elasticity001.png

    Unless you plan to use a Alteryx or Knime, I would recommend a simpler approach to forecasting the impact of price changes. Use a "rule of thumb", meaning a 30% reduction will give a lift of 1.5X using historical promotions as a guide. I know this may sound ridiculous to a top-analyst but I've found it's just as accurate as going through all the data science. It's more intuitive, it's "explainable", and much simpler. I wouldn't underestimate the explainability of your model.Most retailers I've worked with prefer a more intuitive understanding of the forecast recommendations.

    Great topic @ArunManickam - let's keep this conversation going!

  • Thanks @JaredDolich . This is useful.

     

    Lets keep the conversation on, I also like to be simplistic on this. Customer is not willing to invest more on newer AI/ML technology at the moment. Great approach you have shown, let me try to apply this for few SKUs and circle back with customer.

     

    Thanks

    Arun

Categories