Reading Notes: “How Business Science Revealed Hidden Pricing Opportunities During Covid-19”
This is a commentary on the article by Fabrizio Fantini found here. The article was published to Towards Data Science, a popular data science blog.
There are many ways in which to measure and adjust pricing. This article outlines a model for measuring and adjusting the price of products within a given product category, and then shows how price changes across a category can be made without the consumer noticing, and ways to convince the consumer that the price has been changed without actually making a change.
The model, which the author refers to as the “4 and 9 Pricing Framework” is based on the view that the average price for a product group is controlled by four factors. Changing any one of these 4 factors — list price, range of products, promotion intensity, or promotion depth — can be used to adjust the average price of the product group. Changes to any one of the 4 factors can be visualized through a set of 9 graphs: Average List Price, Average list price for a fixed set of products, change in product mix, average percent discount, percent of items discounted, average percent off when discounted, average paid price, average paid price for a fixed item mix, and change in item mix price. It’s worth noting that this strategy mostly only applies to retail pricing, and works best when applied to a set of similar products offered by a single company.
The most straightforward method of adjusting price is changing the list price of an item or items in the group. This is most clearly identified when you look at the list price for a fixed mix of items, any change in this graph will always be caused by a list price change. It might make sense at first to use the regular fixed price graph, but this can be misleading because changes to the product mix can have a significant impact even if all individual list prices remain the same.
The second method of adjusting the price of a product group is to change the product mix. By taking out less expensive products and adding in pricier ones, you raise the average price without adjusting the price of any one particular product. The author shows how this caused nearly 20% increase in the average price of a bottle of wine in the UK during the first weeks of the COVID-19 pandemic, as cheaper wines sold out. Since this increase was measured by the straight average of a set of different wine bottles, and did not take into account which ones were bought more frequently, it is likely that the real difference is less than 20%, but the change is likely still there to some degree.
The final 2 methods of adjusting price center around discounts or sales. One is to make steeper discounts, say, changing the average discount from 20% to 30%. The other is to change the number of items in the group that are on sale at the same time. I think that last method is particularly clever. If you rotate different items through sales regularly, it wouldn’t be obvious that the product group has a lower average price. Even if you look at a few different items it would be hard to see, but the result is still the same.
While this article is interesting and provides some useful tricks, the model it describes doesn’t take sales for different items in the group into account. I see this as a critical oversight, since some of the methods of adjusting average price become far less effective when sales are taken into account. For example, if you want to lower the average price of a bottle of wine, you could increase the number of bottles on sale or the increase the discount on the ones which are already on sale. Assuming inventory is sufficient, the steeper discount method might lead to significant increases in sales for those bottles, but if you put more bottles on the same 10% sale that other bottles are already on, customers who were already planning to buy a discounted bottle will just have more of a selection, and it’s much less effective.
