Hey Team! I am trying to compile a ranking and prioritization of retail scenario planning use cases. What are you hearing out in the marketplace (or inside you company)?
I can't resist the opportunity to weigh in on this question.
Besides What-If modeling, there's a very specific analysis in Retail called the Efficient Frontier for Assortment Planning. It's a scatterplot that has margin capture rate on the Y-Axis and Inventory Levels on the X-Axis. The most efficient, or optimized scenario would be one that resides on the boundary of the plots. Here's a visual example from Farhad Malik and here is a link to his brilliant article about portfolio management (plotting risk and reward, instead). Farhad provides the math in his article as well as the Python code if you prefer to run the scenarios outside of Anaplan. The best portfolios are the ones that are close to the black line.
For the retail example, you would replace the Return with Gross Margin capture and replace risk with Inventory Levels. I might also point out, a very realistic way to build this in Anaplan would be to take Last Year's items (SKUs or Style-Colors) and plot those instead of an entire assortment. As you go up the product hierarchy you can rescatter the chart to see if there are any changes to the frontier as you reselect different assortments.
Following to stay in the loop! So curious how retailers are using Anaplan for their end to end planning process!
One area in Assortment planning that I've seen a lot of attention recently is pareto analysis - the "80/20" rule to make sure when buying this year's assortment, we don't under buy the top-sellers and avoid buying into markdown for the outliers. Also helps make decisions on subset assortments.
Brilliant! So often I see charts that look like this:
When, in reality, they look more like this:
Agree completely – with SKU rationalization there must be an analysis of demand transference (in Anaplan?), which could be used to identify the best “blue” bars and provide a balanced assortment.
What I saw pareto being used for was slightly different. When a buyer had agreed the new option count for a season/floorset based on volume and space and wanted to spread the average line buy across the assortment by ranking options (using buyer insight) and then a pareto analysis to suggest the volume for each one. Before using the analytics, we typically saw the buyers use more of a flat curve. With the analytics it gave them the confidence to really chase their best sellers, whilst ensuring they didn’t over stock the tail.