Comparing Benchmarking Measures for Financial Data

In this project, the team will be asked to compare several methods of benchmarking individuals to peers using financial data. The team should create benchmarks using the following 3 methods and contrast the results:

1. Mean Analysis – compare a franchisee’s financials against the mean cost (over time).
2. Regression – regress average daily cost on revenue, and benchmark stores based on the slope coefficient (a proxy for variable cost).
3. Data Envelopment Analysis (DEA) – apply DEA to establish an efficiency frontier, comparing how different types of costs related to revenue.

A successful project will contrast the outcome of the state benchmarking methodologies, as well as identify the strengths and weaknesses of each. The result should identify the ideal method (based on the data they have been given). Some key elements of a successful project will be to answer the following:

• How is each method affected by outliers or extreme observations? How did you choose to both identify and address them?
• Is there an ideal benchmarking "period" or sample size (1 month, 3 months, trailing 12, YTD)? Should we annualize data?

One of our current solutions focuses on gathering and aggregating financial reports from various restaurant franchisees. A key element of this tool is to benchmark an individual’s current financial state against their peers. A better understanding of various benchmarking methodologies will enhance the user experience, allowing them to gain better insight through more appropriate comparison groups.

Make sure to install all the Python dependencies that are used in the notebooks as follows:

python -m venv venv
source venv/bin/activate # source venv/Scripts/activate if using Windows
pip install -r requirements.txt

Also ensure that the Jupyter notebook can access the R kernel by installing the IRkernel R package as follows:

R # to run the R console
> install.packages('IRkernel')
> IRkernel::installspec()