Monday, November 5, 2018

To Analyze a Sales Population, Consider the Expanded Percentile Curve

(Click on the image to enlarge)
-- Intended for Start-up Analysts and Researchers --

Analysis of Sales Population

1. Instead of just one parameter (like average) consider the expanded Percentile curve, preferably 1st percentile to 99th percentile. Avoid Minimum and Maximum as they may skew the picture.

2. Normally, 5th percentile to 95th percentile is the most meaningful inlier range, extending out to the outliers on both ends. If the sample is excessively large, 1st to 99th could be used as well.

3. So, while studying outliers, consider below the 5th percentile and above the 95th percentile. In case of a large sample, below 1st and above 99th could be experimented.

4. When the time series is extended (e.g., sales between 1/2017 and 3/2018 are used here), sales must be time-adjusted. If the sample comprises 3-4 years' of sales, quarterly adjustments will be in order.

5. Since the growth rates vary by the market, time adjustment factors (coefficients) must be derived at the market level (e.g., 6% in this example). Applying national or even regional factors could result in flawed results.

6. If you have to choose one parameter, it must be the Median, as it is less prone to outliers (Average is heavily influenced by outliers, distorting the analysis).

7. In this example, Median ASP and Median Bldg SF are mutually exclusive (may be connected to get a general idea of the ASP/SF, but not for any serious analysis).

8. In order to analyze ASP/SF, create the organic variable (row-wise) and run percentile stats on it (auto-regressive).

9. Ideally, ASP should be modeled (regression) and the resulting output variable (smoother and statistically more significant) used in all analyses, including defining and removal of outliers.

I picked the above table from TownAnalyst.com as I own and operate it, to avoid having to deal with any copyright issues. My TownAnalyst site is totally free (no strings) and requires no login or registrations whatsoever. Please use the site/system that works best for you.

http://www.townanalyst.com/

homequant@gmail.com    

- Sid Som MBA, MIM
President, Homequant, Inc.

No comments:

Post a Comment