Project profile

To reduce SKUs, this company used multiple sales variables ($ sales, units sold, and $ margin) to determine which items to discontinue. The CEO of the company called this project a “triage” project because our job was to quickly determine which SKUs to drop and which to keep. The project was intended to provide quick answers on how to get back to the historical SKU count.

Performed a sales Pareto analysis to confirm that there are underperforming SKUs (we nicknamed the project team “The Biggest SKLUsers”)

Identified which variables define a high performance SKU

Calculated an “Overall Yield Factor” for all items using multivariate Pareto analysis based on $sales, unit sales, and $gross margin

You ordered the SKUs based on this overall performance factor from best to worst

Identified the bottom 25% as underperforming (these items were the largest users of SKLU)

While our value stream mapping study showed that the number of SKUs had increased and sales were flat, we wanted to understand the full extent of the problem. We perform a quick Pareto analysis of sales. The drop from the best 25% to the worst 25% was dramatic.

Top 25% of SKUs account for 70.6% of sales

The bottom 25% (or 700 items) represents only 1.5% of sales

We had SKLUsers (our nickname for underperforming SKUs). We then used multivariate Pareto analysis to identify which items were SKLU users based on a combination of measures. Multivariate Pareto was used, rather than just a direct sales analysis, due to the importance of different sales variables.

Sales$ is the most commonly used performance measure. Others argued that Gross Margin $ should be used as it represents cash flow. Others thought that if an item has a low unit price, but had a high turnover, this should be considered because volume is a big driver for this business.

Looking at the data, we saw that different SKUs were strong or weak on various measures. For example, this company loses money on some items to drive sales (loss leaders). These items had negative margin $, but generated significant sales and units (and store traffic).

What defines an excellent or poor performing SKU is multidimensional, which prompted the use of multivariate Pareto analysis.

Pareto analysis of multiple variables

Our multivariate Pareto analysis created an overall return factor. We calculate this factor for all SKUs and rank them from highest to lowest performance. Below are the top 10 and bottom ten SKUs.

Top Ten Items by Overall Performance Factor (OPF)

SKU#1, OPF=7.46, Sales $3,013,442, Units 1,117,009, Margin $313,514

SKU#2, OPF=5.61, Sales $1,015,888, Units 953,248, Margin $453,229

SKU#3, OPF=5.29, Sales $1,398,854, Units 1,252,197, Margin $288,231

SKU#4, OPF=5.02, Sales $2,882,798, Units 8,642,193, Margin -$113,737

SKU#5, OPF=4.43, Sales $1,569,511, Units 1,586,938, Margin $81,269

SKU#6, OPF=4.36, Sales $1,566,789, Units 585,275, Margin $215,773

SKU#7, OPF=4.21, Sales $1,172,974, Units 520,190, Margin $293,701

SKU#8, OPF=3.76, Sales $1,165,273, Units 1,790,616, Margin $29,906

SKU #9, OPF=3.45, Sales $820,379, Units 509,906, Margin $253,890

SKU #10, OPF=3.43, Sales $689,507, Units 817,707, Margin $226,552

Note: SKU #4 was a top seller, but sold at a loss. This item was a strategic loss leader for this retail company.

Bottom ten items by overall performance factor

SKU# 2791, OPF=0, Sales $1.99, Units 1, Margin $0

SKU# 2792, OPF=0, Sales $1.99, Units 1, Margin $0

SKU# 2793, OPF=0, Sales $1.89, Units 1, Margin $0

SKU# 2794, OPF=0, Sales $1.59, Units 1, Margin $0

SKU# 2795, OPF=0, Sales $1.49, Units 1, Margin $0

SKU# 2796, OPF=0, Sales $1.49, Units 1, Margin $0

SKU# 2797, OPF=0, Sales $1.29, Units 1, Margin $0

SKU# 2798, OPF=0, Sales $0.99, Units 1, Margin $0

SKU#2799, OPF=0, Sales $0.69, Units 1, Margin $0

SKU# 2800, OPF=0, Sales $3.01, Units 3, Margin -$0.94

What was most surprising was not the performance of the top 10 articles, but the extremely poor performance of the bottom 10. These SKLUsers sold around 1 unit each during the last 12 months. In stores and in the warehouse they were gathering dust and would eventually have to be scrapped or heavily discounted.

Results

Labor costs have been reduced by $1.3 million due to the elimination of handling of these low-volume, slow-moving items. Inventory will be reduced by $5 million after these slow-moving items are removed from distribution centers and not replaced.