We will do one analysis from raw numbers to decision, slowly, with nothing skipped. The data: a small stationery distributor in Coimbatore sells ten product lines. Last year’s revenue in lakh rupees, already sorted from the biggest to the smallest, was 52, 24, 14, 9, 6, 5, 4, 3, 2 and 1. The total is ₹120 lakh.
Step one: share of total, then running total
Divide each product by 120. The first is 43.3%, the second 20.0%, the third 11.7%, the fourth 7.5%, and so on down to 0.8% for the tenth. Now add them up as you go down the list. The cumulative share after product one is 43.3%, after two 63.3%, after three 75.0%, after four 82.5%, after five 87.5%, after six 91.7%, after seven 95.0%, after eight 97.5%, after nine 99.2%, and after ten, of course, 100%.
That running-total column is what a Pareto chart plots as a line over the bars. The sorting is not optional. If you compute running totals on an unsorted list, the curve means nothing.
Step two: test the famous ratio
The 80/20 rule says about 80% of results come from 20% of causes. Here, 20% of ten products is two products, and they deliver 63.3% of revenue. Not 80. You need four products, 40% of the range, to cross 80%. So the pattern is real, the revenue is strongly concentrated, but the neat ratio is not a law of nature. Vilfredo Pareto’s original observation concerned land ownership in Italy, and it has become a rule of thumb ever since. Treat 80/20 as a hypothesis to check against your own data, and expect 70/30 or 90/10 just as often.
Step three: draw the ABC lines
ABC analysis turns the curve into classes. A common convention puts items into A until cumulative share reaches about 80%, B until about 95%, and the remainder into C. The cut-offs are a convention, and some firms use 70/90 or 75/95. Using 80 and 95 on our table:
- Class A: products 1 to 4. Four lines, 40% of the range, 82.5% of revenue.
- Class B: products 5 to 7. Three lines, 12.5% of revenue (87.5% minus 82.5% is 5, 91.7 minus 87.5 is 4.2, 95.0 minus 91.7 is 3.3, which together make 12.5).
- Class C: products 8 to 10. Three lines, 5% of revenue.
Three lines are 30% of the range and bring in 5% of the revenue. That is where the tail is.
Step four: decide what each class gets
Here the analysis stops being arithmetic. For class A, the distributor should never run out of stock, negotiate hardest with suppliers, and review the sales forecast monthly. For B, a standard reorder policy. For C, ask whether each line deserves the shelf space and the working capital, or whether it can move to made-to-order, a slower reorder cycle, or exit.
The same method works far beyond inventory. Rank support ticket categories by volume and you find the few issue types that dominate the queue. Rank defects by count, customers by revenue, or cloud services by monthly bill. In each case, the table tells you where a small amount of effort buys the largest change.
Step five: check what revenue hides
Ranking by revenue alone is a trap. A product with ₹1 lakh of revenue might be a cheap loss-maker or it might be a high-margin accessory that makes customers return. Repeat the Pareto on gross profit and compare the two lists. Products that are class A by revenue but class C by profit deserve a pricing review. Products that are class C by revenue but sold alongside class A items may be pulling their weight in baskets.
Also watch for a single-year bias. A class C item might be new, and a class A item might be on its way out. Run the table across two or three years, or at least split the latest quarter, before cutting anything.
Ranking by volume tells a different story from ranking by value too: an item with high unit counts and low price takes up warehouse effort without moving the revenue needle, and that is a handling cost the revenue Pareto will never show. Pareto tells you where to look. It does not tell you what you will find.
Figures and names in the examples are illustrative.
Leave a Reply