Guide
ABC analysis: putting your effort where the money is
Most catalogues have a small number of lines carrying most of the value and a long tail carrying almost none. ABC analysis finds the split so you stop treating both the same way.
The idea
Managing 2,000 products identically is impossible and, more importantly, pointless. Effort spent perfecting the reorder point on a line that sells four units a year is effort not spent on the line that turns over ₹40 lakh.
ABC analysis sorts the catalogue by annual consumption value — units sold per year times unit cost — and splits it into three classes. The usual shape:
| Class | Share of lines | Share of value | How to treat it |
|---|---|---|---|
| A | ~10–20% | ~70–80% | Tight control, frequent counts, calculated reorder points, high service level |
| B | ~20–30% | ~15–20% | Moderate control, quarterly counts, simple reorder rules |
| C | ~50–70% | ~5–10% | Loose control, annual counts, flat cover rules, larger order quantities |
The exact percentages do not matter and are not a law of nature. What matters is that the distribution is steep, and almost every catalogue's is.
Scroll the diagram sideways →
The method, step by step
- Pull annual usage per product. Units sold or consumed over twelve months. A shorter period works if you scale it, but avoid periods dominated by one season.
- Multiply by unit cost. This gives annual consumption value. Use cost, not selling price — you are measuring capital at risk, not revenue.
- Sort descending by that value.
- Add a running cumulative percentage of total value.
- Draw the lines. Everything up to ~80% cumulative is A; from there to ~95% is B; the remainder is C.
- Sanity-check the boundaries. If A comes out at 45% of your lines, either your catalogue is unusually flat or the usage data is wrong.
Do it in a spreadsheet the first time. Export the catalogue with annual usage and cost, sort, and add one cumulative column. It takes about twenty minutes and you will learn more about your business in those twenty minutes than in most quarterly reviews.
A worked example
A distributor with 500 SKUs and ₹4 crore of annual consumption value:
| Lines | % of lines | Value | % of value | Cumulative | |
|---|---|---|---|---|---|
| A | 62 | 12.4% | ₹3.08 cr | 77% | 77% |
| B | 119 | 23.8% | ₹0.72 cr | 18% | 95% |
| C | 319 | 63.8% | ₹0.20 cr | 5% | 100% |
Sixty-two products carry three quarters of the money. The consequences are immediate and concrete:
- Counting those 62 monthly is about three hours of work. Counting all 500 monthly is a fortnight.
- Calculating proper reorder points for 62 lines is an afternoon. For 500 it never happens.
- A 10% cut in A-class stock releases about ₹30 lakh. The same 10% cut across all C items releases ₹2 lakh.
Every one of those is a decision you cannot even see without the split.
What to do differently per class
A items
- Full reorder point calculation at a 98–99% service level.
- Cycle count monthly.
- Order more frequently in smaller quantities — the carrying cost is worth avoiding.
- Watch supplier lead-time variability closely; this is where reducing it pays most.
- Review demand at least quarterly.
B items
- Simple reorder point: lead time demand plus a few days of cover, at 95%.
- Cycle count quarterly.
- Standard EOQ order quantities.
C items
- Flat days-of-cover rule. Being wrong here is cheap.
- Count annually.
- Order in larger quantities less often — ordering cost dominates holding cost at this value.
- Review for deletion. The most valuable thing about identifying C items is deciding which of them you should stop carrying at all. Each one occupies shelf space, catalogue attention and a small amount of working capital.
When value alone is the wrong axis
Pure consumption value misses things that will hurt you:
- Criticality. A ₹200 gasket that halts a production line is not a C item in any meaningful sense. Many businesses run a second axis for criticality and promote items regardless of value.
- XYZ analysis. Classifies by demand variability rather than value. An AZ item — high value, wildly unpredictable — is the hardest thing in any catalogue and deserves individual attention.
- Lead time. A cheap item with a fourteen-week lead time needs planning a C classification will not give it.
- Perishability. Dated stock has a deadline that value ranking ignores entirely.
The practical approach is to run ABC on value, then manually promote the handful of items that are critical, long-lead or perishable. That list is usually short and every business already knows most of it.
Pitfalls
- Using selling price instead of cost. Inflates high-margin lines and misrepresents capital at risk.
- Using stock value instead of consumption value. A product with ₹5 lakh sitting on the shelf and no sales is not an A item — it is a dead-stock problem, and the ageing report is the right place to see it.
- Running it once and filing it. Classes drift. Re-run every six to twelve months.
- Treating the boundaries as sacred. 80/95 are conventions. If your data has an obvious cliff, put the line at the cliff.
- Forgetting new products. A line launched two months ago has no annual usage and will land in C. Exclude and manage new lines separately until they have history.
Once the classes exist, they should change how the system behaves — different service levels, different count frequencies, different order sizes. An ABC analysis that produces a spreadsheet nobody acts on has produced nothing. In SmartShelfKart the ABC report runs off the same transaction record as everything else, so it stays current instead of being a one-off exercise.