The Problem: Treating All Products Equally
Most inventory planning systems apply the same rules to every SKU. Same safety stock formula, same reorder logic, same review frequency. On the surface, this seems fair and systematic.
In practice, it's a disaster.
Your top 50 products might generate 80% of revenue. Applying "average" attention to them means under-managing your most critical items. Meanwhile, you're over-managing hundreds of slow movers that barely move the needle.
ABC analysis fixes this by classifying products based on importance, then applying different management strategies to each class.
The Pareto Principle in Inventory
ABC analysis is built on the Pareto principle: roughly 80% of effects come from 20% of causes. In inventory terms:
- ~20% of SKUs typically generate ~80% of revenue
- ~50% of SKUs typically generate ~95% of revenue
- The remaining ~50% of SKUs generate ~5% of revenue
These aren't exact percentages—your business will vary. But the pattern holds across almost every product catalog.
The Three Classes
A-Items (Vital Few):
- Top 10-20% of SKUs by revenue
- Generate 70-80% of total revenue
- Require the closest management attention
B-Items (Middle Ground):
- Next 20-30% of SKUs
- Generate 15-25% of revenue
- Require moderate management attention
C-Items (Trivial Many):
- Remaining 50-70% of SKUs
- Generate 5-10% of revenue
- Require minimal management attention
How to Conduct ABC Analysis
Step 1: Gather Your Data
Pull the last 12 months of sales by SKU. You need:
- SKU identifier
- Total units sold
- Total revenue (units × price)
Use revenue rather than units. A SKU selling 10,000 units at $1 is less important than one selling 1,000 units at $50.
Step 2: Sort and Calculate
Sort SKUs by revenue, highest to lowest. Then calculate:
- Cumulative revenue
- Cumulative revenue percentage
Step 3: Apply Classification Cutoffs
Typical cutoffs:
- A-Items: SKUs representing the top 80% of cumulative revenue
- B-Items: SKUs representing the next 15% (80-95% cumulative)
- C-Items: SKUs representing the final 5% (95-100% cumulative)
Example with 500 SKUs:
| SKU Rank | Revenue | Cumulative % | Class |
|----------|---------|--------------|-------|
| 1-15 | $2.4M | 80% | A |
| 16-100 | $450K | 95% | B |
| 101-500 | $150K | 100% | C |
15 SKUs are A-items, 85 are B-items, 400 are C-items.
Step 4: Validate and Adjust
Review your classifications. Sometimes the math produces odd results:
- A new product might land in C when it should be managed like an A
- A declining product might be an A historically but should be demoted
- Strategic products (loss leaders, gateway products) may need different treatment
Use the analysis as a starting point, then apply business judgment.
Management Strategies by Class
A-Items: Intensive Management
These are your vital few. Stockouts here hurt the most. Apply maximum attention:
Forecasting: Use the most accurate methods available. Review forecasts weekly. Track forecast accuracy and adjust.
Safety stock: Target 95-98% service level. Higher is justified given the revenue concentration.
Reorder frequency: Order more frequently with smaller quantities. This reduces average inventory while maintaining availability.
Review frequency: Weekly at minimum. Daily during peak seasons or promotions.
Supplier management: Prioritize reliable suppliers, even at premium cost. Build relationships. Have backup sources identified.
B-Items: Moderate Management
The middle ground. Important but not critical:
Forecasting: Standard statistical methods. Review monthly.
Safety stock: Target 90-95% service level. Balance cost and availability.
Reorder frequency: Standard order cycles. Optimize for ordering efficiency (hitting minimums, consolidating shipments).
Review frequency: Bi-weekly to monthly.
Supplier management: Standard vendor relationships. Price competitive bidding.
C-Items: Simplified Management
These contribute little to revenue but consume disproportionate management time if you let them:
Forecasting: Simple averages. Don't invest time in sophisticated forecasting for items that barely sell.
Safety stock: Target 85-90% service level. Occasional stockouts are acceptable.
Reorder frequency: Infrequent, larger orders. Accept higher average inventory in exchange for less ordering activity.
Review frequency: Monthly to quarterly.
Supplier management: Price-driven. Consider consolidating C-items with fewer suppliers for simplicity.
Beyond Simple ABC: Adding Dimensions
Basic ABC uses revenue only. More sophisticated approaches add dimensions:
ABC-XYZ Analysis
Combines revenue importance (ABC) with demand predictability (XYZ):
- X items: Stable demand, easy to forecast
- Y items: Variable demand, moderate forecast difficulty
- Z items: Irregular demand, hard to forecast
This creates a 9-cell matrix:
| | X (Stable) | Y (Variable) | Z (Irregular) |
|---|---|---|---|
| A | Tightest management | Close watch | Special attention |
| B | Standard process | Monitor | Evaluate necessity |
| C | Automate | Simplify | Consider discontinuing |
An A-Z item (high revenue, unpredictable demand) needs different treatment than an A-X item (high revenue, stable demand).
Adding Profitability
Revenue doesn't equal profit. A high-revenue, low-margin item may contribute less to the business than a moderate-revenue, high-margin item.
Consider classifying by:
- Contribution margin (revenue minus variable costs)
- Gross profit dollars
- Unit economics
Adding Strategic Value
Some products are strategically important beyond their direct contribution:
- Gateway products that introduce customers to your brand
- Products that drive cross-selling
- Products that serve key customer segments
Flag these for elevated management regardless of their revenue classification.
Implementation Tips
Start simple. Basic ABC is better than no ABC. Add complexity later.
Update regularly. Run ABC analysis quarterly. Products migrate between classes as the business evolves.
Communicate classifications. Your team should know which products are A, B, and C. Make it visible in your systems.
Don't over-engineer C-items. The temptation is to "optimize" everything. The whole point of classifying C-items is to spend LESS time on them.
Watch for class migration. A-items that are declining should be demoted. C-items that are growing should be promoted. Set up alerts.
Common ABC Mistakes
Using units instead of revenue. Selling lots of cheap items doesn't make them important.
Never updating classifications. The analysis is a snapshot. Business changes. Update regularly.
Applying ABC to too narrow a scope. Classify your full catalog, not just one category. Importance is relative to your whole business.
Ignoring new products. New items don't have history. Classify based on expected performance, then validate with actual data.
Being too rigid. ABC is a guide, not a law. Use judgment.
Key Takeaways
- ABC analysis classifies inventory by importance to focus management effort
- A-items (top 20% of SKUs) typically generate 80% of revenue and need intensive management
- C-items (bottom 50%) generate minimal revenue and should be simplified
- Apply different forecasting, safety stock, and review approaches by class
- Add dimensions (predictability, profitability, strategy) for more nuance
- Update classifications quarterly as business evolves
Frequently Asked Questions
How often should I recalculate ABC classifications?
Quarterly is standard. Monthly if your business is highly seasonal or rapidly changing. Annual is too infrequent—products can migrate significantly in a year.
What if my business doesn't follow the 80/20 rule?
That's fine. The exact percentages don't matter. The principle—that some products matter much more than others—holds even if your split is 70/30 or 90/10. Adjust your classification cutoffs accordingly.
Should I use the same ABC classification across all channels?
Not necessarily. A product might be an A-item on Amazon but a C-item in retail. Consider channel-specific analysis if your channels have very different product mixes.
How do I handle product bundles in ABC analysis?
Classify bundles as their own SKUs based on bundle-level revenue. Don't try to allocate bundle revenue back to component products—it overcomplicates the analysis.
What about brand new products with no history?
Classify new products based on expected performance (using similar products as proxies) and review after 3-6 months of actual data. Many companies default new products to B-class until data proves otherwise.