Ganesh Subramanian

Advancing effectiveness, eliminating waste universally

[Ganesh Subramanian @ganeshsub]

The industry standard 50% failure rate is not inevitable. It's the result of treating every SKU as deserving equal commitment instead of proportional commitment based on evidence.

[Ganesh Subramanian @ganeshsub]

50% of products fail every season. Not because retailers pick wrong. Because they bet wrong.

[Ganesh Subramanian @ganeshsub]

Result: 22% reduction in markdown dollars and 19% improvement in full-price sell-through within two seasons.

[Ganesh Subramanian @ganeshsub]

The shift was not in what products they carried. The shift was in how much capital each product received.

[Ganesh Subramanian @ganeshsub]

Tier 4: No demand signal. Pure speculation. Supplier push with no consumer pull. Under 5% of capital.

[Ganesh Subramanian @ganeshsub]

Tier 3: Weak signals. Experimental silhouettes. Poor test results but executive sponsorship. These got 10-15% of capital.

[Ganesh Subramanian @ganeshsub]

Tier 2: Emerging signals. New styles from proven silhouettes. Early market traction. These got 20-30% of capital.

[Ganesh Subramanian @ganeshsub]

Tier 1: Validated demand. Bestsellers with 85%+ sell-through. These got 40-50% of category capital.

[Ganesh Subramanian @ganeshsub]

A sportswear brand applied this logic by creating four capital allocation tiers based on demand signals.

[Ganesh Subramanian @ganeshsub]

Kelly's Criterion from the 1950s answers the question every merchant should ask: How much capital does this product deserve based on demand signal strength and risk of being wrong?

[Ganesh Subramanian @ganeshsub]

This is the equivalent of betting your entire bankroll on a coin flip because the casino requires a minimum wager.

[Ganesh Subramanian @ganeshsub]

Most merchants commit the same inventory depth to proven bestsellers and untested line extensions. Same capital to fill shelf space. Same depth to meet supplier minimums.

[Ganesh Subramanian @ganeshsub]

The compounding comes from repetition, not the size of the initial bet. Small habits, executed consistently, until the advantage becomes impossible to close. What decision rhythm are you repeating that compounds?

[Ganesh Subramanian @ganeshsub]

Most retailers think they have a data problem. They actually have a habit problem. The winners I work with don't have better analytics. They have better decision rhythms that compound every season.

[Ganesh Subramanian @ganeshsub]

What makes this hard to copy: Competitors see the gap and try to close it with one big analytics investment. Doesn't work. You can't buy in one transaction what someone built through repeated multiplication.

[Ganesh Subramanian @ganeshsub]

Most retailers won't do this. They'll keep chasing the hero product, the transformational platform. By year three, the ones who built the habits are operating in a different universe. Higher accuracy, lower markdowns, faster turns. The advantage becomes structural.

[Ganesh Subramanian @ganeshsub]

A sportswear brand went from two drops to monthly launches. Smaller bets, faster feedback. Winners got scaled. Losers got killed before they became expensive. The learning from each cycle fed directly into the next decision.

[Ganesh Subramanian @ganeshsub]

The gap between retailers isn't data volume. It's cycle speed. Two seasonal bets per year = two learning cycles. Monthly micro-launches = twelve. More cycles = faster compounding.

[Ganesh Subramanian @ganeshsub]

If you only watch your category, you see the trend when it arrives. If you layer signals across categories, you see it coming. One global retailer adjusted home buying based on fashion signals. Secured better pricing. Entered the season ahead of the curve.

[Ganesh Subramanian @ganeshsub]

Working with brands across fashion, sports, and home, I see the same pattern. Cross-category signal layering reveals trends months early. Earth tones accelerate in apparel, then appear in home textiles, then in auto accessories.

[Ganesh Subramanian @ganeshsub]

Result: Assortment didn't shrink. It got sharper. Forecast systems got cleaner data. Accuracy improved because they stopped trying to optimize around products that shouldn't exist.