CASE STUDY · 2025 · DTC
5 WEEKS · E-COMMERCE ANALYTICS

Turning a returns problem into a margin lever.

In fast fashion, returns are treated as a fixed cost. I modeled them at the SKU level to show they're a signal instead, which styles, sizes, and reasons drive them, and turned that into merchandising and PDP moves that protected real margin.
ANNUALIZED MARGIN PROTECTED
+$1.1M

From return-driven decisions

LEVEL RETURN MODEL
SKU

Style, size, and reason isolated

OFFENDERS FLAGGED
TOP 5%

The styles quietly bleeding margin

Figures are approximate and rounded to protect client confidentiality; exact numbers are withheld.

The problem.

Returns were logged as a cost of doing business, a single blended rate nobody could act on. That average hid the truth: a small share of styles and sizes drove a large share of the returns, and no one had isolated them.

The approach.

I built a SKU-level returns model: return rate by style, by size, and by reason code, joined to margin. That surfaced the worst offenders and the patterns behind them, sizing that ran off, descriptions that oversold, photography that misled, and translated each into a concrete merchandising or product-page move.

The outcome.

Returns became a decision input instead of a write-off. Acting on the worst offenders protected roughly $1.1M of annualized margin and gave merchandising a repeatable signal, the styles to fix, re-shoot, or retire before they shipped losses.

+
ARTIFACTS

From the work.

Treating returns as a fixed cost?

If returns are a blended rate you can't act on, I'll break them down to the SKU and show you the margin hiding inside. One slot open for Q3 2026.

BOOK A 20-MIN CALL