Assortment review: what is dragging you down and why

An assortment is a portfolio, not a list

Most catalogues follow a harsh asymmetry: a fifth of the positions generate most of the revenue, roughly the same share loses money, and the middle hovers near zero while consuming all your attention. The point of a review is to see those three groups explicitly and assign each of them a fate, rather than trying to lift everything at once.

Classification without magic

Take an export covering the last six months and calculate four numbers per item: revenue, margin in currency, days of inventory turnover, and return rate. The split then becomes straightforward.

Workhorses. High absolute margin and fast turnover. You do not touch their content on a whim, only through controlled tests.
Ballast. Margin near zero or negative, slow turnover. Candidates for delisting or for renegotiating the purchase price.
Potential. Healthy margin but low traffic. This is exactly where the model and listing work belong, because improvements have something solid to stand on.
Traps. High revenue, decent margin on paper, but returns a third or more above the category norm. These items are frequently loss-making once recalculated.

A prompt for the first pass

Prompt: Here is an export of 200 SKUs over 6 months: units sold,
revenue, margin, days on hand, return rate, category.
Split them into 4 groups: workhorses, potential, ballast, traps.
For each group: number of items, total margin,
and how much inventory value is frozen in it.
For the ballast group, list the 10 worst and state what is sinking each one:
price, turnover, or returns. Use only the data in the table.

The model's answer is a list of hypotheses, not a verdict. You then open the ten worst by hand and look at what is actually going on: often half of them turn out to have wrong dimensions in the product data, which means shipping is being charged at an inflated tier.

Insight. Frozen inventory is the quietest leak in the business. Ballast rarely shows up in a profit report, yet it ties up the cash that could be working inside your workhorses.

Common mistake. Delisting an item based on one month of data. A seasonal product looks exactly like ballast in its low season, and two months later you rebuy it at a higher price.

Pro tip. Measure not just an item's own margin but its role in the basket: some products lose money on their own yet are almost always added to an expensive main item. Check basket composition before you delete anything.

Cheat sheet

Four groups: workhorses, potential, ballast, traps.
Content and ad spend go into potential, not into the whole catalogue.
High return rates turn a profitable item into a loss-maker.
Check an item's basket role before delisting it.

1. Which group is the most sensible place to invest listing and model work?

2. What makes the traps group dangerous?

3. Why can you not delist a SKU on one month of data?

Task — checked by AI

Export your own data for at least six months (or the full life of the catalogue if it is shorter) and split your assortment into four groups: workhorses, potential, ballast, traps. Run the table through the model using the prompt from the lesson, then open the ten worst items by hand and write down what is sinking each one: purchase price, turnover, returns, or a data error such as dimensions, category or weight. State how much cash is frozen in ballast and which three items you are delisting.

← Back

🔒 Answer the question correctly to move on to the next lesson.