Choosing What to Sell

01

A category is a business model in disguise

Two sellers with identical skill, identical work rate, and identical capital will earn very different amounts if one sells vintage denim and the other sells textbooks. Not because one is better at reselling, but because the categories have different structures — and the structure does most of the work.

What a category actually fixes for you, before you make a single decision of your own:

• Average sale price, which sets how many transactions you need to reach any given income.
• Handling time per unit — research, photography, description, packing. This is the ceiling on your throughput.
• Postage and packaging cost, which is roughly fixed per parcel and therefore hurts cheap items disproportionately.
• Return and dispute rate. Clothing that has to fit and electronics that have to work carry structurally higher rates than a sealed book.
• Storage footprint per dollar of inventory. Furniture and jewellery hold wildly different amounts of value per shelf.
• Authentication and liability exposure, which is near zero in some categories and business-ending in others.
• Seasonality, which determines whether your cash flow is smooth or violently lumpy.

You can be excellent at every skill in this course and still lose to those seven variables. Choose them deliberately.

02

Margin per handling hour is the real number

Almost every seller who stalls has optimised the wrong number. They chased average sale price, or gross margin percentage, and ended up with a category that pays well per item and poorly per hour.

The calculation is simple, and doing it honestly is the hard part:

1. Net proceeds per unit Estimate net proceeds per unit — sale price less commission, less postage and packaging, less your cost. The fees guide covers what actually comes out.
2. Time the whole cycle Time yourself on the full handling cycle for one representative item: sourcing time amortised, research, photography, listing, and packing.
3. Count the invisible time Include the invisible time — the messages, the occasional return, the trip to post.
4. Then divide Divide. That number is what the category pays you per hour of your life.

A common and instructive result: bulk-listable items with a small unit margin frequently beat individually researched high-value items, because the research is the expensive part and it does not shrink with price. The opposite result is equally common in categories where one photograph and three words are enough. There is no general rule. There is only your measurement.

Neither column is inherently right. The point is that the column you would have guessed at — the one with the bigger number in it — is not reliably the one that pays more. Measure before you commit inventory.

03

Check demand and supply separately

These are two different questions and sellers routinely answer only the first. A category is workable when both are true, and a category where only one is true will read as an opportunity right up until you try to buy stock.

Demand: are people buying this, repeatedly, at prices that work?

• Look at completed sales only, never active asking prices. The valuation guide explains why active listings are a record of hope.
• Count how many sold in the last month, not the last year. A category with four sales a year has a price but no demand.
• Check the spread. If comparable items sell anywhere between $20 and $200, something you cannot see is driving the price, and until you know what it is you cannot buy safely.
• Look at the sell-through rate — sold versus listed. It is the single most informative number available to you before you commit.

Supply: can you actually buy this, at a price that leaves room, more than once?

• Go and try. Not hypothetically — actually attempt to buy three units at your target cost this week.
• Ask whether the supply is repeatable or a one-off. A single lucky find is not a category.
• Check whether your competition includes the people who make the thing. Competing with a manufacturer on a current product is a different business.
• Note whether supply is geographically constrained. Some categories work beautifully in one region and not at all in another.

04

The traits that make a category workable

Beyond the arithmetic, some structural traits make a category friendlier to a small operation. None is decisive alone; together they explain most of why one seller's year is calmer than another's.

• Identifiable. Items with a model number, ISBN, catalogue number, or barcode are faster to research, faster to list, and much less likely to be described wrongly. Unidentifiable items push all the work onto your judgement, every single time.
• Robust. Things that survive being dropped cost less to pack, cost less to post, and generate fewer damage claims. Fragility is a permanent tax, as the packing guide makes concrete.
• Light and small. Postage is charged on the greater of actual and volumetric weight. Density of value is a direct input to your margin.
• Objectively gradable. Categories with established condition vocabularies produce fewer disputes than categories where "good condition" means whatever the buyer hoped.
• Not size-dependent. Anything that has to fit a body carries a structurally higher return rate that no amount of description fully removes.
• Not seasonal to the point of famine. Some seasonality is fine and even useful; a category with one buying window a year is a cash flow problem.
• Low counterfeit exposure. Some categories are saturated with convincing fakes, and the authentication guide should be read before entering one, not after.

The inverse list is worth stating too, because attractive-looking categories often fail on it: heavy, fragile, unidentifiable, size-dependent, heavily counterfeited, and requiring specialist knowledge you do not yet have. Any two of those together is a hard category. Any four is a category best left to the people already good at it.

05

How narrow to go

Advice on this splits into two camps that are both right in different conditions, which is why it is confusing.

Narrow wins on skill. A seller who handles one kind of thing gets faster at every step: they recognise value instantly, grade accurately, write descriptions that read as expert, and build the kind of following the seller brand guide describes. Depth compounds in a way breadth does not.

Broad wins on supply. Sourcing is the binding constraint for most small sellers, and a narrow specialist walks past ninety percent of what is available. In a market where stock is scarce, breadth is what keeps the shelves full.

The resolution most successful small sellers arrive at is a core plus an opportunistic edge: one category deep enough to be genuinely expert in, plus a willingness to buy obvious value outside it when it appears. The core builds the skill and the reputation; the edge keeps the volume up. What does not work is uniform shallowness — five categories, expert in none, competing on price with people who are.

06

Test it before you commit to it

A category test is a small, time-boxed experiment with a decision written down before it starts. Without the decision date it is not a test — it is just the beginning of an inventory problem.

1. Buy a deliberately small batch Ten to twenty units, at the cost you believe the category supports. Not one unit, which tells you nothing about variance, and not a hundred, which is a commitment wearing a test's clothes.
2. List them properly A test run listed badly measures your listing, not the category.
3. Set the decision date first Set a decision date before you start — sixty or ninety days out — and write down what result would make you continue.
4. Track four numbers How many sold, how long they took, what they netted, and how many minutes each consumed end to end.
5. Decide on the date On the decision date, decide. Continue, adjust, or stop and clear the remainder using the ladder in the aged stock guide.

The failure mode this prevents is the slow one: a category that never quite worked and never quite got abandoned, absorbing shelf space and attention for a year because no moment ever arrived at which anyone had to say so. Set the moment in advance.

One caution on reading results: a batch of twenty is a small sample, and small samples are noisy. Treat a clear result as informative and a marginal one as inconclusive — if it comes out marginal, the honest answer is usually that the category will pay you about market rate for your time, which is a real answer and worth knowing.

07

Changing your mind later

Categories are expensive to change but not impossible, and the cost is almost entirely in inventory you already own. Knowing that in advance changes how you should buy while you are still deciding.

• While testing, buy shallow rather than deep. Ten different things beats ten of one thing when you are still learning what sells.
• Prefer stock that clears quickly over stock that maximises margin, until you are committed. Liquidity is worth paying for during the uncertain phase.
• Keep the records that let you compare later — cost, date, channel, net proceeds. The COGS guide is what makes a category comparison possible at all, and it has to be running before the comparison, not after.
• Do not let sunk cost hold the decision. What you paid for stock is irrecoverable; the only live question is what your next hour and next dollar should go into.

The most common real-world path is not choosing a category and staying there forever. It is choosing one deliberately, measuring it honestly, and letting the measurements move you — which requires only that you were tracking the right things from the start.