THE CURRICULUM / CH 1 — Products

1.1Finding Product

Find what is already selling by reading long-running ads, then put it through demand, competition and margin.

6 min read

You open the research tool with nothing in mind, so you type in the thing you saw on your feed last night. Forty minutes later there are eleven tabs open, one product you already liked, and no reason you could say out loud. That is shopping, not research.

Three bands of ad duration and what each is worth as a signal. An ad running for days tells you nothing. One running for weeks has survived a cull and is worth a second look. One running for months is a profit confession, because someone is paying for it every day. Duration says nothing about how much.
A launch-week ad proves a budget. An ad still live months later is a profit confession.

START WITH THE ADS, NOT THE PRODUCT

You do not invent a product. You find one already selling, and the only public evidence that anything sells is that somebody keeps paying to advertise it.

DURATION IS THE SIGNAL.

a launch week proves a budget; a long run proves the maths

One competitor advertisement broken into its five parts, with what each part tells the reader about the advertiser. The hook, the first two seconds before any argument, tells you what they think stops this audience. Who it talks to, read from the casting, setting and pronouns, tells you who they have decided the buyer is. The promise, the change it says it will make, tells you which problem they sell against. The proof, meaning what is shown rather than what is said, tells you which objection they expect to meet. The ask, what it wants done next, tells you how warm they think the viewer already is. This is a reading key: you are reading their assumptions, not their results, because no result is visible from outside an ad.
The buyer, the problem, the promise, the price.

THREE GATES EVERY CANDIDATE PASSES

A nine-point tick list a product must clear before you spend anything on it, drawn as a form with a hairline box beside every line. Demand, competition and margin each take a line; then demand evidence from outside your own feed spread across months, every guessed figure marked on the sheet as a guess, delivery inside the patience of the person buying, a buyer described as more than anyone, one thing you will change that is not being cheaper, and a test budget you can lose with the judging date set now. The verdict rule at the foot is absolute: if any box is unticked it is not a candidate. Acceptable on all three gates beats spectacular on one. No scores, thresholds or targets appear anywhere.
What a candidate clears before a dollar goes near it.
  • Acceptable on all three beats spectacular on one. Spectacular on demand alone is the product with a thousand sellers.
  • Saturation is identical presentation everywhere and a sliding price, not a high seller count.
  • Your feed is not a source. It is a delivery result, and it reached every other beginner too.

Then ask when it sells. A store built on six weeks of the retail calendar is a graveyard for the other forty-six.

The Australian retail year as a vertical ladder, one row per month. January, back to school and the post-sale lull, with term starting late in the month. February, Valentine's Day on the fourteenth, and summer ending. March, autumn, with Easter falling here or in April. April, Easter, school holidays and Anzac Day on the twenty-fifth, with public holiday trading rules differing by state. May, Mother's Day on the second Sunday, and end-of-financial-year campaigns opening in the back half. June, the financial year ends on the thirtieth and business buyers bring deductible purchases forward. July, the new financial year begins on the first, tax returns open, and deep winter begins. August, no national retail event — the month to build, shoot and fix. September, Father's Day on the first Sunday, which in Australia is September and not June. October, the run-up: stock, creative and cash for the peak are decided now. November, Black Friday and Cyber Monday late in the month, plus Singles Day and Click Frenzy, and the cyclone season running from the first of November to the thirtieth of April, which can delay freight into northern Australia. December, Christmas and then Boxing Day on the twenty-sixth, with Australia Post publishing its cut-off dates each year and international closing earliest. The courier cut-off is the only date here that cannot be moved. Seasons are inverted, moveable dates are flagged as moveable, retailers are named only as anchors, and no sales figure is attached to anything.
The year a product has to survive, and the parts that move on their own.
Statement poster: "We do not lose money. We purchase information."

MARGIN IS A STACK, NOT A NUMBER

A worked example taking one selling price apart. An insulated stainless steel drink bottle is listed at seventy nine dollars ninety five in Australian dollars. Seven lines are removed in order: unit cost ex works eleven forty, freight and import duty three sixty, GST remitted seven twenty seven which is the price divided by eleven, payment processing one seventy at one point seven five per cent plus thirty cents, pack and deliver in Australia twelve ninety nine, returns and replacements one twelve, and apps and platform one eighty seven. Those seven come to thirty nine ninety five, leaving forty dollars of contribution before advertising. Advertising then takes twenty eight dollars per order, leaving twelve dollars, which is fifteen per cent of the price. It is a worked example, not a quote and not a result.
Worked example. Run the same arithmetic on yours.

Where the stack goes wrong

LineHow it gets missed
Product costA price quoted for a volume you will not buy
Delivery of one unitA domestic rate on something shipping abroad
Platform, app and processing feesRounded to nothing because they look small
Returns and replacementsAssumed to be zero because you have no history
Duties and taxesFound out after the first delivery

Build it before you like the product too much to be honest. Ad spend is not in the stack; it fits inside what the stack leaves.

A worked example of one order at seventy nine dollars ninety five drawn as a single vertical bar cut into slices, each sized by the dollars it takes. In order down the bar: the ATO takes seven dollars twenty seven in GST, the factory eleven forty, freight and customs three sixty, the courier twelve ninety nine, the payment processor one seventy, refunded buyers one twelve, the platform and apps one eighty seven, the ad platform twenty eight dollars, and the seller keeps twelve dollars. The ad platform slice is thirty five per cent of the order and the seller's slice is fifteen per cent. The nine slices add to seventy nine dollars ninety five exactly. It is a worked example, not a quote and not a result.
Every hand that takes a piece of a single order.
The chain from an order placed with a supplier to a parcel delivered, with the break points marked between the links. After the order is placed it can turn out the item was never actually in stock, which leaves you having promised a date nobody else agreed to. After it is picked and packed the wrong variant can be in the box, which costs a return you pay for and a review you cannot edit. In transit the tracking can stop updating, so the customer asks you and you know nothing either. At the border it can be held, taxed or turned back, so the customer is asked for money nobody mentioned. The parcel is then delivered, or marked delivered and never received. Every link is somebody else's job and every break is still yours. No failure rate or transit time is shown.
Every failure point between order and delivery lands on that stack.

Break-even on ad spend is one divided by the share of price that survives as contribution: if a third survives, advertising has to return three times what it costs before you are level.

A curve of break-even return on ad spend against contribution margin. Break-even ROAS is one divided by contribution margin, so the curve falls steeply as margin rises: a thin margin needs a high multiple to break even, a fat margin needs a low one. This is arithmetic, not a target.
Arithmetic, not a target. The curve is steep at the thin end.
Store analytics for one day: $169.96 in total sales, 2 orders, 4.26% conversion.
One day in a real store. That is the whole day, and nothing is wrong with it.
Store analytics for 13 to 17 June 2025: 136 sessions, $544.65 in sales, 6 orders, 4.41% conversion.
Five days in the same store. Real, unscaled, not benchmarks.
  1. Confirm demand outside the ads, and count the sellers with a line each on what they do badly.
  2. Get a real quote at your real order quantity, then build the stack with every guess marked.
  3. Decide the one thing you will change. If the answer is nothing, take the next candidate.
One refund traced through every line it lands on. The product is not back on the shelf: it is with the buyer, in the post back, or not coming back at all. The shipping out is already spent, because the carrier moved it once and that work is done and paid for. The shipping back is paid by someone, and if you take the order back the return leg is a second freight cost on one order. The ad spend is already spent: it bought this order, the order went away, and the spend did not go with it. The processor fee depends on your own agreement, which sets what comes back and what is kept. Your time is the line nobody bills you for: the message, the decision, the paperwork, and the next one arriving behind it. The sale reverses; the costs it already made do not. No amounts, fee percentages or refund rates are shown.
The sale reverses. The costs it already made do not.

TAKEAWAYS

  • You find a product already selling. Duration of its ads is the signal.
  • Working backwards from a proven ad is research. Cloning it cheaper is the trap.
  • Demand, competition and margin must all be acceptable. Excellence on one forgives nothing.
  • Supplier cost to retail price is the top of a stack, not the margin.

CHECK YOURSELF

A product keeps appearing on your feed and several accounts are posting it. What has that actually told you?

A supplier quotes a unit cost and you plan to sell at four times it. What is missing before that difference is margin?

DO THIS BEFORE THE NEXT MODULE

  1. 01Run the Majorka steps on one category and record the ten longest-running ads.
  2. 02Build the full landed cost stack for your three best candidates, marking every guess.
  3. 03Choose one candidate for the next module and write the single thing you would change.
Enrol free