The Price on the Item Is Not the Price You Got
A catalogue check compares one number against another number, both of which somebody typed. It will find the item priced below cost. It will not find the item discounted every afternoon, the supplier who crept eleven percent over a year, or the freight that arrived three weeks after the goods.
Ask a retailer what margin they make on an item and the answer arrives quickly, because it is a subtraction anyone can do: the selling price minus the buying price, both of which are sitting right there on the item record.
Ask what margin they actually made on the last hundred units of it, and the answer takes longer, because that number has been altered by a discount at the counter, by a supplier price that moved in March, by freight allocated after the goods were already on the shelf, and by three units that were returned.
Both numbers are useful. They answer different questions, they live in different places, and treating either as the other is where most retail margin analysis quietly goes wrong.
What a catalogue check can see
The anomaly radar in AWRA runs three checks against the item master, and it is worth knowing exactly what those three are, because their scope is also their limit.
| Check | What it looks for | Severity |
|---|---|---|
| Selling below cost | Selling price lower than buying price | High |
| Thin margin pricing | Margin under five percent, but not negative | Medium |
| Missing selling price | No selling price, or a selling price of zero | Medium |
Each returns the five worst offenders, ordered so the most severe appears first, and the whole thing is cached for a quarter of an hour so a busy dashboard is not recomputing it on every load.
These three catch a specific and genuinely valuable class of error: the typing mistake, the item set up in a hurry, the price never updated after a cost increase. Every catalogue of any size has them, they are invisible until something looks, and each one is losing money on every sale.
An item priced below cost does not announce itself. It sells well, because it is cheap, and every sale makes the problem larger.
The count saturates at five
Worth knowing before you read the number as a scale: the count reported alongside each check is the count of the five examples it fetched, so it reads "5 items are priced below cost" whether you have five or five hundred. The five shown are genuinely the worst five, so as a prompt it works — but treat it as "at least five" rather than "exactly five", and fix them, refresh, and look again until the list comes back empty.
What a catalogue check cannot see
A list price is what you intended. Almost everything that erodes retail margin happens after the intention, and none of it touches the item record.
- The discount at the counter. An item with a healthy list margin, sold at fifteen percent off every afternoon because a cashier has discretion and a customer asked. The item record still shows the intended margin, and always will.
- The supplier price that crept. Nobody raises a price by twenty percent — they raise it by three, four times a year, and the buying price on the item record is whatever it was when somebody last thought to update it.
- The freight that arrived later. Duty, clearing and transport frequently land weeks after the goods, so the buying price on the record was never the landed cost. On imported stock this gap is routinely larger than the whole margin.
- The mix. An item can have an excellent margin and be losing you money, because it is always sold alongside something that does not, in a bundle a customer only ever buys as a bundle.
- The returns. A unit sold and refunded costs you the handling and sometimes the condition, and neither appears anywhere near the item's list margin.
None of that is a criticism of the check. A catalogue check is a hygiene test and it is correctly scoped as one. The mistake is to run it, find nothing, and conclude the margin is fine.
Where the realised numbers actually live
The three questions have three different homes, and knowing which screen answers which saves a great deal of arguing.
-
Intended margin — the item record
Selling price against buying price, checked by the anomaly radar. This is where you find the typing errors and the prices nobody updated after a cost rise. Run it, clear it, run it again until it comes back empty.
-
Realised margin — the sales records
Every sale line stores what was actually charged, what discount was applied, and the tax on it, and cost of goods is accumulated at the moment of the sale on the valuation basis rather than looked up afterwards. That means the margin you actually got is recorded per line rather than reconstructed later.
-
True cost — landed cost allocation
Freight, duty and clearing allocated across a shipment and carried into stock value. For imported goods, comparing a list buying price against a selling price without this is comparing the wrong number.
-
Supplier drift — the vendor scorecard
Weighted average purchase price per month over time, per vendor. This is where a supplier who moved three percent four times becomes a line on a chart rather than a feeling somebody has.
The useful discipline is to check them in that order. A catalogue full of items priced below cost makes every downstream margin number harder to interpret, so clear the hygiene problems first and then ask the harder questions.
What AWRA OpsHub does today
- Three catalogue checks — below cost, thin margin under five percent, and missing or zero selling price — with the worst offenders listed first.
- Realised margin is recorded per sale line: price charged, discount applied, tax, and cost of goods accumulated at the time of sale on the valuation basis.
- Landed costs are allocated across a shipment and carried into stock value, so imported goods can be compared on delivered cost.
- Weighted purchase price drift per vendor over time is computed in the vendor scorecard, alongside on-time rate and lead-time variance.
What it does not do
- The reported count saturates at five. Each check fetches its five worst offenders and reports the size of that list, so the number reads five whether there are five or five hundred. Read it as "at least five", clear them, and look again.
- The radar compares list prices only. It reads the buying and selling price on the item record. An item correctly priced on the master and discounted to nothing at the counter every day is invisible to it.
- No bundle or mix analysis. Margin is per item and per sale line; whether an item is profitable given what it is always sold alongside is a question the system does not attempt.
- No alert on supplier price drift. The drift is computed and charted on the vendor scorecard, but nothing notifies you when a supplier moves — somebody has to open it.
Not ours, by choice
- We will not set your prices or your margin floors. What margin an item should carry depends on your competition, your position and your strategy, and a threshold shipped by a vendor is a guess with a number attached. The five percent thin-margin line is a prompt to look, not a recommendation.
- We will not blend list margin and realised margin into one figure. They answer different questions and averaging them produces a number that cannot be traced back to anything — which is worse than two numbers that each mean something.
A true count rather than a capped one, an alert when a vendor's weighted price moves beyond a threshold you set, and realised-margin anomaly checks alongside the catalogue ones, are all scope rather than ceilings. The sale-line data, the drift computation and the alerting infrastructure all exist and work.
If you import, do the landed cost work before you take any margin figure seriously. Comparing a supplier invoice price against a shelf price on goods that carried freight, duty and clearing is not a margin calculation — it is a subtraction between two numbers that were never comparable.
A quarterly margin review that is worth the hour
Most margin reviews go the same way: somebody exports everything, sorts by margin ascending, looks at the worst twenty items, and finds that nineteen of them are data problems. Doing the data problems first turns the same hour into something useful.
In this order
- Clear the anomaly radar completely — below cost, thin margin, missing price — and refresh until each comes back empty rather than at five.
- Check the buying price on your top fifty items against the most recent purchase order for each. Anything more than a few months stale is a number, not a cost.
- For imported lines, confirm landed costs were allocated to the shipments rather than left on the freight invoice.
- Open the vendor scorecard for your three largest suppliers and look at the price drift chart, not the headline score.
- Compare list margin against realised margin on your twenty best sellers. A consistent gap is a discounting policy nobody wrote down.
- Look at what your highest-discount items have in common. Usually it is one counter, one shift, or one customer, and that is a conversation rather than a pricing change.
The fifth line is where most of the value sits. A gap between what you intended to make and what you made is not a data problem to fix — it is the actual operating reality of your shop, and it is usually the largest single lever available to a retail business.
The discount side of this is worked through in the discount nobody approved, the supplier side in supplier terms, rebates and the margin you actually keep, and the imported-cost side in what is landed cost.
Why below-cost items survive so long
It is worth understanding why this particular error persists, because the answer explains why an automated check earns its place rather than being a nicety.
An item priced below cost sells. It sells better than the items around it, because it is cheaper than it should be, and a fast-selling item generates no complaints from anybody. The buyer sees good volume. The branch sees stock moving. The customer is delighted. Nothing in the ordinary rhythm of a retail week produces a signal.
It surfaces eventually, usually when someone builds a margin report for an unrelated reason and notices a negative number. By then it has typically been running for months, and the total is a figure that would have justified the check many times over.
That asymmetry — an error that produces good news at every point until it is measured — is exactly the shape of problem worth automating. The other two checks are milder versions of the same thing.
Our take
Clear the catalogue checks until they come back empty rather than at five, because the count saturates and "five" may be five hundred. Then stop treating that as a margin review — it is a hygiene test, and passing it means your list prices are sane, not that your margin is. The realised figure is recorded per sale line, the true cost needs landed costs allocated, and the supplier drift is on the vendor scorecard waiting for somebody to open it. Three questions, three places, and the gap between the first and the second is usually where the money is.
See where your margin actually goes
Catalogue checks for below-cost, thin-margin and unpriced items; realised margin recorded per sale line with the discount and the cost of goods; landed costs allocated across shipments; and weighted supplier price drift over time.
Explore reportingFrequently asked questions
What does the anomaly radar actually check?
Three things, all on the item record: items whose selling price is below their buying price, items with a margin under five percent that is not negative, and items with no selling price or a selling price of zero. Each returns the worst five, ordered so the most severe is first, and the result is cached for about fifteen minutes. It is a catalogue hygiene test rather than a margin analysis, and that scoping is deliberate — it catches typing errors, hurried setups and prices never updated after a cost rise.
It says five items are priced below cost. Is that all of them?
Not necessarily. Each check fetches its five worst offenders and reports the size of that list, so the number saturates at five — it reads five whether you have five or five hundred. The five shown are genuinely the worst five, so as a prompt to act it works fine. Read it as "at least five", fix those, refresh, and look again. When the list comes back empty, you have actually cleared them.
Why does an item with a good margin still lose money?
Because the item record holds what you intended, and several things happen afterwards that it never learns about. A discount applied at the counter, a supplier price that crept up in small steps, freight and duty that landed weeks after the goods and were never in the buying price, and returns. On imported stock the landed-cost gap alone is routinely larger than the whole nominal margin, which is why comparing a supplier invoice price against a shelf price is not really a margin calculation at all.
Where do we see the margin we actually made?
In the sales records rather than on the item. Every sale line stores what was charged, what discount was applied and the tax on it, and cost of goods is accumulated at the moment of the sale on the valuation basis rather than looked up afterwards — so the realised figure is recorded per line rather than reconstructed later. Comparing that against the list margin on your twenty best sellers is the single most useful hour in a margin review, because a consistent gap is a discounting policy nobody ever wrote down.
Will we be told if a supplier raises their prices?
Not automatically. The weighted average purchase price per month is computed per vendor and charted on the vendor scorecard, so the drift is visible and reconstructable — but nothing notifies you when it moves, and somebody has to open it. For your three or four largest suppliers, put that on a quarterly rhythm with a name against it. A supplier who moves three percent four times a year is a twelve percent problem that no single invoice ever looks like.
Should we set a minimum margin the system enforces?
That is a real request and it is not what the five percent thin-margin check is. That threshold is a prompt to look, not a floor, and nothing prevents an item being priced below it or sold below it. Whether a hard floor is right depends on your business — promotional lines, loss leaders and clearance all legitimately sit under any threshold you would pick — so if you want one enforced, it is worth scoping deliberately rather than assuming a report is doing it.
