Data Trust levels: how MarginOS grades margin reliability

Data Trust is the grade MarginOS puts next to every SKU’s margin to tell you how much of that number is backed by real evidence. It has three levels — HIGH, MED and LOW — and it answers a question most profitability tools never ask: not “what is my margin,” but “how much should I believe this margin before I act on it.” A SKU whose costs all came from documents is graded differently from one propped up by a store-wide average, and MarginOS shows you which is which.

What is Data Trust?

Data Trust is a per-SKU reliability grade that reflects how much of that SKU’s margin rests on measured evidence rather than approximation. It appears as a HIGH, MED or LOW badge in the Data Trust column on Catalog Profit, and again inside each SKU’s detail drawer.

It exists because a margin number and your confidence in that margin number are two different facts, and showing only the first is how operators get hurt. Discontinuing a SKU, repricing it, or pushing spend behind it are expensive moves. If the profit figure driving that decision was quietly assembled from a store-wide default cost, you deserve to know before you act, not after. Data Trust is the visible half of the Defensible Profit Model: MarginOS grades the evidence behind every cost, and Data Trust is that grading surfaced where you make decisions.

Before you start

• A connected Shopify store, so MarginOS has orders, products and the costs held against your variants.
• At least one cost source for the SKUs you sell — a per-unit cost you set, supplier or freight documents, or a connected 3PL. Data Trust grades the evidence you have supplied, so with none there is nothing to grade.
• Optionally, a connected 3PL or WMS inventory feed. Inventory freshness is one of the inputs, so a feed that stops updating will pull grades down.

Where to see Data Trust in MarginOS

1. Open Catalog Profit and find the Data Trust column. Every SKU carries a HIGH, MED or LOW badge, or -- if MarginOS has not graded it yet.
2. Hover any badge to see what that grade means for that specific SKU and whether the margin is safe to act on.
3. Click the Low Trust filter to show only the SKUs whose margins you should not trust yet. This is the fastest way to turn the grade into a to-do list.
4. Click into a SKU to open its detail drawer, where the Data Trust badge sits alongside the cost stack that produced it — so you can see which layer is weak, not just that something is.
5. Check your Margin Briefing for the data-quality signals behind low grades. Each one names the SKU, what is missing, and what to supply to fix it.

What each level means

How MarginOS grades Data Trust

MarginOS grades each SKU on two things: whether it has cost evidence specific to that SKU, and whether anything in its cost stack had to be approximated.

A SKU earns HIGH only when both hold — there is SKU-level evidence behind its costs, and nothing anywhere in the stack was estimated, defaulted or substituted. A single approximation is enough to drop it to MED. The grade reflects the weakest link rather than an average, because averaging would let one bad input hide behind several good ones. A SKU falls to LOW when the foundation itself is missing: no cost evidence of any kind, or an inventory snapshot stale enough that the stock and cover figures can no longer be trusted.

Things that pull a SKU down to MED include a store-wide default cost standing in for a real one, a fulfillment cost applied globally rather than per SKU, an inbound cost estimated because no document covered it, a product cost inferred from order lines rather than read from a source, and a unit cost that looks implausible against its peers. Each of these also raises a data-quality signal in your Margin Briefing naming the SKU and the fix — the grade tells you something is wrong, the signal tells you what.

One deliberate detail worth knowing: MarginOS decides whether cost evidence exists by presence, not by amount. A documented cost of $0.00 is real evidence and counts. A cost that is simply absent does not, and is never quietly treated as zero. This is the same stance that governs the rest of the cost model — an unknown cost is reported as unknown, because a zero cost turns an unprofitable product into a star.

Grades are recomputed as your data changes. Supply a missing cost, reconnect a stale feed, or send in the invoice MarginOS was waiting on, and the SKU is re-graded on the next reconciliation. You never have to clear an old grade by hand.

Example

Two SKUs show a similar margin on Catalog Profit, and their grades tell you those are not the same number.

SKU A — HIGH. Its unit cost comes from a supplier document, its fulfillment cost from a connected 3PL at the SKU level, and its stock count from an inventory feed that synced this morning. Nothing was estimated. The margin is as good as your source data, and you can price against it.

SKU B — MED. Same product-cost evidence, but no per-SKU fulfillment figure ever arrived, so MarginOS applied your store-wide pick-and-pack default instead. That single substitution is the whole difference. The margin is close enough to rank SKU B against its peers, but if this item’s real fulfillment cost runs above your store-wide average — heavy, bulky, or awkward to pack — its true margin is worse than shown. Supply the SKU-level cost and it re-grades to HIGH.

The practical read: a MED grade is not an error, it is a disclosure. It tells you how far to lean on the number, and what to supply to make it stronger.

FAQ

What does Data Trust actually measure?

Data Trust measures the quality of the evidence behind a SKU’s margin, not the size of the margin itself. A SKU can be highly profitable and graded LOW, or barely break even and be graded HIGH. The grade answers how much to believe the number, not whether the number is good news.

Why is my SKU showing MED instead of HIGH?

Something in its cost stack was approximated rather than measured. The most common causes are a store-wide default cost standing in for a SKU-specific one, a fulfillment cost applied globally instead of per SKU, or an inbound cost that had to be estimated. Your Margin Briefing names the specific SKU and the specific gap.

Why did my Data Trust drop when my costs did not change?

Inventory freshness is part of the grade. If an inventory feed stops updating, the stock counts behind the number go stale and MarginOS lowers the grade even though the costs are untouched. Reconnect or re-upload the feed and the grade recovers on the next reconciliation.

Does a LOW grade mean the margin is wrong?

Not necessarily — it means MarginOS cannot vouch for it. The figure may well be close. But it rests on estimates or is missing inputs, so it is not a number to make an irreversible decision on until the underlying data is fixed.

How do I improve a SKU’s Data Trust?

Give MarginOS SKU-specific evidence for whatever it is currently approximating: set a per-unit landed cost, send in the supplier or 3PL invoice that covers it, or connect the fulfillment source that reports costs per SKU rather than in aggregate. Then make sure your inventory feed is syncing. Grades update automatically as the evidence arrives.

Does a $0.00 cost count as missing data?

No. MarginOS distinguishes a cost that is genuinely zero from one that is absent. A documented $0.00 counts as real evidence; a missing cost is reported as missing and never silently treated as zero, because a zero cost would make an unprofitable product look like a winner.

Is Data Trust graded per SKU or across my whole store?

Per SKU, because cost evidence varies item by item. Use the Low Trust filter on Catalog Profit to see how much of your catalog is currently untrustworthy, and work the list down.

Related

• The Defensible Profit Model: how MarginOS grades cost accuracy
• How MarginOS calculates Gross Profit (CM1)
• Why a cost landed in Unmapped Costs
• Set your store-wide COGS, shipping and pick-pack fallbacks
• How It Works