"Sharon, what are the 3 KPIs I should track to catch inventory data drift?"
I hear this at least once a week, usually from owners and operations leaders who have realized their inventory numbers do not match reality. They want a dashboard. 3 metrics. A simple alert.
There are no KPIs that catch inventory data failures early. By the time a KPI shows something is off, it is not catching a problem. It is reading the aftermath.
The warning sign that arrives after the damage
The pattern repeats. Revenue is growing. Gross margin looks healthy. The dashboard is green. Then operating cash flow does not move, or it declines while revenue climbs.
That combination is the tell: revenue up, margin up, cash not following. Most teams explain it as timing. "Next quarter will normalize." It usually does not, because cash has been slipping for months into inventory that looks profitable on paper and does not convert back into cash.
A researcher who builds data algorithms asked me the same question: which 2 or 3 KPIs would detect an early failure between inventory and finance data? My answer surprised him.
"The issue isn't which KPIs to watch. The issue is that once a KPI reflects the problem, inventory has already been steering financial decisions for months."
Gross margin can look fine while the cost basis under it is stale, incomplete or in the wrong month. The KPI is not lying. It is reporting fiction with perfect consistency.
What was found, on 5 products
A beauty brand I worked with had healthy turnover and a green dashboard. We traced 5 products from purchase order to financial statement and found $147K in costs recorded incorrectly: wrong month, missing components, some counted twice.
The KPI said "healthy." The cash position told the truth.
Where to look instead
The real question is not "which KPI should I watch?" It is: where did the data and the physical inventory stop matching?
Inventory has a physical journey: ordered, arrives, received, stored, picked, shipped. The data has a parallel one: purchase order created, invoice posted, costs recorded, cost of goods sold recognized, margin reported. When the two drift apart, reports still reconcile while the cash protection disappears.
First confirm the gap exists. The simplest signal is COGS variance: what the system thinks things cost against what was paid, fully loaded. 3 quick tests:
- Spot-check a few fast-moving products. Compare the system cost to what was paid, fully loaded.
- If shelves are empty and the system says stock is healthy, the records and the warehouse are no longer describing the same business.
- If the forecast says 100,000 units are needed and 60,000 are on hand, the purchase plan should show 40,000. If it shows 70,000, that is not a planning miss. That is broken logic upstream.
The 5 places data and reality separate
With the gap confirmed, the break is in one of 5 places.
- Data hygiene. Costs hit the wrong period, landed costs never fully attach to the product, and small manual fixes stack up.
- Workflow discipline. Placeholder codes, manual overrides and "we'll clean it up later" entries become the system.
- Data flow. Sales, finance and operations each produce a credible number that disagrees with the others, from different sources at different times.
- System logic. The rules behind cost, valuation and reorder were built for an older version of the business. The system stays consistent while the business changes around it.
- Decision logic. Reorders and financing move quickly on inputs nobody has verified. The warehouse knows what moved. Finance knows what posted. The risk lives in the space between them.
The researcher put it well: "Inventory issues are frequently data and mechanism problems, not just execution issues."
The fix: follow the cost story
Pick a small set of top products and follow the cost from commitment to recognition. The search is not for one mistake. It is for the moment the dates, costs and assumptions stop lining up. That moment is the place to fix, and it is where the decision can still be held.
If COGS variance is growing, or physical counts will not reconcile cleanly to valuation, it is not a KPI problem. It is an inventory-to-finance data problem, and that is a different conversation. The free adjustment test takes 12 numbers and says whether your month-end adjustment is noise or a pattern.
Before the next large commitment
A major purchase order, a change in buying strategy, borrowing to fund inventory: that is the moment to slow the decision down. Not to debate dashboards. To confirm the numbers describe the same reality the warehouse and the bank account are living in, so the number can be defended when someone outside asks.
Is this your number?
If your close balances and the number still moves, this is the work: a month-end figure that matches what actually happened.