Inventory Turns
- Formula
- Cost of goods sold divided by average inventory value
- Typical range
- 6-12 turns/year is typical for discrete manufacturers; under 4 usually means excess or slow-moving stock
- Direction
- Higher is better
Open the Calculator Generate the DAX
Inventory turns measure how many times a company sells through and replaces its average inventory over a period, showing whether working capital is converting into revenue or sitting on a shelf. A higher number means inventory dollars are moving faster; a lower one means cash is tied up in stock that isn’t selling.
What Good Looks Like
Discrete manufacturers commonly run 6 to 12 turns a year. Under 4 turns usually points to excess or slow-moving stock somewhere in the mix, and it’s worth checking whether the drag is raw material, WIP, or finished goods before assuming it’s a demand problem. On the other end, extremely high turns (20 or more) aren’t automatically good news either; they can signal a plant running so lean it’s exposed to stockouts and expedited freight every time demand ticks up.
The number is easy to flatter by using a single point-in-time inventory snapshot, usually taken at period end, instead of a true average. A plant that pushes hard to draw inventory down right before month-end close will post a great turns number that has nothing to do with how it operates the other 25 days of the month.
Inventory Turns in Power BI (DAX)
With a trailing COGS fact and a monthly inventory snapshot fact:
COGS (TTM) =
CALCULATE (
SUM ( fact_cogs[cogs_amount] ),
DATESINPERIOD ( dim_date[date], MIN ( MAX ( dim_date[date] ), TODAY () ), -12, MONTH )
)
-- Average of the month-end balances in the last 12 months. Each month's
-- snapshot date is found with every filter but the date cleared, so a part
-- with no stock that month counts as zero instead of dropping out.
-- MonthDates is taken first: CALCULATETABLE applies the month from the row,
-- and REMOVEFILTERS would otherwise clear that month along with the rest.
Average Inventory =
CALCULATE (
VAR Months =
FILTER (
ADDCOLUMNS (
VALUES ( dim_date[month] ),
"@Snapshot",
VAR MonthDates =
CALCULATETABLE ( VALUES ( dim_date[date] ) )
RETURN
CALCULATE (
MAX ( fact_inventory_snapshot[snapshot_date] ),
REMOVEFILTERS ( fact_inventory_snapshot ),
MonthDates
)
),
NOT ISBLANK ( [@Snapshot] )
)
RETURN
DIVIDE (
SUMX (
Months,
VAR Snapshot = [@Snapshot]
RETURN
CALCULATE (
SUM ( fact_inventory_snapshot[inventory_value] ),
fact_inventory_snapshot[snapshot_date] = Snapshot
)
),
COUNTROWS ( Months )
),
DATESINPERIOD ( dim_date[date], MIN ( MAX ( dim_date[date] ), TODAY () ), -12, MONTH )
)
Inventory Turns = DIVIDE ( [COGS (TTM)], [Average Inventory] )
Days of Inventory = DIVIDE ( 365, [Inventory Turns] )
Average Inventory takes the last snapshot in each month, then averages the 12 month-ends
in the same window as COGS (TTM), so the ratio holds on a monthly visual. A part with no stock in
a month counts as zero rather than dropping out of its average, so part rows add up to the
total; Inventory snapshots shows the difference on demo data.
Without that window, a month on the axis divides 12 months of COGS by one month’s snapshot. Averaging
several points across the year is far harder to game than any single snapshot date.
dim_date[month] needs one value per month of each year, such as 2026-03, and dim_date
should stop at the last month you have loaded, or the window runs past your data.
In the middle of a month the 12-month window covers 13 calendar months, so the average
takes 12 month-ends plus the latest snapshot of the current month. To average exactly 12
closed months, use MIN ( MAX ( dim_date[date] ), EOMONTH ( TODAY (), -1 ) ) as the end date in both Average Inventory and COGS (TTM), so the two windows still match.
Common Mistakes
- Using a period-end snapshot instead of an average. A single low balance at month close can flatter the ratio without reflecting how inventory actually ran all month.
- Blending raw material, WIP, and finished goods into one number. Each has a different target and a different owner, and a blended turns figure obscures which one is actually the problem.
- Comparing turns across product lines with different margins or lead times without normalizing. A build-to-order line and a distribution-heavy line will never share the same reasonable benchmark.
Inventory Turns Calculator
Runs in your browserResult
DAX Generator
Use My Own Table and Column NamesUpdates as You Type
COGS (TTM) =
CALCULATE (
SUM ( fact_cogs[cogs_amount] ),
DATESINPERIOD ( dim_date[date], MIN ( MAX ( dim_date[date] ), TODAY () ), -12, MONTH )
)
-- Average of the month-end balances in the last 12 months. Each month's
-- snapshot date is found with every filter but the date cleared, so a part
-- with no stock that month counts as zero instead of dropping out.
-- MonthDates is taken first: CALCULATETABLE applies the month from the row,
-- and REMOVEFILTERS would otherwise clear that month along with the rest.
Average Inventory =
CALCULATE (
VAR Months =
FILTER (
ADDCOLUMNS (
VALUES ( dim_date[month] ),
"@Snapshot",
VAR MonthDates =
CALCULATETABLE ( VALUES ( dim_date[date] ) )
RETURN
CALCULATE (
MAX ( fact_inventory_snapshot[snapshot_date] ),
REMOVEFILTERS ( fact_inventory_snapshot ),
MonthDates
)
),
NOT ISBLANK ( [@Snapshot] )
)
RETURN
DIVIDE (
SUMX (
Months,
VAR Snapshot = [@Snapshot]
RETURN
CALCULATE (
SUM ( fact_inventory_snapshot[inventory_value] ),
fact_inventory_snapshot[snapshot_date] = Snapshot
)
),
COUNTROWS ( Months )
),
DATESINPERIOD ( dim_date[date], MIN ( MAX ( dim_date[date] ), TODAY () ), -12, MONTH )
)
Inventory Turns = DIVIDE ( [COGS (TTM)], [Average Inventory] )
Days of Inventory = DIVIDE ( 365, [Inventory Turns] )