A mixed cost — also called a semi-variable cost — has a part you pay whatever the volume and a part that grows with activity. A shipping department with a fixed lease on its packing area and per-order labor and supplies is a typical case. Break-even and budgeting need the two parts separately, but your bills show only one total each month. The high-low method is the simplest way to pull them apart using nothing but your own records of activity and total cost.
Total cost (Y) = Fixed cost (a) + Variable rate (b) × Activity (X)
Four steps
-
Pick the high and low activity periods
Choose the months with the highest and lowest activity — orders, units, machine hours — not the highest and lowest cost.
-
Find the variable rate
Divide the change in cost by the change in activity: (cost at high − cost at low) ÷ (high activity − low activity).
-
Find the fixed part
Take the total cost at either point and subtract the variable rate × that activity. Both points give the same answer.
-
Write the cost equation
Y = a + bX predicts the cost at any activity level inside the range you observed.
Worked example: a shipping department
Six months of hypothetical records are below. The highest activity is April with 1,200 orders and the lowest is May with 400. March has the highest cost, but it is not the highest-activity month, so it is not used: the method is anchored to activity, because activity is what drives the variable part.
| Month | Orders | Total cost | Predicted by Y = $2,600 + $6X | Difference |
|---|---|---|---|---|
| January | 620 | $6,520 | $6,320 | +$200 |
| February | 480 | $5,900 | $5,480 | +$420 |
| March | 900 | $9,950 | $8,000 | +$1,950 |
| April (high) | 1,200 | $9,800 | $9,800 | $0 |
| May (low) | 400 | $5,000 | $5,000 | $0 |
| June | 750 | $7,350 | $7,100 | +$250 |
Variable rate: ($9,800 − $5,000) ÷ (1,200 − 400) = $4,800 ÷ 800 = $6 per order. Fixed part: $9,800 − $6 × 1,200 = $2,600 a month, and the low point confirms it: $5,000 − $6 × 400 = $2,600. The cost equation is Y = $2,600 + $6X, so a month with 900 orders is planned at $2,600 + $5,400 = $8,000.
Check the estimate
The method uses only two points, so compare its predictions with the months in between. Here January, February and June come in $200 to $420 above the line: the two extreme months may sit slightly low, and the true fixed part could be a little higher than $2,600. March is $1,950 above the line — when you look up the records you would expect to find something unusual, such as a one-off repair. A month like that should not be chosen as a high or low point, and it should not be built into the equation either.
Three limits matter. First, if the high or low month was unusual, the whole equation inherits the distortion; pick the next most extreme normal month instead. Second, the method assumes a straight-line relation between cost and activity, which a quick scatter plot of all months can confirm or reject. Third, the equation holds only inside the activity range you observed: forecasting 2,000 orders from data between 400 and 1,200 is an extrapolation, and a step cost such as extra packing space may appear on the way. When you have many months of data, least-squares regression uses all of them and is more accurate.
Once you have the split, the fixed part joins your other fixed costs and the $6 rate joins the variable cost per order — exactly the inputs a break-even or contribution calculation needs.
Method reference: OpenStax: Estimate a Variable and Fixed Cost Equation and Predict Future Costs. Checked October 26, 2026. All records are hypothetical.
Related calculators & guides
Report an error
Thank you! Your feedback helps us make this page better.
Could not send. Please try again.