Use the high-low method to separate mixed costs into fixed and variable components, write the cost equation, choose an activity basis.
Last updated
Estimate mixed-cost behaviour from two observations Use the highest and lowest activity points to estimate variable cost per activity unit, fixed cost, the mixed-cost equation, and whether the forecast sits inside the observed relevant range.
Display currency
Switch the currency for cost rows before entering high-low method costs. Activity counts and the slope calculation stay unchanged.
Example scenarios
Assumptions
The high-low method uses only the highest and lowest activity observations. It assumes cost behaviour is roughly linear inside that range and ignores any intermediate outliers.
Result
$6.00 variable cost per unit produced
The high-low estimate implies a mixed-cost equation of $1,200.00 fixed cost plus $6.00 for each unit produced.
Estimated fixed cost
$1,200.00
Activity spread
1,800
Cost spread
$10,800.00
Forecast total cost
$15,600.00
Forecast variable cost
$14,400.00
Fixed-cost share at forecast
7.69%
Cost per 100 activity units
$600.00
Midpoint cost estimate
$15,000.00
Forecast stays inside the observed relevant range This forecast is interpolation between the low and high observations, so the estimate is usually more defensible than a forecast pushed beyond the observed range.
Cost equation summary
Estimated mixed-cost equation
$1,200.00 + $6.00 × activity
Forecast cost change from low observation
$6,000.00
Forecast cost change from high observation
-$4,800.00
Forecast range status
Interpolation
Observed-range planner
Point
Activity (unit produced)
Variable cost
Total cost
Use
Lowest observation
1,400
$8,400.00
$9,600.00
Observed low-activity anchor used in the high-low estimate.
Midpoint of observed range
2,300
$13,800.00
$15,000.00
Interpolated cost halfway between the observed low and high activity levels.
Highest observation
3,200
$19,200.00
$20,400.00
Observed high-activity anchor used in the high-low estimate.
Forecast activity
2,400
$14,400.00
$15,600.00
Forecast sits inside the observed activity range, so the estimate is interpolation.
Interpretation note
The variable-cost estimate comes from the slope between the two selected observations. Estimated fixed cost is the remainder after backing variable cost out of either observation. If the high or low point is abnormal, a scattergraph or regression usually gives a more reliable cost equation.
High-low method calculator guide: mixed cost equation, relevant range
A high-low method calculator separates a mixed cost into estimated variable and fixed components by using the highest and lowest activity observations in a dataset.
What the high-low method measures
The high-low method estimates how much of a mixed cost changes with activity and how much stays fixed. Instead of using every data point, it looks only at the highest activity level and the lowest activity level, then treats the difference between those two observations as the slope of the mixed-cost line.
That simplicity makes the high-low method popular in managerial accounting, teaching, and quick planning. It also creates the main risk: if the high or low point is unusual, the entire cost equation can be pulled away from what the rest of the dataset suggests.
Choose the highest and lowest activity points, not just the highest and lowest costs
A common mistake is to sort the data by total cost and then choose the biggest and smallest cost rows. The high-low method is based on activity level, so the high point should be the period with the most units, machine hours, service calls, labour hours, or other cost driver being analysed. The low point should be the period with the least activity.
This distinction matters because a high-cost month is not always the highest-activity month. A repair event, rush shipping, overtime premium, or unusual supplier invoice can make a moderate-activity month look expensive. If that distorted month becomes one of the two selected endpoints, the variable cost per unit and fixed-cost estimate can both become misleading.
High-low method formula and cost equation
The first step is to find the change in total cost between the high and low activity observations. Divide that cost change by the change in activity to estimate variable cost per activity unit. Then subtract the variable portion from either the high observation or the low observation to estimate fixed cost.
Once you have those two pieces, you can write a simple mixed-cost equation. That equation is the part managers actually use because it turns historical observations into a forecast model for another activity level.
Variable cost per activity unit = (Highest cost - Lowest cost) / (Highest activity - Lowest activity)
This estimates the slope of the mixed-cost line using the two selected observations.
Estimated fixed cost = Total cost - (Variable cost per activity unit x Activity level)
Subtracting the variable portion from one observation leaves the fixed-cost component.
Forecast total cost = Estimated fixed cost + (Variable cost per activity unit x Forecast activity)
This applies the estimated mixed-cost equation to another activity level.
Worked example: from the high and low points to a forecast
Suppose the highest activity level is 3,200 units with total cost of 20,400 and the lowest activity level is 1,400 units with total cost of 9,600. The activity spread is 1,800 units and the cost spread is 10,800, so the estimated variable cost is 6.00 per activity unit.
Backing that slope out of the high observation leaves estimated fixed cost of 1,200. If forecast activity is 2,400 units, forecast variable cost is 14,400 and forecast total cost is 15,600. The estimated mixed-cost equation is therefore 1,200 plus 6.00 times activity.
Relevant range: interpolation versus extrapolation
A high-low estimate is usually easier to defend when the forecast sits between the observed low and high activity levels. That is interpolation inside the observed relevant range. Forecasts outside that span are extrapolation, which increases the chance that overtime, step costs, supplier discounts, idle-capacity effects, or other non-linear behaviour will break the estimate.
This is why the calculator now shows whether the forecast sits inside the observed range or outside it. A manager looking at a cost forecast for 4,200 units based on observations between 1,400 and 3,200 units should not treat that answer the same way as a forecast for 2,400 units that sits comfortably inside the historical span.
How to read the calculator output
The headline variable cost per activity unit is the slope of the high-low cost equation. The estimated fixed cost is the intercept, or the part of the mixed cost that remains after the variable cost is removed from one selected observation. Together they form the cost-volume equation used for the forecast.
The observed-range planner is designed to make the estimate easier to audit. It shows the low anchor, midpoint, high anchor, and forecast activity side by side, so you can see whether the forecast is interpolation or extrapolation and whether the total cost estimate moves in a sensible pattern across the activity range.
When the high-low method works well and when it does not
The method works best when cost behaviour is approximately linear and the high and low observations are normal operating periods. It is especially useful for preliminary budgets, teaching examples, and first-pass cost behaviour analysis when time or data quality is limited.
It works less well when the high or low point is distorted by a one-off event, when the cost has strong step-fixed behaviour, or when multiple cost drivers are changing at once. In those situations, the high-low method is still a useful starting point, but not a final decision tool.
High-low method versus regression analysis
The core weakness of the high-low method is that it ignores the middle of the dataset. Regression analysis uses the full data series, which can make it more reliable when you have enough clean observations and the decision matters enough to justify the extra work.
That does not make the high-low method useless. It just means the method belongs in the quick-estimate lane. If the page helps you spot that your forecast is outside the observed range or that an extreme point might be driving the answer, it has already done something useful that many simple calculators fail to surface.
It separates a mixed cost into estimated variable and fixed components by using the highest and lowest activity observations, then applies the resulting cost equation to a forecast activity level.
How do you calculate variable cost per unit with the high-low method?
Subtract the lowest total cost from the highest total cost, then divide by the difference between the highest and lowest activity levels. That gives the estimated variable cost per activity unit.
How do you find fixed cost after using the high-low method?
Multiply the variable cost per unit by either the high or low activity level, then subtract that variable portion from the matching total cost observation. The remainder is estimated fixed cost.
What is the biggest weakness of the high-low method?
It uses only two observations and ignores the rest of the dataset. If the high or low point is unusual, the resulting fixed-cost and variable-cost estimates can be misleading.
Should the high-low method use the highest cost or highest activity?
Use the highest activity level and the lowest activity level. The costs attached to those activity levels are then used in the formula. Choosing the highest and lowest total costs instead can select the wrong periods if a month had an unusual repair, overtime spike, discount, or one-off invoice.
What activity unit should I enter?
Use the cost driver that best explains the mixed cost. Production units may fit manufacturing materials or variable overhead, machine hours may fit equipment-related costs, labour hours may fit labour-driven overhead, and service calls may fit field-support costs. The label does not change the formula, but it makes the result easier to interpret.
Why does the calculator show interpolation versus extrapolation?
Because a forecast inside the observed activity range is generally safer than one outside it. Forecasting beyond the historical range increases the risk that non-linear cost behaviour or step costs make the high-low equation less reliable.
When should I use regression instead of the high-low method?
Use regression when you have enough reliable data points, the decision is material, and you want the estimate to reflect the full dataset rather than only the two extreme observations.
Can the high-low method still be useful if it is not perfect?
Yes. It is often a good first-pass planning tool for teaching, quick budgets, and early pricing analysis. The important thing is to treat it as an approximation and escalate to deeper analysis when the stakes or data complexity justify it.
