High-Low Method: Calculator, Formula and Worked Example

The high-low method estimates the fixed and variable components of a mixed cost from two observations: the highest activity level and the lowest activity level, together with each observation's total cost. It is a quick preliminary estimate, not proof of a causal relationship or a substitute for reviewing all your records.

For an owner planning costs, the useful output is an explicit equation: estimated total cost = fixed component + variable rate × activity. That can help prepare a budget or improve the cost inputs used in break-even analysis. The estimate is only as useful as the activity measure and observations behind it.

High-low method formula

Variable cost per activity unit = (cost at highest activity − cost at lowest activity) ÷ (highest activity − lowest activity). Then calculate fixed component = either selected total cost − variable rate × that observation's activity. Both selected observations should yield the same fixed component before rounding.

Choose the extremes by activity, not by total cost. A month with a repair bill can have the highest cost without the highest activity. The rate's units matter: dollars per visit, dollars per machine-hour and dollars per mile are different inputs. Keep the cost scope and time period consistent.

A six-month worked example

A fictional service business reviews one monthly mixed vehicle-and-logistics cost pool against completed equivalent visits. The invented USD observations below illustrate the method; they are not expected industry costs. The business must first decide whether visits are a reasonable driver, or whether measured mileage would explain this pool better.

Use month F and month A: ($2,440 − $1,640) ÷ (280 − 80) = $800 ÷ 200 = $4 per visit. The fixed component is $2,440 − ($4 × 280) = $1,320 per month. Check the low point: $1,640 − ($4 × 80) also equals $1,320.

The resulting model is estimated monthly cost = $1,320 + $4 × visits. At 220 equivalent visits, the estimate is $1,320 + $880 = $2,200. A target of 220 lies inside the observed 80–280 activity range, but that alone does not prove the estimate is reliable.

Test the model against observations you did not select

At 240 visits the equation predicts $2,280, whereas month E records $3,000: an unexplained $720 difference. Do not silently discard it. Check repairs, fuel prices, route distances, recording errors, overtime or a different mix of jobs. The example does not establish which explanation is correct. At 160 visits it predicts $1,960 versus $1,980 actual, a $20 difference.

A model that exactly fits its two chosen observations can still fit the business poorly. Plot all months and inspect the residuals before using the estimate for an important decision. If the relationship is not approximately linear, use a better driver, a separated cost schedule or an appropriate multi-observation regression rather than claiming the two-point line is accurate.

Use the calculator above with your own records

1. Collect comparable activity and total-cost records for the same monthly cost pool. Keep annual and monthly amounts separate.
2. Identify the minimum and maximum activity observations from the full dataset. Enter the cost paired with each observation, even when another month has a higher total bill.
3. Check the variable rate, fixed component and target-period estimate. Investigate a negative rate or negative fixed component rather than assuming it is a real saving.
4. Compare the equation with intermediate observations. Rebuild it when contracts, prices, staffing or the service mix changes.

The calculator accepts the two observations you have already selected. It does not inspect your full dataset or identify its extremes automatically. Its default entries reproduce this guide: low activity 80, low cost 1,640, high activity 280, high cost 2,440 and target 220. Currency labels do not convert exchange rates.

High-low method in Excel or Google Sheets

Put low activity in B2, its cost in C2, high activity in B3, its cost in C3 and target activity in B4. Calculate the rate with =(C3-C2)/(B3-B2) in B6, fixed cost with =C2-B6*B2 in B7 and the estimate with =B7+B6*B4 in B8. Require complete numeric inputs and B3 greater than B2 before evaluating these formulas.

Do not use MAX on the cost column to choose the high observation. For a full dataset, select the row with maximum activity and the row with minimum activity, preserving the corresponding costs. If several months tie at the same extreme activity but have different costs, investigate the differences and document a defensible treatment; the two-point method does not decide it for you.

How this changes your break-even inputs

If the selected cost pool truly consists of $1,320 monthly fixed cost plus $4 per completed visit, include $1,320 in fixed overhead and $4 in variable cost per visit. Do not also add the full vehicle cost as another monthly expense. The remaining business costs still need to be included.

For an independent fictional example, assume $100 price, $56 other variable costs, and $6,680 other monthly fixed costs. Adding this pool gives variable cost $60, contribution $40 and fixed costs $8,000. Operating break-even is 200 visits. This calculation assumes the entire combined cost model remains valid at 200 visits and does not forecast demand. See fixed, variable and mixed costs, the contribution income statement and the break-even calculation guide.

When not to rely on high-low

• Abnormal endpoints: a shutdown, unusual repair or promotion can distort either selected point.
• Changing prices or methods: inflation, supplier changes, a new route or new equipment can mean observations no longer describe the same relationship.
• Step costs: another vehicle or supervisor changes the fixed component. One straight line is then misleading.
• Mixed activity: a short visit and a complex installation are not equivalent just because both are recorded as one job.
• Extrapolation: a target outside the observed activity interval has no endpoint support. Even an inside-range estimate needs the other checks.

For high-stakes pricing or staffing, reconcile actual records and compare more observations. A negative fixed component can be a symptom of a poor driver or changing cost behavior; it is not negative rent. A zero rate is mathematically possible, but means the selected points provide no evidence of a positive variable component.

Method and source

OpenStax's fixed and variable cost equation lesson explains selecting the highest and lowest activity levels and using the associated costs. This article's six-month records and business scenarios are original fictional examples.