What the High-Low Method Actually Splits
Most business costs refuse to sit still. A power bill, a delivery invoice, or a maintenance account moves with production, yet never starts from zero — a base charge sits underneath every period. Cost accountants call these mixed costs, or semi-variable costs, and the high-low method is the fastest pencil-and-paper way to pull them apart into a fixed floor and a variable slope.
The method needs exactly two observations: the period with the highest activity and the period with the lowest. Because the fixed component is by definition identical in both, the entire cost difference between them must come from the variable portion. Divide the cost gap by the activity gap and the fixed cost cancels out of the arithmetic completely.
The payoff is a usable cost equation: total cost = fixed cost + variable rate × activity. Once the split is on paper you can quote a job, build a flexible budget line, or feed the variable rate into break even calculator math. This tool runs the arithmetic and returns all three numbers at once: the slope, the floor, and the projection you asked for.
The Formula, Piece by Piece
Variable cost per unit = (cost at high activity − cost at low activity) ÷ (units at high activity − units at low activity). On the defaults, ($9,000 − $4,000) ÷ (1,200 − 400) = $5,000 ÷ 800 = $6.25 per unit. That single division is the variable rate — no regression package, no curve fitting, and it works identically in dollars, rupees, or euros.
Fixed cost then falls out of either endpoint: high cost − (variable rate × high units) = $9,000 − ($6.25 × 1,200) = $1,500 per period. Verify against the low point and you get the same answer: $4,000 − ($6.25 × 400) = $1,500. If the two checks ever disagree, one of the four inputs contains a typo — the method is internally self-checking that way.
The finished model reads total cost = $1,500 + $6.25 × units, the classic y = a + bx line from cost accounting textbooks. Every projection, including the one this calculator prints, is that line evaluated at your target: $1,500 + ($6.25 × 1,000) = $7,750. Managers chasing the revenue side of the same unit economics can pair the split with the contribution margin calculator.
A Full Worked Example: Utility Cost by Machine Hour
A machine shop logs twelve months of power invoices. The busiest month ran 1,000 machine hours and billed $6,700; the slowest ran 300 hours and billed $3,200. Slope: ($6,700 − $3,200) ÷ (1,000 − 300) = $5.00 per machine hour. Fixed piece: $6,700 − ($5.00 × 1,000) = $1,700 per month. Two divisions, one complete cost equation.
Projection time. A rush order next month will consume 800 hours, so expected power cost is $1,700 + ($5.00 × 800) = $5,700. Notice what the fixed piece does to unit cost along the way: at 300 hours the bill averaged $10.67 per hour; at 1,000 hours only $6.70. Volume looked cheaper purely because the fixed meter charge spread across more hours.
That per-unit slide is arithmetic, and it is why average cost claims need a volume label attached. The same effect is what the average fixed cost calculator isolates from the other direction — hold the fixed pool constant and watch the per-unit share collapse as output grows. Variable cost, by contrast, never dilutes: $5.00 an hour is $5.00 at any volume.
The Method's Blind Spots and the Negative-Fixed Warning
The high-low method uses exactly two of your twelve months and discards the other ten. If either chosen month hides a one-off — a storm surcharge, a machine breakdown, a billing correction — the slope inherits the damage directly. The method has no averaging effect and no internal way to notice the distortion. That is the price of its speed.
Concrete sensitivity, still in the shop example: replace the storm-inflated 1,000-hour month with the second-busiest, 800 hours at $5,900. The slope jumps from $5.00 to $5.40 per hour, fixed drops from $1,700 to $1,580, and the 800-hour projection moves from $5,700 to $5,900 — a $200 swing from swapping a single data point.
A negative fixed cost is the loudest possible warning. It means total cost fell faster than activity did, which no straight line can produce, so at least one point is distorted or the cost is genuinely curved. Treat any negative-fixed output as unusable until the inputs are reconciled against the ledger, and only then roll the split into a budget calculator.
High-Low vs Scattergraph and Least-Squares Regression
The scattergraph method plots every period and fits a line by eye; least-squares regression fits it mathematically using all n points and reports how well the line explains the data. High-low sits at the extreme end of simplicity: two points, zero judgment, thirty seconds. For a quick estimate between formal runs, that trade is often worth making.
The statistical cost is real. Regression returns an R² statistic — on well-behaved utility data often between 0.6 and 0.9 — telling you what share of cost variation the line captures. High-low returns no goodness-of-fit signal at all: any two points define a perfect line, even when the ten ignored months contradict it loudly.
A practical middle ground: run high-low on the true extremes, then again on the second-highest and second-lowest months. If the two slopes land within about 10% of each other — the shop example produced $5.00 versus $5.40, an 8% gap — the cost is close enough to linear for budgeting. Then test how the split moves profit with the degree of operating leverage calculator.
Choosing Valid High and Low Points
The rule people get wrong: select on activity, never on cost. In most datasets the most active month also costs the most, so the two coincide — until a cheap surge month (subcontracted labor, donated materials) or an expensive idle month (minimum-order fees, storm charges) breaks the pattern. Rank periods by units, hours, or miles first, then take the costs that go with them.
Any consistent driver works: machine hours, patient days, truck miles, orders shipped, rooms cleaned. The variable rate simply reads in dollars per driver unit — $5.00 per machine hour, or $0.80 per mile in a fleet. One discipline matters throughout: the cost and the driver must share period boundaries, or the slope quietly absorbs the timing mismatch.
Screen both chosen periods for one-off events before committing. Once the split holds up, the variable rate becomes your walk-away floor for quoting — price below $6.25 a unit and every sale destroys cash. Applying floor-plus-margin logic with the unit price calculator turns the split into a defensible price list instead of a guess.
Reading the Output: Rate, Floor, and Projection Together
Three numbers, three jobs. The variable rate ($6.25 on the defaults) belongs in pricing floors and contribution math: every additional unit sold adds exactly this much cost. The fixed floor ($1,500 per period) belongs in period budgeting — it arrives whether the machines run or not, and it annualizes to $18,000 of committed spend.
The projection is the equation applied forward. At the default 1,000 units the model prints $7,750, which works out to $7.75 per unit at that volume — yet at 400 units the same equation says $10.00 per unit, and at 2,000 units $7.00. Watching per-unit cost move while nothing about efficiency changed is the clearest way to internalize fixed-cost spreading.
The two pieces also join up in break even math. Price at $10 and contribution is $10 − $6.25 = $3.75 per unit, so the operation needs $1,500 ÷ $3.75 = 400 units a month to cover the fixed floor. Before quoting, layer a target margin onto the $6.25 variable floor with the markup calculator and the price floor defends itself.
Feeding the Split Into Pricing, Budgets, and Cash Plans
Flexible budgets write themselves once the equation exists: budgeted power at any activity level is literally $1,700 + $5.00 × hours in the shop example. Run separate high-low splits per account — power, maintenance, supplies — each with its own driver, then consolidate the fixed floors and variable rates in a cost of doing business calculator for the full overhead picture.
On the income side, the split tells you which costs scale and which ones wait. Products priced above the variable rate contribute immediately; the fixed floor decides when profit starts. To see the split land on the bottom line, push a projected month through the accounting profit calculator — revenue minus the projected total cost tells the whole story.
Cash is where fixed floors bite hardest. In a quiet 300-hour month the shop still wires out $1,700 before the first unit ships, which is exactly the expense a runway analysis exists to catch — the burn rate calculator converts a $1,700 monthly floor into weeks of survival at any cash balance. Seasonal businesses should hold that cushion before the slow season, when the method's low point arrives on schedule.