Your electricity bill was $1,180 in March and $1,600 in April. Your production went up — but did the whole $420 increase come from running the ovens more, or is part of your bill fixed no matter what you bake? Until you can answer that question, you cannot build an honest budget, set prices that actually cover your costs, or figure out your break-even point. The high-low method gives you the answer with nothing more than two months of data and fourth-grade arithmetic.
This guide walks through what the high-low method is, the exact formula, a worked example you can follow line by line, the mistakes that quietly break it, and when you should graduate to a more powerful technique.
What the High-Low Method Actually Does
Most of the costs in your business fall into three buckets:
- Fixed costs stay flat no matter how busy you are: rent, insurance premiums, the base fee on your internet plan.
- Variable costs move in lockstep with activity: raw ingredients per unit sold, credit-card processing fees per dollar of revenue, mileage reimbursement per mile driven.
- Mixed costs (also called semi-variable costs) have one foot in each camp: a fixed base charge plus a variable piece that rises with usage.
Mixed costs are everywhere. Your electric bill has a fixed service charge plus a per-kilowatt-hour rate. Your delivery van costs a fixed insurance premium plus fuel and wear that scale with miles. Your sales rep earns a base salary plus commission. Your phone plan charges a base rate plus overage or per-line fees.
The high-low method separates a mixed cost into its fixed and variable components by comparing just two data points: the period with the highest activity level and the period with the lowest activity level. The logic is simple. Fixed costs do not change between the two periods, so any difference in total cost must come entirely from the variable piece. Divide that cost difference by the activity difference and you have your variable cost per unit. Plug that back in once, and the fixed piece drops out.
The Formula in Three Steps
You need two things for each period: the activity level (units produced, miles driven, machine hours, sales dollars — whatever drives the cost) and the total mixed cost. Then:
Step 1: Compute the variable cost per unit of activity.
Variable rate = (Cost at highest activity − Cost at lowest activity) ÷ (Highest activity − Lowest activity)
Step 2: Compute the fixed cost using either point.
Fixed cost = Total cost − (Variable rate × Activity level)
Run this on the high point or the low point — both give the same answer, and checking both is a free error detector.
Step 3: Write your cost equation.
Total cost = Fixed cost + (Variable rate × Activity level)
Accountants write this as Y = a + bX, where Y is total cost, a is fixed cost, b is the variable rate, and X is activity. Once you have the equation, you can forecast the cost at any activity level inside your normal operating range.
A Worked Example: The Bakery's Electric Bill
Say you run a small bakery and want to understand your electricity cost. You pull six months of loaves baked and electric bills:
| Month | Loaves baked | Electric bill |
|---|---|---|
| Jan | 8,000 | $1,240 |
| Feb | 10,500 | $1,455 |
| Mar | 7,200 | $1,180 |
| Apr | 12,000 | $1,600 |
| May | 9,400 | $1,372 |
| Jun | 11,300 | $1,543 |
Step 1: Identify the high and low activity points. The highest activity is April (12,000 loaves, $1,600). The lowest is March (7,200 loaves, $1,180).
Variable rate = ($1,600 − $1,180) ÷ (12,000 − 7,200) = $420 ÷ 4,800 = $0.0875 per loaf — about 8.75 cents of electricity per loaf.
Step 2: Solve for the fixed cost. Using the April point:
Fixed cost = $1,600 − ($0.0875 × 12,000) = $1,600 − $1,050 = $550.
Verify with March: $1,180 − ($0.0875 × 7,200) = $1,180 − $630 = $550. Same answer — the math checks out. That $550 is your service charges, lighting, refrigeration baseline, and everything else that runs whether you bake or not.
Step 3: Write the equation. Monthly electricity cost = $550 + ($0.0875 × loaves baked).
Now you can forecast. Planning a 13,000-loaf July? Budget $550 + ($0.0875 × 13,000) = $1,687.50. Quoting a wholesale contract that adds 2,000 loaves a month? The incremental electricity cost is just 2,000 × $0.0875 = $175 — the $550 fixed piece does not budge, so it should not factor into the marginal price.
Four Mistakes That Quietly Break the Method
The high-low method is easy to compute and easy to misuse. These four errors account for most bad results.
1. Picking high and low by cost instead of by activity
This is the single most common mistake. The method requires the periods with the highest and lowest activity, not the highest and lowest cost. Most of the time they coincide — but when they do not, using the cost extremes silently corrupts everything.
Imagine February's bill spiked because the utility added a one-time meter fee, making it your highest-cost month even though production was middling. If you anchor on February, the "difference in cost" you attribute to variable activity includes a fee that had nothing to do with activity, and your variable rate comes out too high. Always sort by the activity column first.
2. Letting an outlier anchor the whole estimate
The method uses exactly two data points and ignores the rest, which makes it fragile. If your highest-activity month was the freak holiday rush when you rented a second oven and paid emergency repair rates, your variable rate inherits all of that weirdness.
Before computing, eyeball your data or plot it. If the high or low point looks wildly out of line with the rest — a strike month, a shutdown week, a billing error — either adjust it, drop that period and use the next-most-extreme one, or switch to a method that uses all your data (more on that below). A quick scatter plot takes thirty seconds and catches what a table hides.
3. Forecasting outside the relevant range
Your cost equation is only trustworthy inside the activity range you measured — accountants call this the relevant range. Our bakery equation came from months of 7,200 to 12,000 loaves. Forecasting 13,000 loaves is a small, reasonable step beyond the data. Forecasting 30,000 loaves is fantasy: at some point you would need a second oven, a bigger electrical panel, or a new location, and the fixed cost would jump to a new plateau.
Any forecast far above your high point or below your low point deserves a big asterisk. If you are planning a major expansion, rebuild the estimate from data that resembles the future operation, not the past one.
4. Mixing periods where the underlying rates changed
The method assumes the fixed cost and the per-unit variable rate stayed constant across both periods. If the utility raised its per-kWh rate in April, or you renegotiated your phone plan mid-year, the cost difference between March and April reflects the rate change as well as the activity change, and the formula cannot tell them apart.
Whenever prices, wage rates, contracts, or processes changed between your two anchor periods, either pick anchors on the same side of the change or adjust the older figures to current rates first. Stale data plus new rates equals a wrong answer delivered with false precision.
When High-Low Is Enough — and When to Upgrade
The high-low method's strength is speed: two data points, one division, one subtraction, done. That makes it ideal for quick what-if estimates, sanity-checking a budget line, smaller businesses with limited history, and any situation where a rough split beats no split at all.
But ignoring all but two observations is also its weakness. Two upgrades worth knowing:
The scattergraph method. Plot every period's activity and cost as a dot, then draw the line that visually fits the cloud best. The line's intercept with the cost axis is your fixed-cost estimate; its slope is the variable rate. It uses all your data and makes outliers visually obvious, at the cost of being somewhat subjective — two people can draw two different lines through the same cloud.
Least-squares regression. This is the statistical version of drawing the best-fit line: it finds the one line that minimizes the squared distance to every data point. It uses all your data, is fully objective, and reports an R-squared value telling you how well activity actually explains the cost. In a spreadsheet it is two functions — SLOPE for the variable rate and INTERCEPT for the fixed cost — or one LINEST call for the full statistics. If you have a year or more of monthly data, regression almost always beats high-low, and it costs you about sixty extra seconds.
A practical rule of thumb: use high-low for a first-pass estimate or when you only have a few periods of data, and switch to regression once you have twelve or more clean observations and the decision actually matters — annual budgets, pricing overhauls, and make-vs-buy calls all qualify.
Putting Your Cost Equation to Work
Splitting mixed costs is not an academic exercise. Once fixed and variable pieces are separated, several everyday decisions get dramatically easier:
- Budgeting. Plug planned activity into each cost equation instead of guessing next year's bills from last year's average. A budget built on cost behavior flexes correctly when volume changes.
- Pricing. The variable rate tells you the true incremental cost of one more unit — the floor below which a special order loses money. Fixed costs, already covered by existing volume, should not inflate that floor.
- Break-even analysis. Contribution margin (price minus variable cost per unit) is the engine of break-even math, and you cannot compute it until every mixed cost is split.
- Outsourcing and make-vs-buy. Comparing an in-house cost against a vendor quote is only fair when you isolate the costs that would actually disappear — the variable piece plus any avoidable fixed cost.
- Variance investigation. When a bill comes in over budget, the equation tells you whether activity was simply higher than planned (expected) or the rate itself moved (worth investigating).
None of this works without clean inputs, and that is where your bookkeeping earns its keep. The method needs matched pairs — activity and cost for the same period — recorded consistently month after month. If miles driven live in a glovebox notebook, utility bills pile up unopened, and machine hours exist only in someone's head, no formula can save you. Tracking both sides in one version-controlled ledger turns cost analysis from a quarterly archaeology dig into a query you can run any afternoon. If you want your numbers in plain text you fully own, the Beancount documentation shows how to structure accounts so activity-linked costs stay separable from day one.
Track the Pairs, and the Math Takes Care of Itself
The high-low method asks very little of you: find your busiest and quietest periods, divide one difference by the other, and subtract once. In return you get a cost equation that budgets, prices, and forecasts better than gut feel ever will. Just remember its three conditions — anchor on activity, watch for outliers, and stay inside your relevant range — and upgrade to regression when the stakes justify sixty extra seconds in a spreadsheet.
Simplify Your Financial Management
As you sharpen your cost analysis and budgeting, maintaining clear financial records is what makes techniques like the high-low method possible — matched activity and cost data, month after month, with nowhere for an outlier to hide. Beancount.io provides plain-text accounting that gives you complete transparency and control over your financial data — no black boxes, no vendor lock-in. Get started for free and see why developers and finance professionals are switching to plain-text accounting.
