Cumulative Variance Calculator
One bad month rarely sinks a budget. What sinks it is a small miss, repeated, that nobody totals up until the year is half gone. Cumulative variance is that total: every period’s gap between what happened and what was planned, added together, so the drift is visible while it is still cheap to fix.
This calculator works in two modes. Budget variance mode pairs your actual and planned figures period by period and shows each gap plus the running total. Statistical variance mode treats your entries as a data series and shows how its variance builds as each new value arrives. Both modes show their full working in a table, not just a headline number.
Positive variance means over plan; negative means under. Whether over plan is good or bad depends on whether you are tracking costs or revenue, so check your own convention before reporting the number.
Quick Answer
How to use the two modes
Pick the mode that matches your question. Budget variance answers “how far off plan are we in total?” Statistical variance answers “how has the spread of this series grown as it arrived?” They share a name and nothing else, so the tab matters.
Enter values in period order, comma-separated. In budget mode the two lists must cover the same periods in the same order: the third actual is compared with the third planned value, so keep them paired.
Read the running total, not just the headline. The table shows the cumulative figure after every period. A total that worsens steadily tells a different story from the same total caused by one bad period, and the response should differ too.
Entries that are not numbers are ignored rather than guessed at, so if the tool reports fewer periods than you expected, look for a typo or a missing comma in your list. Budget mode refuses mismatched lists outright instead of pairing the wrong periods, which is the single most common way cumulative variance figures go wrong in spreadsheets.
Budget mode: the formula and a worked quarter
The budget formula is deliberately plain. Each period contributes one subtraction, and the cumulative variance is the sum of those subtractions, which is identical to cumulative actual minus cumulative planned.
| Period | Actual | Planned | Variance | Cumulative |
|---|---|---|---|---|
| 1 | 90.00 | 100.00 | -10.00 | -10.00 |
| 2 | 105.00 | 100.00 | +5.00 | -5.00 |
| 3 | 98.00 | 100.00 | -2.00 | -7.00 Under plan |
Read the running column as the story of the quarter. Period one missed by 10. Period two clawed half of it back. Period three slipped a little again, leaving the quarter 7 under plan. A manager who only saw the -7 might hunt for a problem spread evenly across the quarter; the table says the problem was period one, and it was already half fixed. The percentage puts the size in context: -7 against a plan of 300 is a 2.33% underrun, while the same -7 against a plan of 20 would be a 35% miss. Same variance, completely different urgency.
A second case, in the form the tool is most often used: monthly sales against forecast. Enter actuals of 4200, 3900, 5100, 4800, 4600, 5300 against forecasts of 4000, 4000, 4500, 5000, 5000, 5000 and the cumulative variance finishes at +400, or +1.45% against a cumulative forecast of 27,500. The running total peaks at +700 after month three, dips to +100 after month five, and recovers in month six: a half-year that looks calm in total was anything but calm in the middle, which is exactly what the cumulative view is for.
Reading the sign: over plan is not automatically bad
This tool reports actual minus planned, in that order, always. What the sign means then depends on what the numbers are. For revenue or output, positive is usually good news and negative is a shortfall. For costs, the reading flips: positive means you spent more than planned. The calculator’s verdict sticks to the neutral words “over plan” and “under plan” for this reason, and you should too when you pass the figure on.
Teams that need the component view of a cost overrun, split into price and quantity effects, are asking an absorption or overhead question rather than a cumulative one. The absorption variance calculator and the controllable overhead variance calculator break a single period’s overhead variance into its drivers, while this page totals the periods. Likewise, a variance measured on cash movements rather than budget lines belongs in the cash flow variance calculator, where timing differences, not plan misses, explain most of the gap.
Statistical mode: variance of each prefix, summed
Switch tabs and the same words mean something else. For a data series, the tool takes the first value on its own, then the first two, then the first three, and so on: each of these expanding groups is called a prefix. It computes the population variance of every prefix and adds those variances together. Population variance means the squared deviations from the prefix mean are averaged by dividing by the count, n, because each prefix is treated as the complete data available at that moment, not as a sample from something larger.
| Prefix | Mean | Variance | Running total |
|---|---|---|---|
| first 1 | 5.0000 | 0.0000 | 0.0000 |
| first 2 | 7.5000 | 6.2500 | 6.2500 |
| first 3 | 10.0000 | 16.6667 | 22.9167 |
| first 4 | 12.5000 | 31.2500 | 54.1667 |
For 5, 10, 15, 20 the cumulative variance is 54.1667. Two properties are worth knowing before you use the figure. First, it can never decrease: variances are never negative, so every new point adds zero or more, and a sudden jump in the running total marks the arrival of a point far from the earlier values. Second, because it is a sum across prefixes, longer series produce larger totals even at the same spread, so compare cumulative variance only between series of the same length, or compare the prefix table row by row instead. Used that way it is a process-monitoring view: it shows when a series started behaving differently, not merely how spread out it ended up. Readers who need the variance of a conditional or updated distribution rather than a running sum should see the conditional variance calculator, which answers that different question directly.
What competing calculators miss
The competing cumulative variance tool reviewed for this page takes the same two lists and returns a single total. That is where it stops, and where its problems start.
Miss 1: no per-period table. A lone total cannot show which period caused the miss or whether you are recovering. This calculator’s running column showed the worked quarter recovering in period two and slipping in period three, information the total of -7 hides completely.
Miss 2: no percentage context. The competitor never divides by the plan, so a variance of 7 reads the same against a plan of 300 and a plan of 20. Here the variance versus plan percentage is part of the result, because size without scale is how small misses get escalated and large ones get shrugged off.
Miss 3: a confused sign story. The competitor’s own explanation calls a positive result favorable and under budget in one paragraph, which cannot both follow from actual minus planned for a cost budget. This page fixes the order (actual minus planned), uses the neutral over/under plan wording, and warns where the earned value convention differs, so the number survives being quoted in a report.
Miss 4: the statistical meaning is absent or faked. Statistics tools that use the word cumulative often return the plain variance of the whole series. This one shows its definition openly, with every prefix’s mean, variance, and running total in the open, so the specialized meaning is auditable rather than hidden behind a label.
Frequently Asked Questions
The total deviation between actual and planned results across periods: the sum of (actual minus planned) for each period. It is the headline number in budget reviews and earned value analysis.
It depends on what you measure. For costs, positive variance (actual above plan) is usually bad; for revenue, it is usually good. Always label which one you mean.
How the spread of your data evolves as observations arrive: the sum of the variances of the expanding prefixes. Use it for process monitoring, not for budgets.
Each variance pairs one actual period with its planned counterpart. Mismatched lists cannot be paired, so the tool asks you to fix the counts instead of guessing.
Yes. Enter actual sales and forecasted sales per month; the cumulative variance is your total forecast miss, and the table shows which months drove it.
Variance is the average squared deviation; standard deviation is its square root, in the original units. This tool works in variance units throughout.
Because each prefix is treated as the complete set of data available at that point, which is the population variance definition: squared deviations averaged over the count itself. Dividing by n minus 1 is the sample variance correction, used when your data is a sample meant to estimate a larger population. The two give different numbers, so do not compare this mode’s output with a sample variance from a spreadsheet’s VAR.S function.
Decimals and negative values are fine in both modes: enter them as plain numbers separated by commas. Text entries are ignored rather than treated as zero, so a label typed into the list silently reduces your period count. If the result covers fewer periods than you entered, check the list for a stray word or a missing comma.
Add This Calculator to Your Website
References
- Project Management Institute, “Earned Value Management Systems” (PMI library): defines cost variance as EV – AC in the earned value system and explains how its sign convention differs from a plain actual-versus-plan total. PMI: Earned Value Management Systems. Accessed October 2026.
- Project Management Institute, “Managing Firm Fixed Price Projects Using Earned Value Management” (PMI library): the basic EVM data elements (planned value, earned value, actual cost) and the variance measures derived from them. PMI: Managing Fixed Price Projects Using EVM. Accessed October 2026.
- Project Management Institute, “Earned Value Analysis in the Cost Management of Construction” (PMI library): applied variance analysis against the ANSI/EIA-748 earned value standard. PMI: Earned Value Analysis in Cost Management. Accessed October 2026.
- OpenStax, “Business Statistics,” section on measures of the spread of the data: variance as the average of squared deviations, and the population-versus-sample distinction behind dividing by n or n – 1. OpenStax Business Statistics: Measures of the Spread of the Data. Accessed October 2026.
- Penn State, STAT 200 elementary statistics course, “Measures of Spread”: population variance and standard deviation formulas used in the statistical mode’s prefix calculations. Penn State STAT 200: Measures of Spread. Accessed October 2026.
- Reliasoft, “Statistical Background on DOE”: population mean and variance definitions and the sample estimator distinction, the reference for why complete prefixes divide by their own count. Reliasoft: Statistical Background on DOE. Accessed October 2026.