Planning variance index: how often your estimates actually hit the mark

Reviewing last month's closed work orders, a maintenance manager notices that roughly one in four came in with costs more than ten per cent away from estimate. Some ran over, some came in well under, but either way the schedule built on those numbers was shakier than anyone realised.

Where the single-job metrics tell you how far off one estimate was, the planning variance index answers a steadier question: across the whole month's work, how often did the planning function actually hit its target?

What it actually measures

The planning variance index is the proportion of closed planned work orders whose actual cost fell within a set tolerance, commonly plus or minus ten per cent, of the planned cost, expressed as a percentage of all planned work orders closed in the period. There is no single agreed definition of the window handed down from on high, so treat the tolerance you choose as a house rule and, above all, apply it consistently. A high score means the planning process is consistently accurate, not just occasionally lucky.

How to work it out

Planning variance index (%) = (Work orders within tolerance × 100) / Total planned work orders closed

In a month, 4,694 planned work orders close; 1,254 of them miss the ten-per-cent window, leaving 3,440 inside it.

Planning variance index = (3,440 × 100) / 4,694 = 73.3%

Roughly one job in four missed the target that month. The team sets an improvement goal for next quarter and starts with the largest-deviation orders, hunting for the causes they share.

Why the portfolio is the only honest place to judge planning

This index exists because of a hard truth about estimating: any single job can be off by half again or double, and that is perfectly normal, since real work throws up surprises that no estimate can foresee. Judge a planner on one job and you are reading noise. But those surprises cancel out across many jobs, so the proportion that lands on target, taken over a whole month, becomes a stable and revealing number. That is what the variance index captures: it measures the planning process as a system, not the luck of any one estimate. It is about the most direct read available on whether your planning genuinely works, and a clean way to see whether new estimating techniques or planner training are moving the needle, precisely because it rises above the job-to-job scatter that makes single estimates so misleading.

What good looks like

There is no universal benchmark for this one, so start by establishing your own baseline, then aim to improve it by roughly ten per cent each period, working steadily toward hitting the window across more and more of your work. As maturity grows, the ambition is twofold: push the hit rate higher, and then tighten the tolerance itself, so the same score comes to mean a sharper estimate.

Where it can mislead you

• Reactive work is never formally planned, so exclude it from both the top and bottom of the fraction.
• Leave out blanket and standing work orders too, even those that happen to close during the period, since they were never meant to carry a precise estimate.
• Track scope changes and use the revised planned cost as the reference, or a legitimate change of plan will look like an estimating error.
• The index reflects the planning system, not individual planners. How time is booked, how work is executed, and which priorities get pushed all shape it, so it belongs in a review of the process, never on one person's appraisal.

The planning variance index is the portfolio view of planning accuracy, and the portfolio is the only honest place to judge it. It turns a drawer full of individual hits and misses, each of them noisy on its own, into a single stable number that tells you how much faith the rest of the business can reasonably put in your plans.