Earned value management was invented for work that comes in countable units. Civil construction mostly cooperates: linear feet of pipe, each structures, tons of asphalt. Earthwork is the exception, and it is the exception on almost every job.
Why dirt is harder than pipe
A cubic yard is not a fixed thing. The same material occupies different volumes depending on where you measure it, and the difference is not small.
| State | Where it is measured | Typical relationship |
|---|---|---|
| Bank (BCY) | In place, before disturbance | The reference volume |
| Loose (LCY) | In the truck or the stockpile | Larger than bank; the swell |
| Compacted (CCY) | Placed and compacted in the fill | Smaller than bank; the shrink |
Swell and shrink factors vary by material and by how well it is compacted, and your geotech report or your own history should supply them rather than a rule of thumb from the internet. The point for EVM is not the factor itself. The point is that you must fix one measurement basis for the pay item and never change it. If the contract pays in bank yards from cross sections and your crew is counting truckloads of loose material, your earned value will be wrong by the swell factor, consistently, in the flattering direction. That is the single most common earned value error on earthwork.
Pick the basis the contract pays on, and measure earned value in that basis. Truck counts are an operational metric, not an earned value input, unless the contract pays by the load.
Do not earn twice on the same yard
On a balanced site, a single cubic yard gets excavated, hauled, placed, and compacted. If your budget has separate cost codes for excavation, haul, and embankment, you can track production on each. But if the pay item is a single line for site grading, you have one budget to earn against, and splitting a yard's value across three codes requires deciding how much of the item's value each step represents.
The clean approach is to define, up front, the completion event that earns value. Two defensible options:
- Earn at final placement. A yard earns value when it is placed and compacted in its final location. Conservative, simple, and it will understate progress in periods with heavy stockpiling.
- Earn in defined stages. Assign fixed percentages of the item's value to excavation, haul, and placement, based on how the estimate was built. More accurate through the middle of the job, and it requires the discipline to hold the percentages fixed rather than adjusting them when the numbers look bad.
Either works. What does not work is deciding case by case, which converts earned value from a measurement into an opinion.
A worked example
Pay item: 40,000 BCY of unclassified excavation at a $6.50 per BCY cost budget. BAC is $260,000, planned at 4,000 BCY per week over 10 weeks. Value is earned at final placement.
End of week 4. Surveyed and cross-sectioned quantity in place: 13,000 BCY. Actual cost charged to the code: $97,000.
EV = 13,000 BCY × $6.50 = $84,500
AC = $97,000
CPI = 84,500 ÷ 97,000 = 0.87 | CV = −$12,500
SPI = 84,500 ÷ 104,000 = 0.81 | SV = −$19,500
EAC = 260,000 ÷ 0.8711 = $298,500 | VAC = −$38,500
Now the part that only matters on dirt. Suppose the crew also has 2,800 BCY excavated and sitting in a stockpile, not yet placed. Under the rule you chose, that material has earned nothing. It cost money to move, so it is in AC, and it is dragging CPI down.
Is that a problem with the method? No, and this is worth sitting with. The stockpile is real work with real value, and your rule says it does not earn until it is placed. The consequences are:
- CPI will look worse than reality during periods of heavy stockpiling, then improve when the stockpile is placed.
- That swing is predictable, so it is manageable: note it, expect the correction, and do not panic-react to a dip you created with your own accounting rule.
- If stockpiling is a large and persistent part of your operation, the staged-earning approach is the better choice and you should have picked it at the start.
Where the quantity number comes from
On earthwork you have several possible sources of installed quantity, and they do not agree with each other:
| Source | Strength | Weakness |
|---|---|---|
| Survey cross sections / drone or GPS surface comparison | Usually the contract authority for payment | Periodic, so it is not available daily |
| Truck counts | Available in real time, easy to capture | Loose measure; needs a conversion factor and is easy to over-count |
| Machine control production data | Continuous and detailed | Depends on model accuracy and does not exist on every machine |
| Foreman estimate by station | Always available | Least precise; fine for daily management, weak for billing |
The practical answer for most contractors is to use a fast daily source to manage and the survey to settle. Capture truck counts or station progress daily so CPI and SPI are current enough to act on, then reconcile to survey at the pay period. Having the daily record turns that reconciliation into a comparison of two numbers rather than a reconstruction of a month.
Other earthwork-specific traps
- Over-excavation and unsuitable material. Extra yardage removed at the engineer's direction is usually a separate item or extra work. Letting it flow into the base excavation code destroys the CPI on the base item and buries the entitlement.
- Weather and standby. Rain days generate cost with no earned value. They belong in the record as their own event, not smeared across production codes, or you will diagnose a productivity problem that is really a weather problem.
- Haul distance changes. A budget built on a 900-foot haul does not survive a 2,400-foot haul. This shows up as a CPI drop that no amount of crew pressure will fix. It is an estimate or a claim issue, and the field record of haul routes is what supports it.
- Topsoil and strip. Frequently a separate item, frequently coded to general excavation by a crew that does not think about pay items. This one is nearly universal.
- Item completion is not surface acceptance. Grade tolerance, proof rolls, and density all have to pass. Earning 100 percent at rough placement overstates progress.
Where to go next
For the underlying formulas, see what is earned value in construction and CPI and SPI explained. For turning the same data into crew-level production rates, see production rates and how to track them.
The hard requirement in everything above is a daily, dated, attributed quantity record. PM Axsus captures quantities against pay items at the point of installation, alongside the hours that produced them. See quantity tracking and PM Axsus for excavation contractors.