ROCK COST $180,000. REPORTED PROFIT WENT UP.
Percent complete on almost every civil WIP schedule is cost to date divided by total estimated cost. Leave the bottom number alone when the dirt turns out different and the schedule reports a job getting more profitable while it loses money.
Percent complete is a cost ratio, so anything that raises cost also raises reported completion. A civil sub $790,000 into a $2M dirt job hit rock, spent another $180,000, and the WIP schedule moved from 50 percent complete to 61 percent and reported $257,800 of gross profit earned. Updating the cost-to-complete estimate for the same condition on the back half put true completion at 51.6 percent and true earned gross profit at $61,900. The schedule was overstating the job by $196,000 while the rock was still under the back half of the site.
Nothing about this is a bookkeeping error. The schedule did the arithmetic it was given correctly, using a total cost figure that stopped being true the week the hammer came out.
This post covers one line of the WIP schedule. That page covers the three things civil margin is lost to and how the whole trade is run. Read Civil Operating System for the complete treatment, worked figures included.
HOW DOES A WIP SCHEDULE DECIDE PERCENT COMPLETE?
Cost to date divided by total estimated cost. That single ratio drives everything else on the schedule. Multiply it by the contract value and you get earned revenue. Multiply it by estimated gross profit and you get earned gross profit. Compare earned revenue to what you have billed and you get overbilling or underbilling.
Look at what the ratio is made of. The top number comes out of your accounting system and it is real money already spent. The bottom number is an estimate somebody typed once, usually the estimator, usually at bid time. One of those two updates itself every week and the other one sits there until a person changes it.
That's the whole problem in one sentence. A civil job runs eighteen months, the site gets a vote, and the number that decides how complete the job is was written before anybody put a bucket in the dirt.
THE $2M DIRT JOB, AND WHAT THE ROCK DID TO IT.
Take a $2,000,000 site package. Civil subs at $1M to $5M run about 21 percent gross, so the estimate carries $1,580,000 of cost and $420,000 of gross profit. At $790,000 spent the schedule reports 50 percent complete, $1,000,000 of earned revenue and $210,000 of earned gross profit. Everything agrees and everybody is comfortable.
Then the excavation hits rock the geotech report described as weathered shale. Hammer time, longer haul cycles, a second truck on the route. Another $180,000 of cost goes into the job over six weeks. Cost to date is now $970,000.
Run the same ratio with the same bottom number. $970,000 divided by $1,580,000 is 61.4 percent complete. Earned revenue jumps to $1,227,800 and earned gross profit to $257,800. The schedule just reported that the job earned $47,800 more profit than it had six weeks earlier, during the six weeks it was bleeding.
WHAT THE HONEST NUMBER LOOKS LIKE.
Rock doesn't stop at the halfway line. Whatever sits under the back half of the site is the same material, so the honest revised total cost is $1,880,000, the original $1,580,000 plus $180,000 already spent and another $120,000 coming. Revised total gross profit on a $2,000,000 contract is $120,000.
Now the ratio: $970,000 divided by $1,880,000 is 51.6 percent complete. Earned gross profit is 51.6 percent of $120,000, or $61,900. Against the $257,800 the stale schedule reported, the job was overstated by $195,900.
Ten points of percent complete and $196,000 of gross profit, on one job, from one number nobody updated. A civil sub running four jobs at once can carry that four times over and still show a WIP schedule that foots and ties to the general ledger.
MOST SUBS MISS THIS: AN OVERRUN READS AS PROGRESS.
Every contractor knows a cost overrun is bad. Almost nobody has sat with the second-order result, which is that a cost overrun makes the WIP schedule more optimistic, immediately, in the same month the money is lost. Cost is in the numerator. Spend more and completion goes up.
That's why profit fade gets discovered at closeout and never at 60 percent. The schedule wasn't warning anybody, because it was reading the overrun as work getting done. It kept reporting the original 21 percent margin against a growing base, so the worse the job got, the more profit it claimed to have earned.
Send this to your PM and ask one question about the job he is least comfortable with: when did you last change the total estimated cost. If the answer is bid day, the percent complete on that job is a guess wearing a decimal point.
THE COST-TO-COMPLETE REVIEW THAT FIXES IT.
The fix is a monthly cost-to-complete review, done by the person who runs the job, before the WIP schedule is built. Never the estimator and never the bookkeeper. The PM is the only one who knows the haul cycles got longer.
Four questions per job, and they take about fifteen minutes each. What has this job cost so far, by cost code. What quantity is left in each code. At today's production rate and today's unit cost, what will the rest of it cost. Does the total of those two figures still equal the number in the WIP schedule.
When the answer to the fourth question is no, the schedule gets the new total before anybody reports earned revenue off it. Do that and the rock turns up as margin erosion in September, while you can still price the change order, chase the claim, and requalify the back half of the work.
WHAT TO CHANGE THIS MONTH.
Pull your current WIP schedule and add one column: date the total estimated cost was last revised. Fill it in honestly for every open job. Any job over 30 percent complete with a bid-day date in that column is a job whose percent complete you can't use.
Then look at the change orders. On the example above, $300,000 of changed conditions is a claim worth pursuing, and the WIP review is what surfaces it while the survey shots, the daily reports and the truck tickets are still easy to pull together. A stale schedule costs the profit twice: once on the job, and once on the claim nobody filed in time.
THE SHORT LIST.
FREQUENTLY ASKED.
