Construction bid leveling an owner can check in an afternoon

Construction bid leveling is the step where three trade prices for the same package are made comparable, and it is where most scope gaps are either caught or created. An owner cannot price the work. An owner can absolutely read the leveling sheet, and six checks find most of what goes wrong.

What leveling is for

Three trade contractors bid the same package and return three numbers. The numbers are not comparable, because each bidder has read the documents differently, excluded different things and assumed different things.

Leveling is the process of adding the missing scope back to each bid so that all three cover the same work, and only then comparing them.

Done properly it is the most valuable hour on the whole buyout, because it finds the scope nobody included. Where two of three bidders exclude the same item, that item is almost certainly not in anybody price and it is not in the guaranteed maximum price either.

Done poorly it does the opposite. A leveling sheet that simply lists three numbers and recommends the lowest hands the project a package with holes in it, and the holes surface as change orders between months eight and eighteen.

The owner interest here is direct. Under a guaranteed maximum price the contractor carries the overrun, but only up to the guarantee, and only where the item genuinely was in the scope. An item nobody priced is usually an item nobody agreed was included, which makes it a change rather than an overrun.

The six checks in a construction bid leveling review

One. Is there a column for adds. A leveling sheet with three bid numbers and no adjustment rows has not leveled anything. Every bid should show the base price, the identified exclusions priced back in, and an adjusted total.

Two. Do the exclusions appear in someone else scope. Each exclusion should be marked as either carried by another package, carried by the owner, or unresolved. Unresolved is the finding.

Three. Is the low bidder low before or after leveling. It is surprisingly common for the apparent low bid to be the highest once exclusions are priced in, and equally common for the sheet to recommend it anyway.

Four. How many bidders. One bid is a negotiation. Two is a comparison. Three or more is a market. A package awarded on a single bid should be flagged whatever the number says.

Five. Spread between bidders. Three bids within four percent of each other suggest a well defined scope. A spread of thirty percent means the bidders read different scopes, and somebody is wrong about what the package includes.

Six. Alternates and unit rates. Whether they were requested, whether all bidders priced them, and whether the recommendation accounts for them.

Why the spread is the most informative number

Owners look at the lowest price. The more useful figure is the distance between the bids.

A tight spread means the documents were clear enough that three experienced estimators reached similar conclusions. That is a statement about the design as much as about the market.

A wide spread means they did not, and there are only three possible reasons. One bidder is buying the work. One bidder has misread the scope. Or the documents genuinely do not say what is included, in which case all three have guessed and two of them have guessed low.

The third reason is the common one on packages with a wide spread, and it is the one that produces change orders later.

It is also diagnostic about the design set generally. A project where several packages come back with wide spreads is telling you that the documents were not ready to be priced, which is a finding about the guaranteed maximum price rather than about any individual trade.

That is the same underlying problem described in scope gaps between trade packages, seen from the procurement side instead of the drawing side.

A worked example

A mechanical package on a 52 million dollar project. Three bids: $8.15 million, $8.44 million and $9.10 million.

The leveling sheet recommends the low bidder at $8.15 million, a saving of $290,000 against the carried figure.

Reading the exclusions rather than the totals: the low bidder excludes equipment rigging, roof curbs, seismic restraint and the temporary heating required by the schedule. The middle bidder excludes only rigging. The high bidder excludes nothing.

Priced back in, the three become $8.71 million, $8.52 million and $9.10 million. The recommended bidder is no longer lowest.

The sheet is not dishonest. The exclusions are listed on it. They are simply listed in a column nobody totaled.

Awarded as recommended, the excluded items surface over the following year as four change orders totaling $410,000 net of the apparent saving, and two of them carry schedule impact because roof curbs were discovered after the roofing package was complete.

What an owner is entitled to see

Under an open book guaranteed maximum price, the leveling sheets are part of the cost of the work record and an owner can ordinarily ask for them. Under a closed arrangement the position is weaker.

Either way, the useful request at signature is narrow: the leveling sheet for any package above a stated value, at the time of award rather than at the end of the job.

Asking at award is the point. A leveling sheet reviewed before the contract is signed allows the unresolved exclusions to be closed by the trade contractor who is still competing. The same sheet reviewed at closeout is a historical document.

Most contractors do not object to this, because a competent buyout process produces sheets they are content to show. Resistance is itself informative.

The review does not require the owner to price anything. It requires reading a column of exclusions and asking who carries each one, which is a question of scope rather than of cost.

The packages worth reading, and the ones that are not

Nobody should review every leveling sheet on a project with ninety packages. Four criteria identify the ones that matter.

Value. Anything above roughly three percent of the contract.

Interfaces. Packages that meet many others: envelope, mechanical, electrical, elevators. Scope gaps concentrate where packages meet, so leveling errors there are worth more.

Timing. Packages bought late, where the market has moved since the price was carried and where the buyout gap is most likely to be adverse.

Spread. Any package where the bids differ by more than about fifteen percent, whatever its value, because the spread says the scope was unclear.

On a typical project that is eight to twelve packages out of ninety, and reading them takes an afternoon. The rest can be taken on the buyout log alone, which is the reasoning behind ranking by exposure rather than reviewing everything, as in a budget review in an afternoon.

There is one further category worth adding: any package where the design was still moving when the tender went out. Bidders price what they were given, and where they were given a set that changed two weeks later, the leveling sheet is comparing three answers to three slightly different questions.

The tell is in the addenda. A package issued with five addenda during a three week tender period has bidders who each stopped reading at a different point, and the spread will show it.

What we do

We read the leveling sheets on the packages that carry the exposure, not all of them, and we read the exclusion columns rather than the totals. Each unresolved exclusion is written up as a finding with a value and the question to put to the contractor while the trade contractor is still competing. That work sits in the schedule and procurement risk review. Where an owner’s representative carries this reading, support for owner’s representatives does the reading while they keep the relationship.

Questions people ask