Why a $3,994,000 Envelope Cut Left 530 Units at HERS 44.59

A $3,994,000 cut put 530 units on the edge of failing energy code. The project is a 530-unit luxury multifamily building in Westford, Massachusetts, all electric. The team looked at dropping the continuous insulation and swapping triple pane windows for double pane, and the savings were real. What the budget sheet did not show is that the resulting design landed at a HERS score of 44.59 against a code limit of 45. That is a gap of 0.4 points on a building with 530 units in it.

The Real Numbers

The building is 530 units, luxury multifamily, all electric, in Westford, Massachusetts. The value engineering exercise was straightforward and the kind that happens on every large project. Drop the continuous insulation. Swap the triple pane windows for double pane.

The savings came to $3,994,000. On a project of this size that is not a line item anyone ignores. It is the difference between a deal that works and a deal that gets restructured, and the analysis showed the reduced envelope still cleared the code target.

That is where most reviews stop. The model says the design passes, the savings are documented, and the decision looks made. Nobody asks a second question of a design that already cleared the target, because the target is binary in the way it gets reported. Pass or fail, and this one passed.

The number underneath the pass is what changes the picture. The double pane design put the building at a HERS score of 44.59. The code limit is 45. The design cleared the target by 0.4 points.

The Margin Problem

A pass is a pass right up until the score moves, and scores move. Energy modeling software has a built-in swing of plus or minus 2 points depending on which program you use. That is not a modeling error. It is the spread between accredited tools running the same building with different engines, different default assumptions, and different rounding behavior on the way to a single number.

Run the worst-case unit through that range and the score lands anywhere from 42.6 to 46.6. The low end passes comfortably. The high end fails code.

So the $3,994,000 stays in the budget on paper. But one unit that tests on the wrong end of that swing and the whole building is out of compliance. On a 530-unit building the exposure is not theoretical, because the rating does not get carried by an average. It gets carried by the units that get rated, and a large all-electric building has units with more exposed wall, more glazing, and worse orientation than the building as a whole.

The envelope was carrying real risk that the cost savings did not reflect. That is the gap between the two numbers. The budget sheet showed $3,994,000 in savings and no line for what a 0.4 point margin costs to defend if the score moves the wrong way.

What a Failed Score Actually Costs

The reason a thin margin matters is that the fix arrives late and the leverage is gone by then.

A compliance shortfall does not surface during design when the envelope is still a drawing. It surfaces after construction, when the rating is being finalized and the building is standing. At that point the continuous insulation that was cut is behind finished cladding and the double pane windows are installed in 530 units.

Buying back 2 points at that stage does not happen through the envelope. It happens through whatever is still accessible, which is usually mechanical equipment, water heating, lighting, or added renewables. Those are the levers that stay open late, and they are priced as a change order rather than a bid.

There is a schedule cost on top of the hard cost. Certificates of occupancy, lender requirements, and incentive program milestones tend to sit downstream of the rating. A building that cannot close out its rating on schedule is a building carrying construction financing longer than the model said it would.

We looked at the cost levers on this same 530-unit project from the other direction, ranking which envelope move returned the most dollars. Every path in that exercise cleared the model. This is the part that analysis does not capture. Clearing the model and clearing the model with room to absorb tool variance are two different results, and only one of them is a plan.

When a Thin Margin Is Worth Taking

None of this means a 0.4 point margin is automatically the wrong call. There are projects where a thin pass is a reasonable position, and the difference is whether the thinness was chosen or discovered.

When the rating tool is locked and known. The plus or minus 2 point range is a spread across programs. If the project is committed to one accredited tool and the rater has a track record with it, the practical range narrows considerably from the theoretical one.

When the worst-case unit has already been modeled. A margin measured against the building average is fragile. A margin measured against the worst orientation, the top floor corner, and the highest glazing ratio is a much more defensible 0.4 points.

When there is a documented buy-back plan. A thin margin with a costed fallback is a managed risk. Knowing in advance which mechanical or water heating change closes a 2 point gap, and what it costs, turns a surprise into a decision.

When the savings genuinely carry the project. Almost $4,000,000 is not a rounding error. On a deal where that money determines whether the project gets built, accepting a thin margin with eyes open is a legitimate business call, provided somebody actually made it.

When field performance is expected to beat the model. Some teams consistently test tighter than their modeled air leakage assumption. Where that history exists and the air sealing scope is real, the as-built score tends to land below the modeled one rather than above it.

The Decision Framework

The useful question on a project like this is not whether the design passes. It is how much room the design has when the score moves. A HERS target should be treated as a number you clear by a margin you chose, not a number you clear by whatever is left over after value engineering.

Before signing off on a $3,994,000 envelope cut, model the worst-case unit rather than the typical one, ask what the score does under the tool spread, and price the buy-back for a 2 point miss. If that buy-back is affordable, the thin margin is a risk you can carry. If it is not, the savings are smaller than they appear, because part of that money is holding open an exposure nobody has funded.

The Westford numbers are specific to Westford. A 530-unit all-electric luxury building with that much glazing behaves differently from a 40-unit building with a modest window-to-wall ratio, and a project with a lower HERS target has room this one does not. What carries across is the discipline of asking where the score sits relative to the limit, not just which side of it. The only way to know how much margin your building is buying with its envelope dollars is to model your building, including the unit that tests worst.