Why an Unlisted Product Is Approved One Project at a Time

By Michael Stern, JDS Development Group

On every building I have worked on that had few precedents, there has been at least one assembly sitting in a laboratory with nothing behind it in any code table. The thing is real. It is fabricated, instrumented, waiting for a load, and it has no permission to be installed in the building it was made for.

That gap, between a product that exists and a product a city will accept, is the subject here. I am writing it for developers about to commit to something unusual, for architects of record who will carry the filing, and for construction managers holding a schedule around it. The route through the gap is well worn, and almost nobody plans for it the first time.

A code carries what has already been proven, written down as classifications, standards and acceptance criteria. A product outside that record is not forbidden, only unrecognized. Unrecognized is a condition with a procedure attached to it.

What happens when the building code has no listing for a product?

It goes down an alternative route and is approved for that one project. The reviewing agency decides what evidence it will accept, measured against criteria that either already exist or have to be established before the laboratory has anything to test to.

In New York the department separates materials the construction codes prescribe from materials the codes do not prescribe even where nationally published standards for them exist. The second group carries an acceptance step of its own, and the criteria have to be in place before anything can be approved against them. That ordering is the part teams get wrong.

Establishing criteria is work in its own right. Somebody has to propose what the product must demonstrate, and the agency has to agree that demonstrating it proves the product fit for the use proposed. Where an accredited evaluation service already publishes criteria for that class of product, a great deal of that argument has been had in public and in advance.

What the agency reads is a package rather than a letter. It holds the criteria, the test reports, the calculations, and a description of the product precise enough that the thing approved and the thing delivered are the same object.

The testing itself is done by an accredited independent laboratory that either performs the tests or witnesses them, never by the people who want the answer. Engineers with the manufacturer seal the calculations, engineers with the evaluation service read them, and the quality system that makes the product is inspected so that what gets shipped is what was tested. That chain is the whole value of the exercise.

A product you can buy is not yet a product you can build with.

The people who run this are the code consultant and the architect of record, working with the manufacturer. A developer who tries to drive it personally slows it down. What a developer can usefully do is ask, before the design is fixed, which elements of this building have no listing behind them at all.

Approval work belongs on the schedule beside the structure

I schedule an approval the way I schedule a structural package. It gets a start date, an owner, a date by which the answer is needed, and a drawn alternative behind it. On a building with few precedents there are usually several of them running at once.

They cluster where the building does something nothing nearby has done. The American Copper Buildings are twin towers joined 300 feet up by a three story skybridge, the first skybridge in New York City in 80 years. When a building carries an element the city has not seen in eighty years, nobody at the counter can look it up, and the team has to arrive with the evidence already assembled.

The same is true of a tower whose proportion has no local precedent. 111 West 57th Street is the tallest and most slender residential building in the Western Hemisphere, built in the courtyard of the landmarked Steinway Hall and designed by SHoP Architects. A building like that asks ordinary questions in a place where the ordinary answer does not reach.

The alternative drawing matters more than the approval does. Every unlisted element on a building of mine has a second version behind it built only from recognized assemblies, because a design with no second version is a design betting its schedule on a review it does not control.

Fabrication makes the timing harder. The elements with no listing behind them are often the ones with the longest lead times, so the approval and the order compete for the same weeks, and releasing the order early means paying for whatever the review changes.

This is the same discipline as independent peer review on a structure with no precedent. Buy the scrutiny early, on purpose, and let it shape the design while the design can still move.

Does an approval on one building carry to the next?

Generally not. The approval attaches to the project it was granted for, so the same product on the next building goes through the same testing and the same review again.

That surprises people and it should not. An approval is a judgment about a product in a context, meaning a use, a position in the building, and a set of conditions the data actually covered. Change the building and the context changes with it.

There are moving parts underneath as well. Code editions change, criteria are revised, and the agencies that do the testing are themselves certified and recertified. The same office that evaluates new materials in New York also oversees the technical certification of the approved agencies whose inspections, tests and material approvals the file depends on.

What does carry forward is the file. The data, the criteria and the reports exist after the first project, so the second approval is faster and cheaper than the first without being free. On a detail a developer expects to use again, that is usually the argument for staying with the same product.

The approval also does not end at the permit. Inspection during construction is what keeps the installed product tied to the tested one, and that regime arrives with the approval rather than after it. A product approved on paper and installed without that check is an open question inside a finished building.

It is also the argument for finding out, early, whether a product already holds a published evaluation report. A report that a reviewing official can reference is worth more on a schedule than a better product nobody has assessed yet.

Criteria are written for a class of product and not for one building

Given a choice between a better product with no criteria behind it and an adequate product with published criteria, on a tight schedule I take the adequate one. I have taken it more than once, and a structural engineer is entitled to tell me that is how a city stops getting better buildings.

The objection is a real one. Criteria get written because somebody pushed a product through the first time and paid for the work, and a market where every developer defers to what is already listed is a market where the list stops growing. I take the point and I still schedule around it, which is a position rather than a principle.

The compromise I accept is a duller detail in places where a better one was available. More than once the version that shipped was the recognized assembly and the drawing of the better one stayed in a file. Nobody who lives in the building will ever know which one they got.

The second compromise is time. An approval that depends on fabricating a specimen, testing it, reviewing the report and publishing criteria runs in months, and those months sit in the middle of the design rather than at the end of it. A schedule that assumes otherwise is a schedule nobody can carry.

There is a limit to what I can tell you about any particular product, because the route a given one takes depends on what it is, where it goes in the building and which edition of the code the project is filed under. Read the rule rather than the summary. The building code and the rules a department writes under it are where the route for a specific product is actually set.

Retain a code consultant before the design is fixed rather than after a filing comes back, and let the architect of record own the submission. Ask the manufacturer in the first meeting what has already been tested, by whom, and against which criteria. Those three answers tell you most of what the schedule is going to look like.

A building code is a record of what has already been proven. Everything a difficult building does for the first time has to be proven again, in a laboratory, for that building alone.

Michael Stern is the founder and chief executive officer of JDS Development Group.

Photos by Renan Guedes on Unsplash

Sources

Material Acceptance
Technical Certification and Research
Construction referees Evaluation processes for alternative building products