The Tile Is Not the Floor: What ANSI A326.3 Can and Cannot Tell You About a Floor in Service
By Walkway Management South Florida
Grant Davidson, a standards development engineer at IPA Laboratories, has published a two-part series on slip resistance. Part 1 introduces the dynamic coefficient of friction and the ANSI A326.3 test method. Part 2 covers the product use classification system built on top of it.
Both parts are correct. I want to say that plainly before I add anything, because what follows is not a correction of either one.
What interests me is what the two parts add up to when you read them together. Each article states something true and unremarkable on its own. Put side by side, they define the outer boundary of what this system can tell you, and that boundary sits a good deal closer than most people using these numbers assume. Davidson does not draw the line himself, probably because he is writing for people upstream of where it matters. I get called downstream, after somebody has already fallen, which is where the line shows up.
What Part 1 establishes
Davidson defines DCOF using the standard’s own language: the ratio of the force necessary to keep a surface already in motion sliding over another surface, divided by the weight of the sliding object. He notes A326.3 was developed through the ANSI consensus process with input from a committee of more than sixty professionals, that it requires a tribometer and evaluates the surface wet, and that while it was originally developed for ceramic tile it is now applied broadly across hard surface flooring.
Two things in that article deserve more attention than they usually get.
The first is the word dynamic, and Davidson is right to lead with it. A dynamic measurement describes a surface with the foot already moving. A static measurement describes the force needed to start a slide from a dead stop. Nobody slips from a dead stop. People slip mid-stride, and a static number on a wet floor can read comfortably high for a surface that is genuinely dangerous, because a stationary rubber foot lets the liquid squeeze out of the contact area while it sits there. That is not a small technical distinction. It is the reason ASTM C1028 was withdrawn in 2014, and we have written about the mechanism in why C1028 was withdrawn. Part 1 gets this right and it is worth agreeing with loudly, because plenty of reports still in circulation do not.
The second is a single sentence that Davidson states almost in passing:
Different contaminants such as dirt, water, soap, oil, or grease can change this value.
Hold onto that one. It becomes the whole argument further down.
What Part 2 establishes
Part 2 explains that A326.3 is more than a test method: it carries a five-category product use classification, where manufacturers declare their products into one or more categories using their own internal selection criteria. Davidson describes this manufacturer-declared framework as unique in the North American marketplace, and that is accurate.
He then makes the argument that most needs making in this field. Slip resistance is influenced by a long list of factors and cannot be reduced to a single number. He separates the available methods by what each is actually good at: wet DCOF under A326.3 for nominally flat surfaces, in lab or field; shod ramp testing under EN 16165 Annex B for three-dimensional or heavily textured surfaces a dragged sensor cannot read properly; pendulum testing under EN 16165 Annex C, and ASTM E303 on this side of the Atlantic, for flat surfaces with small-scale texture, in lab or field.
That is the correct way to think about methods. A method is a tool with a working range, not a competing brand. And it follows from this, as EN 16165 states directly, that results from one annex are comparable only to results from the same annex. A pendulum value is not a DCOF in different units. Anyone offering a conversion chart between them is selling something, and we have written about why in SCOF vs DCOF vs PTV.
I would add one working preference to his framing rather than a correction. We lead with the Pendulum Slip Tester as our reference method and bring the BOT-3000E alongside it, because the pendulum measures the floor in the building in the condition we found it and has a deep international research base behind it. The DCOF measurement earns its place next to it for comparison and monitoring, which is what it is genuinely good at.
What the two parts add up to
Here is the synthesis, and it is built entirely out of Davidson’s own material plus the standard’s own method.
Part 1 tells you the number moves with the contaminant. Dirt, water, soap, oil, grease. Change the contaminant and you change the value.
The method holds the contaminant constant on purpose. A326.3 specifies an SBR sensor and a 0.05% sodium lauryl sulfate solution. That is a deliberate and defensible engineering choice. SLS is a surfactant, and it breaks the surface tension of the water so that even a hydrophobic surface takes a thin, uniform film. Without it you would be measuring how well your water happened to spread that morning. With it, two labs on two continents can test the same tile and agree.
Part 2 tells you the category is declared in advance, by the manufacturer, using internal criteria, before anyone has seen the building.
Now put those three together.
The classification describes how a product performed against one standardized surfactant solution, under a specified rubber, in a specified geometry, and was then assigned to a use category by forecast. Every part of that is exactly right for the job it was built to do, which is letting a specifier compare candidate products on a level field.
None of it describes the floor in a commercial kitchen carrying a film of fryer grease. Or a pool deck carrying sunscreen and body oils. Or a grocery entrance carrying rainwater and tracked-in soil in February. Those are the contaminants Davidson’s own Part 1 named as things that change the value, and the classification was never measured against a single one of them.
That is not a flaw in the standard. Holding the contaminant constant is the only way to make products comparable at all. It does mean the resulting number answers a comparison question, not a condition question, and those get confused constantly.
The sentence in the standard that almost nobody quotes
A326.3 says this about itself:
The measurement can provide a useful comparison of surfaces, but does not predict the likelihood a person will or will not slip on a hard surface flooring material.
The standard is not claiming to tell you whether someone will fall. It is claiming to let you compare one surface against another under controlled conditions. That is the same distinction the contaminant point produces, stated by the document itself, and it is the least-cited line in it.
The standard also tells the specifier what to weigh: type of use, traffic, expected contaminants, expected maintenance, expected wear, and the manufacturer’s guidelines. Read that list slowly. Three of them carry the word “expected” in the standard’s own phrasing. Two more describe how a building that is not yet occupied will be used. Exactly one, the manufacturer’s guidelines, is a document that exists on the day the tile is specified. Not one of the six is a property of the tile.
The declaration is a forecast about a building that does not exist yet
Section 3.4 asks manufacturers to declare a product use classification based on manufacturing parameters, internal quality control criteria, their experience with similar surfaces, and the criteria in the standard.
Now look at the factors the standard offers as examples of what a manufacturer might weigh. Expected footwear. Expected contaminants. The flatness or incline of the flooring after installation. How the flooring is used and maintained. Flooring structure. How drainage occurs where liquids are involved.
A manufacturer cannot know any of those things.
A plant producing porcelain does not know the slope of the shower pan it will end up in, whether the installer will hold the specified grout joint, what the housekeeping contractor will switch to when the budget tightens in year two, or whether the drain will be set correctly. It is being asked to declare a category against an assumed installation and an assumed maintenance program, using experience with similar surfaces as its best proxy.
That is not a criticism of manufacturers or of the standard. It is the only thing a manufacturer is in a position to do. Someone has to make an informed forecast at the front of the process, and the party that knows the material best is the sensible party to make it.
But it is a forecast, and a forecast has a property a measurement does not. It can turn out wrong later, quietly, with nobody at fault. The tile can be exactly what the manufacturer declared and the floor can still be a problem, because the floor is the tile plus everything that happened to it.
Nothing downstream checks whether the forecast held.
Interior Wet Plus is the honest illustration
Four of the five categories carry a wet DCOF figure a product can be measured against. Interior Wet is 0.42. Interior Wet Plus is 0.50. Exterior Wet is 0.55, and so is Oils and Greases. Interior Dry is assessed dry rather than wet. Our full breakdown is in the A326.3 minimums.
Interior Wet Plus is the interesting one. It covers barefoot and standing-water environments, pool surrounds, locker rooms, shower areas. It carries the 0.50 figure and it is also a manufacturer-declared class, resting on manufacturing parameters, internal quality control criteria, and experience with similar surfaces.
Two things are true about that at once, and the industry tends to say only one of them.
The first is Davidson’s point, and it is fair. A single dragged measurement on a flat coupon is a poor description of how a heavily textured barefoot surface behaves under standing water. Letting the manufacturer bring knowledge a coupon test cannot capture produces a better classification than a bare number would.
The second is that a classification built on internal criteria is not independently verifiable from the outside, and is not necessarily comparable between two manufacturers who weigh their criteria differently. That is fine for its purpose. It is a product selection aid. It is a materially different kind of evidence from a measurement taken on a specific floor on a specific day, and it should not be asked to do that second job.
Wet Plus covers the exact environments where people are barefoot, the surface is wet by design, and a fall lands on hard material. It is worth knowing that this is the category where the standard leans hardest on the manufacturer’s judgment, and where the real contaminant, sunscreen and body oil and soap, is furthest from the SLS solution the number came from.
The method’s reach has grown, and verification should grow with it
Part 1 notes that A326.3 was developed for ceramic tile and is now applied broadly across hard surface flooring. That is true and it is stated neutrally, which is right.
It is also worth sitting with. A resinous coating, a polished concrete slab, a sealed natural stone and a luxury vinyl plank do not present the same thing to a dragged SBR sensor that a fired ceramic body does. They differ in hardness, in porosity, in how they hold a surface treatment, and in how they wear. The method travelling to those materials is a reasonable development, and the tile industry deserves credit for producing the one North American method everyone else could borrow.
But as a method moves further from the material class it was developed around, the case for checking the result on the actual installed surface gets stronger rather than weaker. That is an argument for more field verification, not for less confidence in the standard.
What happens between the pallet and the fall
The floor someone walks on is not the product that was classified. Between the two sit steps that each move the friction, and none of them are visible afterward.
Installation sets the slope, the drainage, the lippage between units, and the grout ratio, all of which change how water behaves on the surface. A surface treatment can arrive after the fact, a sealer or a coating or a polish, and a treatment that was never part of the manufacturer’s declaration is now the actual walking surface. The cleaning program comes next, and in my experience it changes more floors than traffic does: the wrong product, the wrong dilution, or a residue that never gets rinsed will move a floor’s wet friction measurably. Then wear, concentrated in traffic lanes rather than distributed evenly. Then contamination that works down into the surface texture and fills the micro-profile that was providing the grip.
I have measured what that looks like. On a restaurant floor we tested, the dining walkway running past the kitchen door read 0.33 DCOF dry, while the entry foyer, same flooring, same building, same morning, read 0.71 dry. More than double the friction from the same product, with no water involved at all, because grease had worked down into the texture in one area and not the other. The full write-up, including the pendulum swing data that identified the mechanism, is in why a greasy floor tests better on the seventh swing.
Whatever category that product carried in the catalog was accurate. It simply had nothing to say about the walkway by the kitchen door.
Match the method to the question, not only to the surface
Davidson’s advice to match the method to the surface is sound. I would put a second filter beside it: match the method to the question you are being asked.
If the question is which of these three products to specify for a hotel pool deck, then product-level testing under whatever method suits the surface is the right tool, and the classification is the right shorthand.
If the question is what the friction of this floor was, in this condition, on the day someone fell, most of the toolkit is structurally unable to answer. Both ramp methods in EN 16165, barefoot in Annex A and shod in Annex B, are laboratory procedures and are not in-situ tests. They cannot be taken to the building. That is not a shortcoming, it is what they were built for. It does mean that if the floor is already installed and someone is already hurt, the ramp result on file describes a sample that is no longer the thing in dispute.
The methods that can answer a question about an installed floor are the ones that travel: the pendulum, and field DCOF. That is the practical reason we run both on a job rather than picking one. Where they agree, the convergence is worth something. Where they disagree, the disagreement usually explains itself, and the explanation is generally the interesting part of the report. We have written about where field DCOF helps and where it stops in the limitations of BOT-3000E DCOF testing, and about how we run the pendulum on our pendulum testing page.
The handoff nobody is standing at the end of
Davidson closes Part 2 by saying the framework should support better decision making and clearer communication throughout the supply chain. I agree, and I want to press on where the supply chain ends.
Manufacture, specify, sell, install. Each link hands the next one a document. The classification travels down that chain intact and does its job at every stage. Then the building opens, the chain stops, and the last document anyone holds is a statement about a product, measured against a standardized surfactant, by a party who never saw the building.
From there the floor has years of life ahead of it, and everything the standard told the specifier to think about, the traffic, the contaminants, the maintenance, the wear, starts happening for real. The forecast either held or it did not. Nobody checks.
I see the far end of that. By the time a floor reaches me there is usually an incident report, often a claim, and the classification document is in the file. It is a genuine and useful document. It is just answering a question nobody in the room is asking. The question in the room is about a floor on a date, and the only thing that answers it is a measurement of that floor.
What I would actually do with this
Use the classification. It is the right starting point, and Davidson is right to want it better understood. It is the wrong ending point.
Specify the category, then specify the verification alongside it. A line requiring the installed surface to be tested at substantial completion, under a named method and condition, turns the manufacturer’s forecast into a checked fact and costs a fraction of what an argument about it costs later.
Measure the installed floor, not the sample. What people walk on is the assembly, after the grout, the slope, the sealer, the cure, and the final clean.
Where the real contaminant is not water, say so and plan for it. A326.3 has a category for oils and greases at 0.55 for exactly this reason. A kitchen is not a showroom, and a classification obtained against SLS is the beginning of that conversation rather than the end of it.
Measure it again on a schedule. The standard already told you wear and maintenance matter. A number from handover describes a floor that no longer exists three years later. This matters most in the Wet Plus environments, where the declaration leaned hardest on judgment and the consequences of being wrong are barefoot.
Keep both documents. The classification and the field measurements are not competing claims, and neither replaces the other. One says what was specified and why. The other says what was actually there. In a dispute you want both, and you particularly want the second one to exist.
None of that argues against the classification system, and none of it argues with Davidson. It argues that a forecast made responsibly at the front of the process deserves to be checked honestly at the back of it, and that right now almost nobody is doing the checking.
Both of his articles are worth reading in full: Part 1 on DCOF testing and Part 2 on product use classification. If you specify flooring and want to know whether the assembly you are planning will actually perform the way the category suggests, or whether a floor already in service still does, that is the work we do. You can reach us through the contact page.
This article is informational and describes slip resistance measurement practice and published standards. It is not legal advice. Standards are cited by designation and edition and may be revised; verify the edition in force for your application. Every friction value given here carries the method and condition under which it was measured, and no single value from any method establishes that a floor is safe.
