How this was read
On the full New Jersey home improvement contractor register was downloaded from the state's own bulk export. It held 145,261 rows covering every registration the state has issued, active and lapsed.
Three filters were applied, in this order:
- Status is Active. That removed 98,887 rows, most of them businesses that closed years ago.
- The registered address is in New Jersey.
- The trading name carries a flooring word — floor, flooring, carpet, hardwood, tile, vinyl, laminate, parquet, LVP or LVT.
That left 936 businesses. Names carrying only tile and no other flooring word were then set aside — 302 records — leaving the 634 registrations published here.
This is a floor, not a census
The register has no trade column. The classification field carries two values — Home Improvement Business Contr (145,194 rows) and Home Elevation Business Contr (67 rows) — and neither of them is a trade: one covers home improvement, the other covers raising a house. Nothing in the state's data says which registrant lays floors. The trading name is the only signal available, and it is a weak one: a flooring installer trading as John Smith Construction LLC is registered, is active, lays floors, and is not in this list. The count published here is a lower bound and should never be read as the number of flooring contractors in New Jersey.
Tile was excluded on a standards argument, not a size one. Tile installation is governed by a different body of standards — subfloor deflection, mortar beds, a separate certification route. Folding it into a schedule built on moisture ceilings and flatness spans would have put two incompatible requirement sets in one table.
Credentials are unchecked, and that is stated rather than implied. The largest flooring credential body blocks automated collection of its directory, and reading an absence as "no credentials" would be publishing a finding nobody measured. Every record here therefore shows credentials as unchecked until a person verifies one and dates it.
Where the limits came from
The figures on the schedule were read from installation documents published by the manufacturers themselves — not from retailer pages, distributor blogs or third-party guides. 8 brands are covered. Each cell carries the document it came from, and a cell with no source does not get published.
A cell on that schedule is in one of three states, and the three are printed with different words. A figure is a limit the document sets. A cell reading none required is a document that set the period at nought — 7 manufacturers do, for acclimation. A cell reading not published is the document saying nothing at all, and it adds what that leaves an installer holding. An absence is never rendered as a nought, a dash or a pass, and a nought is never rendered as an absence.
That second half is here because the page failed it. The acclimation column used to test its value for truthiness, so the 7 manufacturers whose documents set the period at nought were published under the words this table keeps for a document that says nothing — a written figure printed as silence, two sections below a heading reading missing is not zero.
One measured field is not on the schedule at all: a slab pH range, which 2 manufacturers of eight publishes. It is named on the schedule instead of shown, because a column blank for the rest tells a reader less than a sentence does — but it is in the register, and the data file this site publishes carries it.
Checked against a second copy
A bulk download can be mangled between the export and the page without anything looking wrong. On ten registrations were therefore looked up a second time through the state's interactive lookup — a different route to the same authority — and five fields were compared on each: registered name, registration number, date first registered, valid-until date and town — 50 comparisons in all.
Which ten is the half that makes it a check. They were not chosen: each of the 634 registrations published here was hashed with its own number and a fixed seed — nj-teyit-2026-08-23 — and the first ten of that ordering were taken, so the sample is reproducible by anyone holding the same file and was picked by nobody. Ten of ten matched the state on every field compared.
- 13VH07902700 — ISRAEL FLOORING LLC
- 13VH08760500 — KFM FLOORING LLC
- 13VH03309500 — MASTER WIRE MANUFACTURING, INC. T/A PREMIER VINYL COMPANY
- 13VH07733700 — SELECTA FLOORING LLC
- 13VH13357700 — GREAT CHOICE FLOORING LLC
- 13VH12567800 — AVA HARDWOOD FLOORING PROFESSIONALS LLC
- 13VH11503700 — BRUNO COLAHAN INC DBA FLOOR COVERINGS INTERNATIONAL METRO NJ
- 13VH07758800 — FERRY CARPET INC
- 13VH04472000 — J. S. HARDWOOD FLOORS
- 13VH13115500 — CARPET KINGDOM RUGS & INSTALLATIONS LLC
What that shows and what it does not. A drawn sample of ten out of 634 entries shows the download arrived intact and was copied faithfully; it does not show that every entry here is right. One field was left out of the comparison altogether — the county, which the live lookup does not publish — so the register's county column has been checked against nothing.
The lookup also returns a status, and it is reported apart from the comparison rather than counted in it: this file only holds registrations the download had already filtered to active, so the copy's side of that comparison was settled before it was made. What the lookup returned is still worth having — Ten of ten were active at the source on the day it was read — but that tests the state today, not the copy kept here.
An earlier version of this page offered three registrations instead. They were the first three rows in alphabetical order, and a sample taken off the top of a list carries no claim about the rest of it; this site's own record of that exercise said so before the page did. If an entry is wrong, say so and it will be checked against the state and corrected.
Who read all this, and what they sell, is on the about page — it matters here, because the publisher works in the trade this site writes about.