The layer you cannot change later without opening the ceiling again

Decide it per outlet rather than per building, because the cable is a small share of the cost and the labor is identical either way. Cat6 handles 1 Gb to the full channel and 10 Gb to about 55 m, less in a tight bundle or a long parallel run, where crosstalk from neighboring cables eats the margin. For a camera drop, a card reader, a desk phone or a printer, that is plenty and will remain plenty.

Cat6A earns its price in three places: wireless access point drops, because access points are where multi-gig ports actually appear; switch uplinks and anything running between closets; and any bundle carrying high-power PoE, because the larger conductors dissipate heat better and de-rate less. Cat6A is also the only copper category whose certification requires alien crosstalk testing, which is exactly the failure mode that bites dense bundles. A sensible building is usually mostly Cat6 with Cat6A where those three cases live.

One hundred meters for the whole channel: 90 m of horizontal cable in the building plus a combined 10 m of patch cords at both ends. Measured along the actual route the cable takes, not across the floor plan. A run that looks like 70 m on a drawing can be 110 m once it goes up a wall, across a ceiling, around a shaft and back down.

Beyond that, the options in order of how often we use them: a fiber leg to a small switch nearer the devices, which is the right answer surprisingly often and is not expensive; a mid-span extender where the overrun is small and the environment is benign; or relocating the IDF. For anything reaching a separate structure (a gate house, a detached garage, a parking deck, a second building) fiber is the answer regardless of distance, because a dielectric link carries no ground potential difference and no surge between the two structures. Copper between buildings is one of the more reliable ways to destroy equipment at both ends during a storm.

It matters, and the test is not the price of the cable, it is what the ceiling space does. If the space above the drop ceiling is used to move return air back to the air handler without ductwork, that is a plenum and it requires plenum-listed cable. If the return is fully ducted, it is not a plenum and riser or general-purpose cable is permitted. You cannot determine this by looking up at the tile; it comes off the mechanical drawings or from the building’s MEP.

Getting it wrong in the cheap direction is expensive, because the remedy is not a fine, it is removing and replacing the cable. Getting it wrong in the safe direction costs money for no benefit. Note also that substitution only runs one way (plenum cable is acceptable in a riser or a general space, but riser cable is never acceptable in a plenum) and that the same rule applies to access control and intercom cable, which is a different listing family and gets forgotten constantly.

It is a test, and it is the difference between cable that is known to be good and cable that currently links. A field tester measures the parameters that predict future failure rather than present function: insertion loss, near-end and far-end crosstalk, return loss, delay skew, and on Cat6A the alien crosstalk between adjacent cables. Plenty of runs with a crushed tie, an over-tight bend or a marginal termination will link at 1 Gb today and fail certification, and those are exactly the runs that go intermittent under PoE load two years later.

Commercially, certification is also what unlocks a manufacturer’s 20 to 25 year materials-and-labor warranty; no major manufacturer warrants uncertified plant. It adds roughly 5 percent to installation cost. We hand over the test result file, so you own evidence of what was installed rather than an assurance that it was fine. If a cabling quote does not mention certification, that 5 percent is where a competitor found room to be cheaper.

Often, and it is the first thing we check, because in a finished building the pathway is the cost and existing pathway is money already spent. We test what is there rather than guessing: category and jacket rating off the print on the jacket, length, and a certification test on each run we intend to keep. Cat5e that tests clean is perfectly good for cameras and readers and we will happily reuse it.

What rules it out: copper-clad aluminum, which we occasionally find and always replace because it cannot carry PoE current safely; cable with the wrong jacket rating for the space it passes through; runs that are physically damaged, crushed or over-length; and anything terminated so badly that re-terminating is more work than pulling new. In a legacy riser we also look for whether new cable can even use the same route, because a 1926 shaft packed with sixty years of abandoned cable sometimes has to be cleared before anything new can go up it. That clearing work is a scope item that other quotes will not have included.

Because a drop is not a unit of cable, it is a unit of access. The same Cat6 outlet costs $100 to $150 in new construction before the drywall goes on, and $300 to $400 in a finished occupied space where somebody has to open a ceiling, work around furniture and put it all back. Fiber runs $300 to $900 a drop for the same reason plus termination.

Volume moves it as much as difficulty. Mobilization, minimum-day labor, and the time spent getting a crew and materials into a Manhattan building spread across however many outlets you are installing, so a ten-drop job can genuinely be $400 a drop while a five-hundred-drop job in the same building is $175. When you compare two quotes, the questions that explain most of the gap are: what jacket rating did each of them price, is certification included, is firestopping included, is abandoned cable removal included, and does the number assume normal hours or a freight-elevator window at seven in the evening. All figures here are typical ranges and subject to survey.