Four cameras in the right places beat twelve in the wrong ones.

For a single-room storefront under about 1,200 square feet, four to six is usually right: entrance tight at head height, register overhead, rear or stock room door, and one or two wide for context. Adding cameras beyond that generally adds storage cost and viewing time without adding answers, unless the layout has genuine blind corners or a second entrance.

The question that actually sizes it is not square footage but choke points. Every place a person must pass through (the door, the counter, the corridor to the back) is worth a camera. Open floor is worth one wide view, not four.

For budgeting, a properly installed small commercial system of four to six cameras in New York typically lands somewhere in the low thousands to around six thousand dollars all in, with the spread driven almost entirely by how hard the cable routes are rather than by the cameras. That is a typical range, subject to survey. A shop with an accessible drop ceiling and one with plaster over lath and no pathway are different jobs at the same camera count.

Honestly, not much on their own, and any installer who promises otherwise is selling. Visible cameras displace some opportunistic theft and are useful for that. What they reliably do is resolve incidents: give you an identifiable image, a timestamp, and a sequence you can hand to police or to an insurer.

Where cameras do change loss meaningfully is internal, and specifically at the register. A camera framed on the drawer and the hands, timestamps aligned with your point-of-sale export, and a habit of spot-checking voided and no-sale transactions catches patterns that no amount of floor coverage will. That is a process change with a camera attached, not a camera purchase.

The other genuinely effective measure is unglamorous: a door position switch on the rear and cellar doors with an alarm when they are held open. Most back-of-house loss goes out through a door that was open for a legitimate reason and stayed open.

Common, and usually solvable. First establish whether it is a lease restriction or a landmark one, because they have different answers. A lease restriction is negotiable and often becomes a written consent with a make-good obligation at the end of the term. A landmark district restriction is not negotiable and involves a review before any Buildings filing: anything street-visible on a designated facade, including a camera housing, surface conduit or an exterior reader.

Where the exterior is genuinely off limits, the workable approaches are mounting inside the glass line looking out (which brings the infrared and glare problems above, so plan for external white light and true wide dynamic range), or mounting on the underside of an existing awning or recess where the fixing is not into the protected material, or coming through an existing penetration rather than a new one. Getting a camera out of a shed-covered facade often means the second approach.

Default answer: no, and we ship systems with microphones disabled. New York is a one-party consent state, which sounds permissive until you notice what an unattended camera actually is: a device recording conversations between people, none of whom is the recorder. That is mechanical overhearing without any party’s consent, and it is a felony rather than a fine.

The narrow situations where audio can be appropriate involve a participating party (for example a staffed intercom where the employee is part of the conversation), or clearly posted notice. Even then it is worth asking what the audio buys you. In practice, almost every dispute a small shop needs footage for is resolved by the image and the timestamp, and the audio adds legal exposure without adding evidence.

The exterior ones probably are, at least for anything above knee height, and there is no clever setting that fixes it. What we can usually do is relocate a camera to the shed’s underside or to the fascia so you keep coverage of the entrance and the sidewalk immediately in front of it. Sheds have a deck and posts that are generally workable mounting points with the contractor’s permission, and that permission is worth asking for early because scaffold contractors are more agreeable before the shed is fully loaded.

The longer-term lesson is a design one. When we install an exterior position we prefer to leave a service loop and, where the wall is already open, a spare pathway to a second location. It costs very little at install time and it converts a shed from a lost year of coverage into an afternoon of relocation.

One upside worth knowing: sheds add lighting and change how people move on the sidewalk, and a camera relocated under the deck at head height often produces better identification footage than the original higher mount did.

For high-value, low-footfall retail, generally yes. Its value is behavioral rather than mechanical. A locked door with a person deciding who comes in changes who tries. It costs you walk-in browsing traffic, which for some businesses is the entire model and for others is irrelevant, so that trade-off should be made on the retail side before it is made on the security side.

Two things to get right if you do it. The release control belongs where the person watching the door actually stands, with a second one at the counter for single-staff hours. And the exit has to remain free: whatever is on the door, a person inside must be able to leave with one motion and without a key, a code or special knowledge. Any arrangement that fails that test is not a security choice, it is a violation, and we will not install one.