A doorbell camera gives you a hood and a shoulder.

For seeing who is at the door in real time and getting a rough record of deliveries, it is genuinely fine and we install them. For producing an image somebody could be identified from, it usually is not, because of where it is mounted and how wide its lens is. Chest height plus a wide field means a face occupies a small part of the frame and is often shaded by a hood or a hat brim.

The upgrade that changes the outcome is not a better doorbell. It is one additional camera at head height, positioned across the approach and framed on the path rather than on the door, with a narrower field of view so the pixels concentrate on a person rather than a porch. On a stoop that is often a small camera under the entry cornice; on a suburban house it is usually on the garage face looking back at the walk.

Two cameras chosen for two different jobs will outperform four cameras all doing the same one, and it is normally cheaper too.

More than most homeowners assume, provided it goes through review first. Anything visible from the street on a designated facade needs Landmarks approval before a Buildings filing, and most entry and camera work falls under the two lighter permit routes rather than a full hearing.

What tends to get approved: equipment mounted within an existing recess or under an existing cornice or canopy; finishes matched to the surrounding material; a panel that fits within the opening the previous one left; cable routed through existing penetrations. What tends to cause trouble: a larger faceplate than the original opening, new surface-mounted conduit on the facade, and anything that reads as an added object on original stone or ironwork. On a corner house the side elevation is reviewed too, which catches people out.

The practical advice is to settle the physical size and finish before choosing a product. Picking a panel first and then discovering it is two inches wider than the existing opening is how these projects lose a month.

Three things, in this order. First, the safety devices. Since 2016 photo eyes and contact edges are monitored by the operator, so a lens fouled with pollen or road film, a bracket knocked by a delivery truck, water in a connector or a cut wire will stop the gate dead. That is the operator refusing to move because it cannot prove it will not close on someone. It is working correctly, and it must not be bypassed.

Second, power and connectivity at the gate. A gate 200 feet from the house needs power out there and, if it has an intercom or a camera, a data path. That is the real cost of a gate project and it is what a survey is establishing. Copper Ethernet will not reach past about 328 feet, and copper between two structures invites surge damage. On a long driveway, fiber is usually the right call.

Third, the gate itself as opposed to the operator. There is a separate standard governing gate construction (protrusions, gaps, climbability), and an operator can be fully compliant on a gate that is not. If we survey and find that, we will tell you in writing before quoting an operator, because automating a non-compliant gate moves a maintenance issue into a liability one.

Sometimes, and it is worth twenty minutes on the survey to find out. The routes that avoid a trench, roughly in order of how often they work: an existing conduit put in for garage power or an old landscape lighting circuit; a soft route around the driveway edge through a garden bed rather than under paving; an overhead run where one already exists between structures; and a point-to-point wireless link with the camera powered locally at the garage.

Wireless is the fallback rather than the first choice, and we would rather be honest about why. It depends on line of sight, it degrades in heavy weather, and it needs power at the far end anyway, so it removes the data trench, not the electrical one. Where a garage already has power, that is often exactly the right trade.

If none of those exist, the crossing is a bore or a saw cut, and that is a real number that belongs in the quote before you commit rather than as a discovery afterwards.

No. We are not a monitored alarm company and we do not provide central-station service. What we install sends notifications to you and to whoever else you choose, and records to equipment you own or to a platform registered in your name.

For plenty of households that is the right arrangement: a phone alert with a clip attached, from a system nobody is paying a monthly fee to watch. Where you genuinely want a monitored intrusion alarm with a dispatch response, that is a separate licensed service and we will say so and point you to it rather than pretend our scope covers it.

What we do build into every job is the assumption that nobody is watching. Recording is continuous or event-based rather than dependent on someone noticing; notifications go to more than one person; and analytics that distinguish a person from a car from a branch moving in the wind do the filtering, so the alerts you get are few enough that you still read them in month six.

Start with the framing rather than the argument. In most cases a rear camera can be repositioned along your own boundary instead of across it and still cover the access path you care about, often better, because a camera looking along a fence line sees anyone moving down it, while one looking across a yard sees a small figure in the middle distance.

Where the geometry genuinely forces some of a neighbor’s property into the frame, privacy masking blanks that region permanently in the recorded image, not just in the live view. We set it at commissioning and note it in the handover so it is documented rather than something you have to demonstrate later. It is a small thing that has resolved a number of these conversations.

And keep the microphone off, which we do by default anyway. Audio is what turns a neighbor’s objection from a preference into a legitimate grievance, and it adds nothing you need.