Whole-Home Automation at the Crestron & Lutron Tier: What You're Actually Buying

Almost every house in Beverly Hills, Bel Air and the Palisades now has smart devices in it. Very few of them have a control system. The difference isn't brand snobbery — it's an architectural one, and it shows up the first time three people try to use the house at once.

A house full of smart devices is a collection of independent products, each with its own app, its own cloud account, its own firmware schedule and its own failure mode. A control system is one processor that owns the house: it knows what state every load is in, it responds to a physical button in under a tenth of a second, and it keeps working when the internet goes down. Crestron and Lutron HomeWorks sit at the top of that second category; Control4, Savant and Lutron RadioRA 3 sit a tier below it and are frequently the smarter buy.

Here's what actually separates the tiers, and where the money goes.

The Decision That Sets Everything Is Lighting

Nearly every luxury residential system we design is really a lighting control system with entertainment attached. Lighting is the subsystem the client touches two hundred times a day, it's the one with the most wire in the walls, and it is the hardest to change later. Pick the lighting platform first; almost everything else follows from it.

The jump from RadioRA 3 to HomeWorks QSX is the jump from dimmers in the wall boxes to centralized dimming panels in a closet. That matters for two reasons. Panels take the heat, the noise and the serviceable parts out of the living space, and they let you drive a single fixture group from one properly matched dimming module instead of hoping the in-wall device in the switch leg is right for the load.

The mistake: Choosing the platform by home square footage alone. A 6,000 sq ft house with 140 dimmed zones, tunable-white circuits and eighteen shades is a HomeWorks house. A 12,000 sq ft house with mostly switched loads is not.

What works instead: Count loads, not square feet. Have your integrator produce a load schedule from the electrical drawings — every circuit, its fixture type, its driver, and whether it's switched, dimmed or tunable — before anybody quotes a platform. That single document prices the job.

Flicker Is a Specification Failure, Not a Product Defect

The most common complaint we get called in to fix on other people's installs: lights that flicker, buzz, drop out at the bottom of the dimming curve, or won't go below 20% without shimmering. It's almost never a bad dimmer. It's a dimmer and an LED driver that were never specified as a pair.

The mistake: Letting the lighting designer pick fixtures and the electrician pick dimmers, with nobody reconciling the two. By the time it flickers, the drywall is closed.

What works instead: Require the fixture cut sheets with the driver model called out, check each one against the control manufacturer's compatibility list, and specify a dim-to-1% driver anywhere the client will actually sit in low light — bedrooms, theaters, dining. Then buy one of each fixture and bench-test the pairing before the order goes in. That afternoon costs a few hundred dollars and saves a re-lamp of the whole house.

Keypads Are the System; the App Is the Backup

Every integrator demos the app. No one who lives with a good system uses it much. The app is for the client who's away, the guest who needs the pool lights, and the specific weird request. Daily operation happens at engraved keypads placed where a person's hand already is.

Three rules make or break this:

• Engrave in the client's language. "Entertain," "Dinner," "Goodnight," "Away" — not "Scene 3." If a guest can't read the button and guess what it does, the design failed.
• The top button is always all-on and the bottom is always all-off, in every room, without exception. Consistency is worth more than cleverness.
• Keypads go where the doors are. The design question is the path a person walks, not the wall the electrician found convenient. Every entry to a room gets control of that room.

The mistake: A 12-button keypad in a room with four things to control. Extra buttons don't add capability — they add hesitation.

What works instead: Design the keypad layouts on paper with the client before the order, room by room, and keep most keypads at four to six buttons. Order the engraving with the hardware; field-swapping buttons later is a lead-time item, not a service call.

Shading Is the Subsystem That Runs Out of Room

Motorized shades are where residential projects most often collide with architecture, because the decision has to be made before framing and nobody thinks about it until finishes.

A single roller needs roughly a 5" pocket; a dual roller — sheer plus blackout, which is what most primary bedrooms and great rooms actually want — needs closer to 8" with the fascia and the light gap accounted for. If that pocket isn't in the framing drawings, you're choosing between surface-mounted brackets on a $40,000 window wall or no blackout at all.

Two more that matter more than clients expect:

• Openness factor is a view-and-glare tradeoff, not a style choice. 1% blocks the most glare and gives the least view; 5% is the usual living-space compromise; 10% on a west-facing Pacific-facing window will not solve the 4 p.m. problem no matter what the fabric sample looked like indoors.
• Battery vs. wired. Battery shades are genuinely good now and right for retrofits, but a shade that cycles twice a day on a timeclock will eat batteries. Anything on an automated schedule, anything over about 8 feet wide, and anything you'd need a lift to reach should be hardwired.

The Network Is Part of the System, Not a Prerequisite

The single most common cause of an unreliable luxury system is a consumer mesh network doing a job it was never designed for. A whole-home control system is dozens of always-on devices holding persistent connections, and it belongs on infrastructure designed for that — which is the same reasoning behind planning structured cabling at rough-in rather than after.

What we specify on every residential system of this tier:

• Managed switches with VLANs separating control, AV streaming, cameras, IoT and guest traffic. Cameras and IoT are the two categories most likely to misbehave; they don't belong on the same broadcast domain as the processor.
• A real PoE budget. Add up what's actually drawing: 802.3at access points at up to 30 W, touch panels, cameras, doorstations. Size the switch for the total plus about 30% headroom — a switch that browns out its ports under load produces symptoms that look like anything but a power problem.
• Static addressing for every control device, by DHCP reservation. A processor that finds its touch panel at a different address after a power blip is the single most common "the system stopped working" call.
• A UPS on the head end — processor, core switch, modem and router — sized for 15–20 minutes. Most outages in LA are brief, and riding through them is the difference between an invisible event and a house-wide reboot.

Heat, Power and the Rack Nobody Planned For

An equipment rack for a home at this tier dissipates real heat — commonly 800 W to 2 kW of continuous load once amplifiers, processors, switches and video distribution are in it. That is a space-conditioning problem, not a closet problem. Plan for a dedicated exhaust or mini-split, not a louvered door, and give the rack its own 20 A circuits on a sequenced power conditioner so everything starts in the right order after an outage.

Leave physical room, too. The rack that's exactly full on day one has nowhere to put the thing the client asks for in year two — and at this tier there is always a year-two request.

Programming and Handoff Are Half the Value

Hardware is a commodity; the same Crestron processor is available to everyone. What you're actually buying is the design work, the programming, and the documentation that lets someone service the house in five years.

Insist on a closeout package: as-built drawings, the load schedule as installed, an IP address and credential list, the programming source files, and labeled everything. If your integrator won't hand over the program, you don't own your house's control system — you're renting it from them.

And budget for the first ninety days. Every real system gets tuned after the family lives in it: scene levels are wrong at the actual time of day people use them, a keypad is one room off, the shade schedule needs to move an hour. Any integrator who treats the day of substantial completion as the end of the job has mispriced it.

Where to Start

If you're planning a new build or a gut remodel, the useful sequence is: load schedule from the electrical drawings, then platform selection, then keypad layouts, then shade pockets into the framing set, then network and rack space into the architectural set. Everything on that list is cheap on paper and expensive after drywall.

Handsome Group designs and installs whole-home control, lighting and AV systems across Los Angeles' Westside and the surrounding communities — including for clients rebuilding in Pacific Palisades and Malibu, where getting infrastructure into the drawings early is the whole game. If you're weighing platforms, we'll tell you honestly which tier your house actually needs, including when the answer is the less expensive one.