First: what actually happens in the room
Before I look at a single rail profile, I spend twenty minutes just watching or remembering how the room gets used. That sounds soft, but it's the cheapest step in the whole job. In a cafe I did last spring, the chairs come out at seven in the morning and get dragged around all day, so any track over the seating zone has to survive people and mops underneath it, not just look neat on a drawing. In a small shop, the display changes every season and the owner doesn't want to call an electrician for it, so sliding heads matter more than output. In an apartment, the opposite — the track might never be touched again, and what matters is that it doesn't glare into someone's eyes from the sofa. Same product family, three completely different briefs. I've learned to write down the room's behaviour first — where people walk, where they sit, what moves, what gets cleaned, what the owner will realistically adjust themselves — and only then open the sample box. A system, at bottom, is a rail plus heads that click on magnetically and can slide, be added or taken off, plus the power and control bits that make it work; the breakdown in Magnetic Track Lighting System Architecture is the shape I copy onto my own plan before I start moving heads around on it.
The ceiling gets a vote
The ceiling decides more of this than any catalogue page does, and I've ignored that exactly once and paid for it. Recessed track needs a void with enough depth and enough clear space around it, and I've stood in a flat where the concrete slab was exactly where the plasterboard wanted to be. Surface-mounted track is fine almost anywhere but it hangs there in the room, visible, so you need to like the look of it. Suspended track is the rescue option when there's nothing solid overhead to fix into — but someone should check what's actually above before promising that. At home, the bedroom ceiling has a beam running through it that I somehow missed until the samples were already out on the kitchen table. Two days lost. Measure the void, note where the joists run, and mark the pencil on the plan where the track can physically go, not where it would be prettiest.
Picking the family is the real fork in the road
This is where the expensive mistakes live, so I'll be blunt about my own. The first time I specified magnetic track, I picked luminaire heads because they looked good — clean little cylinders, nice proportions — and only afterwards asked what they mounted to. Fine, that worked out. But the second time, I chose a high-voltage family for a shop where the owner wanted to hang heavier grille lights and flood lights over a counter, and I hadn't checked that the family could take that size of head. The small residential family, the one called M25, is built around smaller luminaires and domestic rooms; M35 is the bigger one, with larger heads and commercial spaces in mind. I now decide the family first — voltage and size class — and only then let myself look at heads; the fork itself is the one walked through in 48V vs High-Voltage Magnetic Track: A Practical Comparison, and it is worth reading before any head catalogue gets opened.
With the low-voltage 48V side it's a similar discipline. There's a 20mm system, there's a 26mm ultra-thin one whose rail is 26mm wide and just 6mm thick, there's woven textile track, and there's a semi-flexible track that bends around corners. Each of those answers a different problem. The ultra-thin one suits a ceiling where I want the track to almost disappear. The textile one I've used where the track is going to be visible in a warm room and I wanted it to read as part of the interior instead of a piece of extruded metal. Semi-flexible got me around a curved bulkhead in a small bakery — the track follows the curve, and nobody asks questions about it afterwards.
The point is that the pretty head is the last decision, not the first. If you fall in love with a spotlight before you know which rail family it belongs to, you'll either bend the room to fit it or start over. I've done both. The second one is cheaper.
Something else on the same subject, because I've been bitten: colour temperature lives in the family too, and it doesn't always offer you everything. In one family I know, the options are 3000K, 4000K, 6000K — no in-between. In a cafe I wanted something warmer than 4000K but cleaner than the very warm end, and I had to pick a side and defend it to the owner over a coffee. That's not a disaster, but it's the kind of thing you want to discover while you're still choosing, not after the heads are on site.
Glare, and the view from a chair
Here's the test I trust most: sit where the customer sits and look up. Not standing, not from the doorway — actually sit down. I did this in a small dining room once, after the track was already up, and from that one chair the spotlights were fine until I leaned back, and then every single one of them was a small bright dot straight in the eye. The track was correct. The heads were correct. The aiming was wrong. I re-aimed them all with a stepladder and a lot of muttering, and it was still the right fix, but it would have been free on paper.
So now the chair test goes into the plan. In a home, that means the sofa, the reading chair, the seat at the end of the kitchen counter where someone actually eats breakfast. In a shop, it's the customer's eye line toward the window and toward the till. Flood and linear heads throw a wider, softer pool; narrow spotlights are for pulling one thing out of the dark. Grille lights are the ones I reach for when I want a wall washed without hot spots. Pendants off a magnetic track are a different animal altogether — good over a counter or a table, but they change how the whole system looks, so judge them in the room, not on the floor. And aim everything before you call it done, because a track full of heads pointed straight down is the most common and most boring mistake in this whole business.
Where the power supply lives, and whether anyone can reach it
Nobody wants to think about this, and it's the one that ruins a good install. If the system needs a power supply, that box has to physically exist somewhere — in the ceiling void, in a cupboard, on top of a cabinet, in a service space — and it has to be reachable by a human being later. I had one cheerful conversation with an installer who sent me a photo from the top of a ladder showing the driver tucked behind a plasterboard panel that had been screwed shut. It worked. It was also, in any practical sense, gone. Someone will be up that ladder in two years with a torch, and it's better if that someone can actually get to the thing.
So on the plan I mark where the supply sits and how you get to it: which panel comes off, which cupboard it hides in, how much clearance there is around it. For a shop or a cafe, the supply should be somewhere the owner can find it without moving the counter. At home, I'd rather give up a little ceiling and keep the access. And I check the controls at the same time, because they need power too and they need to be within reach of whoever actually switches the lights on. In one family I've used 2.4G panels and remotes, and Bluetooth controls in the same kit — I tell people what the controls in front of them are and what they do in the room, and I don't promise app or voice control on top of that, because nobody has shown me it working on the kit in front of me.
What a useful sample test looks like
The kitchen table has been covered in samples for a week. That's the actual test, not the showroom. I tape two rail profiles to the ceiling where they'll really sit — one recessed, one surface — and I put one head on each and look at them from the doorway, from the sofa, from the angle where the sun comes in. Then I take the whole thing down and do it again in the evening, because a track looks completely different in daylight and after dark. For the 26mm ultra-thin rail, it was the thickness that decided it more than the width — 6mm sits almost flush, and from a chair you barely see it; the low-ceiling side of that argument is made properly in M26 Ultra-Thin Track for Low Ceilings, and it matches what the chair told me. For the textile track, the question was whether it read as fabric in a warm room or just as a stripe on the ceiling. Both answers were useful, and neither came from a photograph.
One more thing worth doing while the samples are out: handle them like a customer. Slide a head along the rail. Add one, take one off. See how it feels, not just how it photographs. Magnetic attachment is the whole point of the system, and if the sliding is fussy with cold hands in a hurry, that's information.
The drawing you send for a quote
By the time I send anything out, the ceiling plan has pencil marks all over it, and that's the version I convert into a proper drawing for quoting. The plan needs the rail runs drawn to scale, with dimensions from something real — a wall, a beam, a window — not from the edge of a photo. It needs the mounting type marked per run: surface, recessed, suspended. It needs the family and the rail profile named, so nobody substitutes a different rail because it was easier to get. It needs every head drawn in position with its type, and the aiming noted roughly, and the colour temperature written in for the room it's serving. Then the power supply location, the access route to it, and the control position with the switching logic. A schedule of quantities that a supplier can price without phoning me. My own checklist looks a lot like the Magnetic Track Buying Guide, only scruffier and with more crossings-out. And a note at the bottom telling the installer to check the void and the fixing before cutting anything, because the drawing is my hope and their ladder is the truth.
Last week I stood in that same room with the old pencil mark, and the chair had moved again. This time I drew the track around where the chair actually ends up, not where I wished it would stay. Then I sat down in it, looked up, and for once didn't flinch.