Is lubing switches actually worth three hours, or is it something people do because the build is finished and they miss it?

It does something real, it is not the biggest thing you can change, and which switch you have decides almost everything about whether it is worth your evening.

What lubing actually fixes is friction noise — the scratch of the stem sliding in the housing. If your switches scratch, you will hear the difference immediately and it is not subtle. If they are already smooth, you are polishing something that was fine and the result will be exactly the disappointment you are worried about.

So test before you commit. Pull four switches, lube two, put all four back next to each other and type. Ten minutes, and it answers the question for your specific switches better than anyone here can.

The other half is by type:

• Linears benefit the most and are the easiest to get right. This is where the transformation talk comes from.
• Tactiles are risky. Too much on the legs rounds off the bump you presumably bought them for, and people ruin switches this way.
• Clicky — do not. You will muffle the mechanism that is the entire point.

And the honest ranking: stabilisers matter more than switches. A rattling spacebar is the loudest bad noise on most boards, it is four or five parts, and it takes twenty minutes. If you only do one thing tonight, do the stabs. That is the change people actually notice when they pick up your board.

The dissenting answer, which I think is worth having in the thread.

I type on stock switches on a board I built deliberately. Not because lubing does nothing — it does — but because the honest size of the difference, on a board you use rather than record, is smaller than the amount of discussion around it implies. It is a real improvement in the same category as a good desk mat: noticeable in a direct comparison, invisible within a week of normal use.

The reason the videos oversell it is structural rather than dishonest. A microphone thirty centimetres from a board in a quiet room hears differences your ears never encounter at typing distance in a room with other noise in it. And a side-by-side recording removes the one thing that actually governs your experience, which is that you adapt to whatever you type on in about three days.

If you enjoy the process, do it — that is a completely legitimate reason and most of the hobby runs on it. If you are doing it because you expect the board to feel like a different board afterwards, do the four-switch test above first, because the answer for a lot of modern switches is that they are already smooth out of the box in a way they were not five years ago.

Seconding stabilisers first, and adding the thing that gets skipped: a lot of people who want lube actually want films.

If your switches wobble slightly — a bit of play in the housing when you press off-centre — that looseness produces a hollow, slightly rattly sound that no amount of lube fixes, because it is a fit problem rather than a friction problem. A thin film between the housing halves tightens it up. Five seconds per switch while you have them open anyway.

On technique, since the failure modes are consistent:

Use much less than you think. The commonest bad outcome is not a bad-sounding board, it is a sluggish one, from too much lube on the rails. A thin haze, not a coat.

Springs are where most of the audible win is and where bag-lubing is genuinely fine. Spring ping is the ringing you hear after the keypress, and it is the noise most people are actually trying to remove when they say they want a quieter board.

Do the housings and stems by brush if you are doing them at all. Bag-lubing whole switches is how people end up with the sluggish result.

Budget it as: stabs, springs, films, then stems and housings — in that order of return per minute. Most people stop after three and are happy.