Raw Histogram??

I would prefer "lightening" to "brightening".
If we use "brightness" for the scene and "lightness" for the viewed image (on-screen or print) things can be kept clearer.

Don

For raw sooters, 18% is too high for digital sensors, which clip, as opposed to negative film, which has a shoulder. There's an argument to be made that 12.5% is too high.

I normally use matrix or evaluative metering. Hopefully this avoids clipping.
If you have a camera with a more sophisticated metering system, why not use it ?

Don

I have to admit that "18%" is lodged in my brain from the days of film.

Don

In my Z 6, there are two matrix metering modes: regular, and highlight-weighted. The real difference is to where in the scene the exposure is anchored; in regular, it's to the middle gray, in highlight-weighted it's to the assemblage of highlights. I would surmise that, if your matrix mode doesn't have any qualifiers, it's anchored to middle gray...

I use my highlight-weighted matrix mode by default, knowing that it'll preserve the highlights I want preserved. It does let go things like intense in-scene light sources, but that's okay by me. @thom in his user guide recommends also setting a +2EV exposure compensation to take back some of the headroom it leaves, getting closer to a ETTR exposure. I haven't played with that yet, I just work the tone curve in post...

The more sophisticated the metering system, the harder it is to figure out what it's doing. And, IMO, the best ways to get the exposure right for raw are the zebras and the histograms, not any of the metering systems.

Doug Kerr addresses 12.5% vs. 18% here ...

... and in much more detail here

In my view, Kerr's opinion in such matters trumps all others but, for Beginners, he can be a little too detailed.

I took plenty of good photos with a Sony NEX-5N, but the Sigma cameras that I use now are I think nicer to use. Never used a Sony FF camera, but I've seen plenty of good photos from them taken by other people.

Don

Only if you lose interest in where the raw highlight clipping occurs at higher ISOs. You may lose interest, because "only read varies" if you constrain exposure and vary ISO setting, but someone else may be interested precisely because read noise varies.

Practically, the value of "ITTR" (ISO to the right for a fixed absolute exposure) is more certain with consistent ambient lighting and manual ISO, like a situation where the subject is a person sitting in a sidewalk café at night and there are colored neon lights in the scene whose colors you want to maintain without clipping. In that case, you can chose your exposure time and f-number, and dial in the highest ISO that doesn't clip the neon lights, and shoot way until the sun starts coming up. If you are using low absolute exposures because you are photographing a bird that jumps from shade to sun to shade with a slow and from silhouette to highlight and back again, high-magnification lens with action-stopping shutter speeds, then trying to ride the clipping point is far more hazardous, and you might want to step back from clipping with some safety margin.

The 18% card positioned as Kodak are suggesting reflects ≈12.7%.
For meter calibration constants K=12.5 and C≈330 (Canon, Nikon):
reflectance R = 100 * π * K / C ≈ 12%.

The statement that the lightest element in the scene has a reflectance of 100% seems to ignore specularity.

It is an assumption made for the purposes of the following text in the link.

Beginners please note that "the statement" does not apply to all scenes in the Real World

Or even most real world scenes.

Per Jones & Condit, 'brightness scale' ('dynamic range') of typical outdoor scenes varies from 4.75 EV to 9.57 EV.

The Brightness Scale of Exterior Scenes and the Computation of Correct Photographic Exposure
LOYD A. JONES AND H. R. CONDIT
Kodak Research Laboratories, Eastman Kodak Company, Rochester, New York

Covered here

Beginners please note that owning and using a Kodak 18% gray card in the manner referred to by @IliahBorg would be almost unheard of these days, although the principles mentioned by Kerr remain true to this day.

I know ;)

Does it relate to this?

www.aghost.net/images/e0186601/ahistorylessonofrailroadtracks.pdf

It is most definitely incorrect, technically, to conflate "relatively ISO-invariant" with "totally-ISO-irrelevant" or "ISO-less", but there are differences in opinion, based on individual experience, that cause people to draw the line at different points in the sand as to whether the difference is practically relevant. Jim and I probably see the underlying empirical facts in a similar way, but we have very different editorial comments on practice, resulting from our own different subject matter.