Upload & Sell: On
If you use your color checker and look at the black uniform (assumed to be black), you see various areas where the blue channel is stronger than the red & green channels. If this is indeed a neutral black, then all three channels should be closer to the same value.
The highlight values (particularly when above 235) can appear to be neutral while the mids/shadows are not. This is why I like to use the color balance layer rather than levels or curves for WB adjustment. It readily allows for addressing the three different areas ... noting that even when adjusting shadows, it may have some impact on mids/highs (just lesser so).
Also, the closer an area is to 255,255,255 the less impact changes in other areas can make to the white. The point here is that just because a white or specular is neutral, that doesn't mean the scene is neutral. You really have to check your shadows and highs independently. Once highs and shadows are both neutral, mids usually fall in line okay (assuming same lighting).
Working off really bright highs and really dark lows can give a false sense of neutral some times, so I try to find other areas to check that aren't quite at the extremes. I typically try to dial in my neutrals so that each channel is within 5 point or so of each other. In some cases with mixed lighting this isn't always practical, and I'll either go with my majority rules or pleasing to the eye whenever WB is conflicted.
Finding "true" neutrals in a scene isn't always a sure thing when you're shooting without a reference (gray card/color checker), so I have to ask which portions of the scene do I think (or want) should be neutral. Recognizing that we can be shooting in mixed lighting anyway, there can be a bit of "pick & choose" which part of the image/scene is most important regarding its neutrality. That, and sometimes masking at mixed lighting areas is useful.
+1 @ Karen that the variance is not obvious to casual observance (on my monitor) in the first image. For that reason, I've left the two images unidentified.
I'm curious though @ what it was that the others "keyed in on" regarding their assessment of WB being off (in the first one)?
In the second image however, things might be a bit more obvious as we check and see quite a bit of blue in both the lights and darks. If you compare the number 8 in the before vs. after you'll note that the original has about 40+ points more blue than R&G. In the revision, it only has been reduced to a 10 point variance (i.e. less noticeable).
Both images could probably benefit from some additional refinement ... noting that the grass is a different color in each scene. I'm guessing they didn't change the lights or the grass between shots, so at least one of them has to be kinda wrong.