The following is a Mavic 517 rim, used for off-road cycling in conditions that are frequently wet and/or muddy.
When this rim was new, the brake track was probably at least 1.4mm in the thickest part and probably over 1.0mm in the thinnest part. At least, the diagram suggests the thinnest part of the brake track is still thicker than the sloping part of the spoke bed, and the latter is about 0.90mm.
As shown here, the thinnest part on the left is about 0.35mm; it is about 0.50mm near the sloping part, and varies from about 0.80mm to 1.15mm between the web and the hook.
On the right, the thinnest part is about 0.60mm; about 0.70mm near the sloping part, and from 0.75mm to 1.15mm 0.35mm; it is about 0.50mm near the sloping part, and varies from about 0.80mm to 1.15mm between the web and the hook.
Note that some parts on the right are much thicker than similar parts on the left. This could be due to uneven thickness in construction and/or uneven wear in braking, likely both.
The rim is made by pushing hot aluminum through an extrusion die. The die is initially made to close tolerances. As the die wears, the rims that it produces have thicker and thicker walls. The die may also wear unevenly, but wear should be mostly even.
This suggests variability is due only to wear. However, some rims have brake tracks machined at the end of assembly, and machining is less accurate than extrusion, leading to brake tracks of uneven thickness.
For example, the following shows a section of a Sun "Zero Degree" rim, which is nearly new and still shows machining marks on the brake track. At the hook, the left brake track is about 1.25 mm, while the right brake track is about 1.45 mm. At the spoke bed, the left and right are about 1.15 mm and 1.35 mm.
Note the section shown here is taken near the welded rim joint. The rim joint causes misalignment. With a classic pinned and non-machined joint, misalignment does not lead to thinning. With a welded joint it is hard to machine away weld flash without also removing some sidewall material.
If the rim was non-machined and had an original extrusion thickness at the bed of 1.25 mm and if the rim must be replaced at 0.25 mm, that means the rim tolerates 1.0 mm of wear. The machine rim has average thickness of 1.25 mm, but is as thin as 1.15 mm, and the rim still has to be replaced at 0.25 mm, so that means there is only 0.90 mm to wear away, or a loss of 10% of service life during manufacturing.
If the rim is replaced more conservatively at 0.40 mm, the reduction is from 0.85 mm of wear to 0.75 mm of wear is a 12.5% reduction in service life. If the rim is replaced at 0.6 mm, that represents a 15% reduction in service life.