Removing a fixed amount equal to the head mill is only helpful if:
1. this is the only modification
2. the geometry was perfect before (no wear, aftermarket replacement parts, seat recession)
If the cam is changed, in theory the pushrod length should be shortened (in addition to any correction made to cylinder head or barrel height) by 1/2 of the added lobe height (not lift).
Example only:
Original lobe .350"
New lobe .400"
Difference .050"
Shorten pushrod .025"
However, this is only true if the base circle (zero lift) is unchanged.
Since the cam manufacturers, when asked "do you reduce the base circle by 1/2 of the added lobe height to correct the geometry?" answer (quoting, literally) "................................................ what?" you won't know what you have unless you either mike the new cam for both base circle and lobe height and compare it to an NOS or perfect cam, or (better) assemble the engine with the new parts and analyze the geometry.
The welded hardface cams such as Norris and Web-Cam probably use the stock base circle - but I don't know this as fact, it's just more likely.
Reground cams (no weld) cannot use the stock base circle, because the new lobe profile and height are produced by removing metal from the base circle - there's no other way. A change of duration, lift, centerline or any combination of these will always reduce the base circle diameter.
Warning: analyzing the geometry is one of the most misunderstood, badly diagnosed problems on the net. Everybody and his dog thinks it's simple to understand, the original factory stuff was fine, and that it can all be handled with adjustments and/or pushrod length.
Wrong, wrong, wrong.
Rather than give you 10,000 words (no, I'm not kidding) if you assemble the engine some of the easiest things to look for are:
1. adjuster runs off the tip (duh)
2. can't get an adjustment
3. angle for valve closed (lift = 0) is completely different from valve fully open (lift = 100%)
4. the adjuster only moves in one direction during the lift
5. adjuster does not pass through the center of the stem tip during lift
The "correct" geometry is perceived to be when a(n) (imaginary) line drawn between the adjuster contact point and the rocker shaft centerline is at 90° with the valve stem axis at 50% lift. This is called the "mid-lift" or "Miller" method, and it's not the only method, not necessarily the best method, but it works and is the easiest to plan and measure.
If this is correct, the adjuster's "scrub path" or what pattern it makes on the stem during lift, will start slightly inboard (closer to the shaft) of the valve stem tip center at 0 lift, walk across (away from the shaft) until 1/2 lift, then walk back. If it does not do this, it's wrong.
The positions need not be centered (far less important, and cannot be cured by adjustment anyway), but stopping and reversing at mid-way is key.
If you can't actually measure this, anything that shortens the length of the scrub path is good.
You can simulate a shorter pushrod by shimming the rocker box up, use the smallest increment shims your patience will stand (.005" is a waste of time, .060" is too big).
You can simulate a longer pushrod by placing a small piece of metal scrap between the tappet and pushrod end, or pushrod and rocker end. I suggest aluminum, since you can easily dent it with a punch to match the curvature. Cut a small circle, then saw a slit from 1 side to the center to make it easier to dent.