Single pattern ("normal") cams presume that the flow efficiency of the intake port at .75 psi of vacuum (= 1.5" Hg @ WOT throttle) is roughly equal to the flow efficiency of the exhaust port a 50 psi or more (exhaust pressure when the valve cracks open, dropping to almost nothing by TDC).
If it's not (one port works better) adding duration and/or lift to the weak side balances this out.
If measured at the same constant 28" H20 pressure (which never happens in an engine) the exhaust CFM flow is about 70-80% of the intake flow. E.g., intake is 150, exhaust should be 105 to 120 CFM, etc. If the exhaust is 140 (93%), the intake duration should be extended, or the test re-done for higher lift, or both.
For a bad intake port, both lift and duration generally help (which one is more depends on the head design).
For a bad exhaust port, only duration has much effect: open the valve earlier; if still more is needed close it later.
Flow work posted here shows the exhaust port to flow substantially less CFM than the expected 70-80% of intake flow, so extending the exhaust duration works.
How much to extend it depends on how much work you've done to either port.
If only the intake has been modified, you need even more help on the exhaust side. If only the exhaust, less bias.
Your head porter's bench should be your last stop before shopping for a cam.
