The first example that comes to mind is US armed service veterans. Apparently, the medical care they received from the Veteran's Administration was really terrible, with apocryphal stories of veterans dying while waiting for appointments to see Drs. There were many bad actors in the VA medical service who would have been fired in any other organization but due to VA policy, kept their jobs. Through programs such as VA Choice and VA Accountability[0], the approval rating of the VA jumped to 91%, a record.
In the US, there are 17.4 million veterans, and as you indicated, they all have family and friends who are going to see these improvements in a positive light.
It makes me wonder how many other situations there are like this that we don't hear about.
On the other end, the president can pardon one person at a time. Perhaps that's the least number of people who can be positively affected by one act.
[0] https://www.va.gov/opa/choiceact/documents/choice-act-summar...
For one thing, guesstimating the number of affected people is pretty subjective. A policy can be important to me even though it does not nominally affect me directly. Secondly, this method doesn't even attempt to weigh policies by anything else, like the magnitude of benefits/downsides, budgetary concerns, civil rights implications, etc. For instance, in my view, it's fine for a political issue involving state-sanctioned persecution or bigotry of a very small minority to share political attention with a large public health crisis.