What do you mean "no evidence that making ads better targeted has been good"? You mean no evidence as in you A/B tested your on app and found no reasonable increase? Or you mean no evidence, as in you never bothered looking for evidence?
Judging from my personal anecdote with A/B testing targeted ads, there's a lot of evidence that it's good for both advertisers and users. Just try running tests and you'll see it for yourself.
Nice how you left off part of my sentence. Targeting ads is good for (some) advertisers (whether it's good for customers is debatable) but I said "good for the ad industry".
If a/b testing increases conversions all that demonstrates is you can get more revenue per ad dollar. Whether any "good" occurs is an entirely different matter. You might be selling penis enlargement pills.
Judging from my personal anecdote with A/B testing targeted ads, there's a lot of evidence that it's good for both advertisers and users. Just try running tests and you'll see it for yourself.