Based on some personal experience from perhaps 7 or 8 years ago, it can work for driving sales.
However it required quite a lot of effort.
It wasn't just throw some money into the ads and wait for the returns - you needed to basically spend a full-time job pulling levers and prodding buttons to find out where the "sweet spot" was and then make sure you keep up with the moving target over the following days weeks and months. If you were lucky and your business was selling the "right" things (i.e. large margin items) you could do quite well. E.g. it might cost you $15-20 in ads to make one sale - that is net-positive if the margin on your items is large. So some fashion, furniture, holidays etc a $20+ Cost-per-acquisition is a bargain as you just paid $20 to make a $150 profit. If you are competing for common ad space/keywords and you have a low-margin business then you're probably not going to do so well since you might be paying $20 to sell something that costs $9.99 or whatever.
Of course, there is then the question of would you have made that $150 profit without spending the $20? Or did you actually make more than one sale from those $20 worth of ads (e.g. in-store sales etc in addition to purchases made online). I am not involved in online advertising now, but at the time that was the trillion dollar question that everyone was trying to answer... no idea if it is solved now.