Imidacloprid is used far too frequently for insects that don’t exist are not a problem, or can be controlled by other methods. People chuck this stuff around for lawn grubs that have populations too low to cause damage or as preventatives and cure all's for trees and shrubs, without identifying a problem. Indeed I hear one local host of a radio show recommending this for just about every problem. That said there are places and reasons where this chemical is an appropriate choice.

Oil, my first choice is not always easy to apply properly, especially for the homeowner.
You need to get one hundred percent coverage to kill scale and have to treat several times. Timing is critical to get the crawlers (juvenile scale). Even with my spray rig I cannot treat all trees due to size or location. So taking the case of say the Cherry that this discussion originally came from. Yes by all means use the oil first. However, if a one-time treatment with Imidacloprid for a scale that will kill the tree is preformed then I proffer this justification. Assuming that I will knock off a few bees in the short term, but completely rid this tree of the problem, then the tree will live for many more years and supply countless bees with pollen long into the future.

In addition whether I use oil or Imidacloprid, I then will teach my customers how to avoid scale problems. Most scale and many other pests have numerous natural enemies that are frequently killed by unknowing homeowners. One Lacebug problem I had with a customer persisted even with several Oil treatments. It was not until I saw a note on the door that I figured out why. The note was from a Mosquito control company and it read that the yard was treated with Bifenthrin, a notorious predator killer. So I was killing pests and they were killing predators.
Other things that will help the survival of predators are such things as trying to keep something flowering all season, and having some fresh water available.

So before turning to Imidacloprid or any other chemical treatment please at least.
1. Identify the problem.
2. Decide whether the problem needs to be treated. Are the populations high enough to cause damage? Which and how many predators are present? Is the insect at the end of its damaging stage? Is the insect at a treatable life stage? Are other conditions such as weather about to control the pest?
3. See if this insect can be controlled using cultural or mechanical means? And think if you are encouraging it.
4. If you do need to treat, treat with the least hazardous option first.
5. If nothing else works and you have a plant that is worth saving, then you can turn to an effective, and labeled chemical option.
By the way just for the sake of discussion, Spinosad the bio-rational pesticide has a very similar mode of action to Inidacloprid.