I am not of the field, bit did dig into the CBT hole (both reading and consulting a trained therapist), and I have the impression that the effects are overstated (and, at least for me, this was totally the wrong approach). The main issue I see with the studies I tried to look at was that they compared the answers to some psychometric scales between people following a CBT method and people following "talk therapy". My problem with it is that, while CBT is pretty formalized and can be "administered" in a not so terrible way even by an if-statement based chatbot [1], "talk therapy" is a very broad category, and the effects will be mostly dependent on not only the skill of the therapist, but also the fit between the therapist and the client. So yes, CBT will consistently give small results, but comparing it to consulting a random non-CBT therapist is as absurd as comparing the travel time by bus to the travel time by jumping in and out of random trains until you reach the destination.

[1] before I get attacked on this: I do not say that there are no good therapists with CBT training - but I am confident most of them would still be very good even without the training or method.