The stricter schedule in advertising vs work whenever probably has little to do with it. Advertising by its nature is more predictable, it doesn't have long term, world changing impact. The ultimate problem of selling is easy enough to crack, even if a campaign flops you try again. They're satisfied with their efficiency so no need to invest in neuroscience yet.
In software/science/math there is more variety, some problems are hard to crack. There is also more motivation to improve efficiency by orders of magnitude, to disrupt the market, like Google/Facebook do. Ad agencies don't develop new tech, only adapt to it, so they're vulnerable to obsolescence. At some point advances in neuroscience, behavioral economics and tech will outpace the rate at which most agencies can adapt.
There are agencies that spend crazy hours polishing up their work, or more often mismanaging. But the end result of such effort is more polished work, not necessarily effect, or a larger quantity of work thrown out on a whim. In technology the end result of crazy hours has more potential impact.
In software/science/math there is more variety, some problems are hard to crack. There is also more motivation to improve efficiency by orders of magnitude, to disrupt the market, like Google/Facebook do. Ad agencies don't develop new tech, only adapt to it, so they're vulnerable to obsolescence. At some point advances in neuroscience, behavioral economics and tech will outpace the rate at which most agencies can adapt.
There are agencies that spend crazy hours polishing up their work, or more often mismanaging. But the end result of such effort is more polished work, not necessarily effect, or a larger quantity of work thrown out on a whim. In technology the end result of crazy hours has more potential impact.