segunda-feira, março 02, 2020

Analítica versus experimentação

Via Nassim Taleb cheguei a esta sequência:

Entretanto, tinha lido "Two Words That Kill Innovation":
"Over the past 50 years, management practices have become ever more scientific and quantitative. Managing by the numbers, using business analytics and leveraging Big Data are all considered to be unalloyed goods, indicative of enlightened management. Without question, data and analytics have their roles and their benefits. But they have a really important dark side too, and when managers don’t see that dark side, they accidentally kill innovation.
.
The implicit logic behind the scientific management doctrine is that you must prove — analytically, and in advance — that a decision is correct before making it. To be clear, it is not the explicit doctrine — few managers think this themselves, but they’re swayed by their training to be scientifically analytical. This works productively for most of their everyday decisions.
...
But when genuine innovation is required, there’s a problem.
...
it is not possible to prove analytically that a new idea is a good one in advance. Why? It’s pretty simple when you think about it. There is no data about how a genuinely new idea will interact with the world in advance of said new idea actually interacting with the world. Therefore there is no way to prove it will work in advance.
.
This creates a real problem for managers who believe that their job in life is to make sure that a decision should be made only when there is analytical proof that it is the right decision.
...
To change this dynamic, managers need to distinguish between when they are honing and refining an existing system and when they are attempting to create something genuinely new. In the former situations, it is totally fine to come in with analytical guns blazing. In the latter, they need to put away the guns and take an entirely different approach. Here, they need to borrow from the design thinking toolbox by engaging in prototyping. Try innovative ideas, but do so in small ways without a lot of up front investment. Generate data through experimentation rather than assuming that there is pre-existing data to be harvested. Iterative experimentation will migrate the solution to an ever more compelling state — and spin off new data along the way."

Trechos retirados de "Two Words That Kill Innovation".

Sem comentários: