Data Scientists can be turned into expensive "Yes-men"

I wrote a few days ago that Data Scientists can get frustrated when the business only seems to accept analyses that back their gut feel and rejects others.

My post was fairly simple and offered one specific suggestion on what the Data Scientist (D) could do about this.

But what if the business doesn’t respond and this keeps happening?

The data scientists either leave or stay long enough to become confirmists! (Harvey Dent vibes anyone?)

The ones who stay back can combine with the business “experts” to do significant damage.

Cue this news on the recent SVB crisis:

BuccoCapital Guy@buccocapital

From WaPo this weekend: SVB’s internal risk model said the company would blow up if interest rates rose… So they tweaked the model to say the opposite Lol

3:16 PM · Apr 2, 2023

445 Reposts · 3.17K Likes

The cost to shareholders of this particular instance of just confirming business leaders conclusions was $44 Billion (from SVB’s peak valuation).

Every org has a structure. This problem could be hiding in multiple verticals and layers.

There is probably a tipping point in the proportion of data-blind business leaders that sinks a company. And the higher up they are, the more damage they can do.

SVB just fucked around and found out.