They are almost all proxy metrics of course. For engineering you can look at commits, PRs, incidents, deployments, etc and most of those metrics actually improved quite dramatically the first few months after everyone worked from home. Many of the offices are hot-desked high capacity environments and I cannot get anything done in them, so the initial reports were completely unsurprising.

I can't speak to what is being said now as I'm not really tracking any of it. I don't believe it's pure bullshit. It could be that people are just picking up bad habits. It could be stress and distractions from all the stuff that's swirling around now. It could be an exaggerated stance to signal to real estate in the affected markets that JPMC isn't packing up and getting out of town. Who knows. I will say that the technical infrastructure has been very good, due in part to long term investments in remote infrastructure for years now and it's pretty seamless switching from the office to remote.

I personally will quit before I go back full time to the chaos of high capacity office environments. I'm not built for it.

JPMC is a huge institution so it's hard to answer. Certainly there are many roles in trading, working with clients, support, etc. that will have pretty quantified metrics associated with them.

How does anyone take quantified metrics seriously if trying to compare pre-covid market bubble data to in-covid recession/market bubble. The environments are completely different.

One can do differential comparison for day of week.

AggregateOverEmployeesByAgeGroup(AggregateDatesBeforeCovid(Prod(employee_i | day of week)/AggregateDatesAfterCovid(Prod(employee_i | day of week before covid)))

Then see if it is flat or if there is changes (such as it dips on monday/friday)

Thats .. an adjustment. It really doesn't tell you if the work inputs allocated to young vs old groups shifted before and after covid. Also, I question the validity of the Prod( ) function as, like much other information work, the productivity yields on different timeframes and can be very difficult to correlate inputs vs outputs.

There were 3 things mentioned in the article. Younger vs older, Monday and Friday vs rest of week for younger, and creativity. You could theoretically measure younger vs older in the same job and also compare those two groups by M/F vs rest of work week. Compare those groups before and after COVID-19.

Yea, but the Prod function still entirely determines if it’s comparable.

Agreed, this claim feels totally baseless.