If you are only interested in reading hyperbole-laced stories about the latest shiny things in #fintech innovation, then what follows is not for you. But, if you actually care about innovation that results in real impacts, then we invite you to keep reading.
From the following angle, something changed around mid-2012 at @GoldmanSachs. Arguably, the groundwork for such change was laid prior to that time, but the impact of that groundwork (at least to observers like us) didn’t become noticeable until the end of Q2-2012 when a metric that we are starting to track much more closely began to break out.
Here’s the setup: In an article published on February 7, 2016 by MIT Technology Review, “As Goldman Embraces Automation, Even the Masters of the Universe Are Threatened,” a key statement made by Marty Chavez, the company’s soon-to-be chief financial officer and former chief information officer, caught our eye: “Goldman has already mapped 146 distinct steps taken in any initial public offering of stock, and many are ‘begging to be automated.'”
Our between-the-lines translation: Goldman is meticulously mapping each task in each workflow and then figuring out how to automate those tasks one by one. Over time, successful implementation of this strategy begins to impact workflow productivity (and the operational analytics emitted by workflows). In our opinion, the chart below is a picture of such a strategy bearing fruit.
What we have here is 41 quarters, beginning in Q4 2006, of reported “assets under supervision” (AuS) – an average of nearly 90% of which are represented by assets under management (AuM) – divided by total headcount over the same period. Clearly, what we want to point out here is the new trajectory established as of Q2-2012.
Now, playing devil’s advocate, one could argue (if they knew GS well enough) that total headcount is an overly crude instrument to be used to evaluate the productivity of the Investment Management division alone. Fair point. It’s true, Goldman doesn’t report any headcount figures other than totals. For this reason, we have eliminated the y-axis in the previous chart because, while the actual AuS per employee figures may not be indicative of the real level of assets that can be managed per employee (given their mix of investment strategies), at least the trajectory of this metric is valid and illuminating.
Nevertheless, we decided to go a step further to validate the concept by creating an index of the change in both variables, AuS per employee and total headcount. A summary of the chart below illustrates that over the 40-month period, total GS headcount has increased by 15.7% while AuS per employee has increased by 63.5%.
While there are still several reasons why this illustration could represent some kind of false positive – and our modeling of new “AuX per employee” analytics is still admittedly in its early stages – we also know that new intelligence like this rarely comes wrapped in blinking neon lights. Hence, we believe this is a notable finding. We will share more as our own innovations bear fruit…