Institutional Knowledge: A Metric about the Google
July 6, 2021
I read an unusual “I left Google” essay called “Leaving Google.” I am not sure I understand how an apparently valued employee would find bureaucratic processes a reason for leaving what seems to be an okay job. You can read the write up by a Xoogler who worked at the mom-and-pop online ad company for 17 years.
I noted one factoid which struck me as quite interesting. Here it is:
At the time I left, out of ~150k employees, only ~300 had worked there longer than me.
That means that the institutional “knowledge” of Google, its thousands of technical components, its hundreds of thousands contracts, its millions of inter- and intra-process dependencies from the days of Backrub to the weirdness of solving death to the incredibly brilliant but Floc’ed solution to user tracking resides in exactly 0.1935 percent of the staff.
Perhaps some of Google’s more interesting behaviors, products, pronouncements, and personnel decisions have drifted from what the company was first engineered to deliver?
Questions:
- When a component buried deep in code created in 1999 goes wrong, who knows how to fix it?
- What if the issue cannot be fixed? What then?
- How much of modern Google is code wrappers slapped over something that mostly works?
- Are disconnects between engineering and marketing created by this loss of institutional knowledge?
I suppose one can use a Web search engine like Bing, Swisscows, or Yandex to seek answers. Google search is not particularly useful for this type of query.
Stephen E Arnold, July 6, 2021
Comments
-
- Subscribe to Beyond Search
Feature archive
News archive
-
