I wanted to play around with Claude Cowork, so I got it to research every Chief of Staff across Europe's 250 fastest-growing startups (using the Sifted 250 list).
Why? I’m interested in understanding the nature of this role in companies, and this list seemed like a good place to start.
11.6% of the Sifted 250 have a Chief of Staff (CoS). These are more established companies, so that tracks. As a startup scales, parts of the CoS role start needing full-time attention - and therefore full-time hires.
UK companies are more likely to have a CoS (15.3%) than continental Europe (9.7%). This likely reflects the US influence on the UK market. The role is essentially absent from southern Europe, the Nordics outside Sweden, and eastern Europe.
Fintechs are least likely to have a Chief of Staff. This is the largest sector in the Sifted 250 (78 companies) but has the lowest adoption at just 5%. Climate tech leads at 19%, followed by B2B SaaS at 15%. Interestingly, Monzo Bank is the company in this list with the most Chiefs of Staff - it has a distributed model where each CoS reports into a different exec. Only two other companies have more than one CoS.
On average, a Chief of Staff stays in the role for 22.5 months. a16z suggests its 12-18 months, McKinsey leans towards 18-24. The data here supports the longer end.
Internal promotion to VP or C-suite
VC and investing (operator-in-residence roles)
Advisory and CoS coaching
Another CoS role at a different company
A Chief of Staff is more likely to be hired externally (57%) than internally (38%). External candidates typically come from strategy/consulting (16%), operations (14%), and serial CoS pathways (14%).
Internally, the most common routes are EA → CoS, strategy manager → CoS, and ops lead → CoS.
Here’s a quick note on the methodology I followed.
Gave Claude the Sifted 250 list
Got Cowork to run LinkedIn searches on all of the companies and find current or previous CoS at these companies, and tracked those individuals’ career paths
Built a .csv with all the data
Ran an analysis of the .csv to pull numbers and find interesting data points
Create infographics (Cowork created them in HTML, then I edited them in Canva)
Challenges
It took a bit of time to get to the best and fastest way to run this search; it started out slowly with the first 10 companies while Claude worked out a repeatable process. And then it moved to batching Google searches across multiple agents, then only visiting LinkedIn for confirmed hits
Generally hitting a slowdown with the Claude in Chrome integration, and having to prompt Claude to check that it was running the search in the optimal way
I reran the analysis through Claude rather than Cowork, and found some errors in the initial analysis
I'd love to know if these findings match what people are seeing. Is the CoS role growing, shrinking, or just being called something different?