Data. YC has collected tons of data points since 2005. From more than 100k applications, roughly 2k batch startups (11 unicorns, several dozend +100MM startups), 3,5k founders, successful or not. Probably everything from founder CVs to week-per-week growth rates to the amount of office hours during a batch to VC deals/termsheet offerings after demo day to following series A, B, C rounds. Probably every single ratio you could imagine that might be somehow relevant. YC employs 50+ ppl, most of them engineers. They are basically a data-driven software business, with ML pre-selecting new applications.
So ... rather generic-sounding advice like 'do things that don't scale' or 'distribution is king' or 'make sth people love' is maybe as good and objective as it gets. These are probably the necessary heuristics to use while not being sufficient (luck, timing, talent...). From a merit-based point of view I'd say that SA is one of the top-10 global startup gurus. Because of the above.
So ... rather generic-sounding advice like 'do things that don't scale' or 'distribution is king' or 'make sth people love' is maybe as good and objective as it gets. These are probably the necessary heuristics to use while not being sufficient (luck, timing, talent...). From a merit-based point of view I'd say that SA is one of the top-10 global startup gurus. Because of the above.