🎬 Scary Good

One of my favorite ideas from investors like Ray Dalio and Cliff Asness is the idea of out-of-sample tests. The idea is that if something works in more than one place or more than one time or under more than one conditions — it’s an idea with more skill than luck. Asness put it this way, "If you give me enough data I will find you something that has worked in the past - but it might be total gibberish."

The trouble with life (and cause for out-of-sample test) is random correlations along with our tendency (read: evolutionary need) to see causes and effects.

One out of sample area is Hollywood though for this week we’ll go to the less ritzy side. We’ll look at Jason Blum and Blumhouse productions, the team behind movies like Get Out, The Purge, and Paranormal Activity.

Blum’s a smart cookie, approaching Hollywood orthogonally but it’s his meta comments on his model that marvel me. In interviews he frequently notes that his way isn’t the only wya. Imagine for example this choice:

Company A is a major movie producer. Some movies top the box office for the year and there are a number of successful franchises the company can (continue) to update. Along with box-office success are merchandise and other tie-ins.

Company B is a smaller movie producer. Not every film makes it to the box office, some go straight to streaming. The company operates with a smaller budget and has a few successes, no failures, and nothing on the scope of Company A.

The first company is Disney, the second is Blumhouse Productions. Jason Blum emphasizes that importance that both do what’s best for them. If Disney acted like Blumhouse they’d shrink their competitive moat. If Blum acted like Disney he’d bankrupt his company.

This dichotomy is the exact reason Pixar had to be bought. So, while film can provide an out-of-sample test on strategy, it’s not necessarily the best strategy for your sample.