Building on the strong points you and others raised (especially ethai), I think one factor working against data-driven marketing in EA orgs is how cautious EAs tend to be when deciding what to do. There’s often a lot of overthinking and a desire to fully resolve details before launching anything.
This trait is very valuable in many contexts, but it can work against growth, where success depends on fast learning and experimentation.
Saying “we tried growth for a year and it didn’t work” doesn’t mean much if “trying growth” mostly meant spending three months debating Facebook ad copy, running a campaign for a few weeks, and then spending another three months overthinking the results rather than rapidly testing more variations (I’m using ads just as a simple example, but ofc growth marketing goes way beyond testing Ads).
I agree. Many EA norms that are great for reasoning, caution and overthinking before acting, work against growth, which learns through fast, imperfect experiments.
So “we tried growth and it didn’t work” often reflects a very slow learning rate, not strong evidence that the channel itself is low-ROI.
Building on the strong points you and others raised (especially ethai), I think one factor working against data-driven marketing in EA orgs is how cautious EAs tend to be when deciding what to do. There’s often a lot of overthinking and a desire to fully resolve details before launching anything.
This trait is very valuable in many contexts, but it can work against growth, where success depends on fast learning and experimentation.
Saying “we tried growth for a year and it didn’t work” doesn’t mean much if “trying growth” mostly meant spending three months debating Facebook ad copy, running a campaign for a few weeks, and then spending another three months overthinking the results rather than rapidly testing more variations (I’m using ads just as a simple example, but ofc growth marketing goes way beyond testing Ads).
I agree. Many EA norms that are great for reasoning, caution and overthinking before acting, work against growth, which learns through fast, imperfect experiments.
So “we tried growth and it didn’t work” often reflects a very slow learning rate, not strong evidence that the channel itself is low-ROI.