90% chance of failure? Try 10 times
The evolutionary approach applies to work too.
The way to beat low probability is by either trusting your intuition and going narrow (2–3 attempts or choices) or being open to anything and going wide (10+).
Often, we can attempt something many many many times with very low cost. Ads on social media, emails, copy changes, algorithm tweaks,…
If we attempt many times, our chances of success go up and we can create a system based on the approaches that worked. That system may be a business, or a non-profit, or a government. The pay-off of a working solution is often so asymetrical that tens, or even hundreds of attempts are warranted.
If the decision is low-cost (relative to your resources), the outcome unpredictable, and the pay-off asymetrical, try as many times as you need to.