Failing fast has become a common trope by now. It makes a lot of sense in some contexts. Fear of failing causes many problems, the biggest of which is being afraid to try something new. Even if a new thing only has a 1% chance of teaching you something useful, not trying gives you a 0% chance of learning something useful. Fear of failing prevents long term success.

Overcoming this fear isn’t easy, but it can be done. Failing fast can help someone grow accustomed to failure. They start to realize that failure isn’t the end of the world and a lot can be gained by taking chances. They also minimize the cost of taking chances through the “fast” part of failing fast.

However, the key part of failing is being able to learn from that failure to increase the chances of success later. In this respect, failing fast can be taken too far. Being able to tell yourself that you’re “failing fast” can create a false sense of security. Instead of analyzing failures thoroughly, it becomes easier to just shrug off a failure without putting a serious attempt in learning from it.

A perfect example is software development processes. No process is perfect and each one needs to evolve based on the team. And yet, it could be very easy to overreact to the first problem that can come up. “Look at how we let this one bug through! This process sucks. We need to start from scratch.”

The problem in this scenario is that it skips an in depth analysis of the failure. Maybe the right choice is to start from scratch. Or maybe there’s a tweak to the existing process that could work. Gut reactions are not going to make the right decision here. The failure needs to be discussed and dissected. Otherwise the failure gains you nothing. Failing fast becomes failing for the sake of failing instead of failing for future success.

There’s also the issue that for the most part, the number of possible things you can try is near infinite. Using an analogy: let’s say I’m writing a blog post and I’m trying to find the right way to word something. There are hundreds of thousands to millions of words. When phrases/sentences with many words are in consideration, that number grows very large very quickly. Not every permutation is useful. Experience with language provides a rough filter from the vast majority of those permutations. Analysis of previous writings provides a finer filter for the remainder.

But without that analysis, I would just be trying random words blindly. That’s not a path to success. That’s simply failing.