There is no ‘start from scratch’

We need to start over and do it better. Can’t we just throw it out and start over? This system is broken, we need to start over.

I’ve heard these statements a million times in a million contexts, so today I’m going to stick to the tech and product environment for my examples. The reality is, no. No there is no blank slate, but you can make change, even radical change.

There’s a concept called path dependence which has been examined in research across sociology, history, economics, and politics, and while researchers in different fields define it a bit differently, the basic concept is the same. Path dependence means what happened in the past shapes what CAN or DOES happen next. It’s the idea that “history matters”.

In part because of the different fields of research involved, there’s actually a spectrum of path dependence. On one end, it says: What happened in the past determines where we are now, and it locks in some things for the foreseeable future. On the other end, it says: What happened in the past does impact where we are now, but at any moment we can make a different choice that doesn’t at all depend on what happened before it. And then in the massive grey area of the middle it says: What happened in the past determines where we are now, but the future is changeable if we start from where we are.

When I look at this spectrum, I actually see a really common ideological workplace divide. On one end the designers or engineers who believe that the past doesn’t have to impact the future; the legacy systems and the ‘how we’ve always done it’ aren’t working, we can just start over. On the other end the product managers or engineers who believe that the future is locked-in because of the past; we have to work within the legacy system and there are important reasons why we do things this way.

I believe in the middle ground for a lot of reasons, but I’ll share two of them. 1. suboptimal decisions. They happen, often you can’t undo them, and they can absolutely limit what’s realistically possible. Ignoring them, unfortunately, doesn’t make them go away. And 2. a concept called lock-in. This is when a choice or pattern becomes self-reinforcing. For example, we purchased a 3 year contract for this software so our team is going to use it, or (as a less literal example), all our competitors built an AI summary feature so we have to as well regardless of whether it’s useful.

The takeaway here is this: path dependence exists, and I think accepting it gives us more power to shape the future. There are some imperfect decisions that we cannot undo and some lock-in that holds us in place, and we have to see that in order to make change that can actually work. And also, we have to see that path dependence doesn’t mean path inevitability. Thanks for listening.

Sources: https://www.jstor.org/stable/765077 (pdf available via Google Scholar), https://scholar.google.com/citations?view_op=view_citation&hl=en&user=2XymNBEAAAAJ&citation_for_view=2XymNBEAAAAJ:4DMP91E08xMC, and https://www.cambridge.org/core/journals/american-political-science-review/article/abs/increasing-returns-path-dependence-and-the-study-ofpolitics/AC2137B913363E33D97FC5CEC17CC75D (pdf available via Google Scholar)