I recently applied to an analyst role at a VC firm and one of the application questions was; What questions do you spend the most time thinking about?*
It’s a fascinating question, and it was a an illuminating experience to sit with it. I think a lot, almost bordering on overthinking but yet, I had never consciously paid attention to what I think about the most. After a rendezvous with my thought patterns and an audit of the way I engage with people, I realized that I spend a lot of time thinking about diminishing returns.
How can I slow down an object from approaching diminishing returns quickly? When does putting in additional resources start yielding progressively smaller or less impactful results?
I have worked in data analytics for up to five years now and have seen so many projects fall into this pattern where you put in so much to achieve very little result. Let’s take Machine Learning projects for example, say linear regression, there is consistently a huge focus on attaining model accuracy alongside other parameters for measuring if your model is giving a reliable result. For the non techies, Linear Regression is simply a computer’s way of saying – based on what happened before, here’s my best guess at what will happen next
What happens in these scenarios is that there is a tendency to continue to feed more data into the system or tune the model and apply other techniques, if it does not meet certain expectations. Essentially brute forcing these models into submission. Having walked into these rabbit holes quite a bit in my experience, has created in me a warning system for everything. When do we stop and pivot? Do we continue at this pace? What can break and how do we prevent it from breaking? What if we get stuck, how do we proceed with the discomfort?
The decision to continue to put in more resources or time or effort, is often caused by a fear of not being innovative. It is easy to think that if you do not do anything then the product or model or whatever the object in question is, will not produce at its optimal. The best form of innovation sometimes is consistent effort in the same direction. Of course this does not apply in every situation; sometimes it’d be better to stop and pivot, other times applying minimal effort gives the best results (remember the 80/20 principle).
Whichever scenario you get to apply will be dependent on how early you start thinking about diminishing returns aka sustainability. But this time it is not just financial or economic sustainability, but the sustainability of effectiveness, and the wonder that your innovation brings.