Software talent, cost curves, & capital efficiency

Software is typically a highly capital efficient business with a flat marginal cost curve. Each additional unit of software sold is relatively inexpensive after your product has been developed.

However, the marginal cost curve of a software company varies greatly with the level of expertise of your product talent. How cheaply and efficiently your code base scales to accommodate new users is a direct result of architectural decisions made by developers.

Your initial capital expenditure for product development varies greatly with the quality of your designers and product managers. Good designers and product managers drive down initial capex by identifying the most impactful user problems, testing ideas quickly and cheaply, prioritizing features, identifying adjacency, and reducing rework. They also drive organic adoption and increase usage by creating something worth raving about, which improves your capital intensity — how much it costs you to generate $1 of sales.

Sounds simple enough, right? Just hire a bunch of geniuses, architect a great solution, make it beautiful, engaging, and deeply salient to your customer’s most pressing problems and voila: Success awaits.

If only it were that simple. Assessing technical and design talent in order to figure out who will enable you to make the best of the unique cost structure and capital efficiency of the software business is a highly challenging problem.

Daniel Kahneman, the behavioral economist and Nobel laureate, describes an anecdote from his time in the Israeli army when his team was tasked with assessing the quality of potential military leaders. After a given batch of leaders was selected, high command would report back and share the performance of each leader. Over time, Kahneman and his team became increasingly confident about their ability to predict talent, but barely improved at actually gauging who would succeed and fail. I believe this is the state of affairs today in assessing technical talent.

We need to study our recruitment processes longitudinally, attach appropriate metrics to our people to gauge their performance effectively, and be more rigorous in selecting the right people.

Humans have a tendency to substitute an easier problem for a harder one if they are able to do so. Instead of figuring out whether or not someone will dramatically move the needle for our business, we might substitute a simpler question: Where did they go to school? Have they worked for a competitor? These questions are very reasonable to ask, but they ultimately belie the true question we need to answer. Can this person help us maximally realize our potential?

If you enjoyed these musings, follow me on Medium. And check out my website if you like.

Share your thoughts, feedback etc. in the comments.

🏴‍☠ayontech.io🏴‍☠️