Should you be building things that scale?
Week 7Reflection from IEOR 190E, UC Berkeley
Much like ebay or amazon, we would like to build a system simple enough to sustain itself. Related, we’ve been grappling with the issue of scalability for a few weeks now. Since we don’t charge customers 1000s of dollars, we can’t afford to hire only analysts. Everytime we pitch in class, we’ve been asked how we are going to scale? They’re worried that the data security vertical we’re focusing on won’t scale well to other verticals.
Personally, I don’t think you need to do things that scale in the very beginning. Even if Quintus can only offer data security recommendations first, we need to find the small group of customers for who our recommendations solves a serious problem. Then, you need to take extraordinary measures to make that small group of customers happy. Then and only then you keep finding more small groups of customers and start scaling subsequently.
In fact, one of the most common types of advice given Y Combinator is to do things that don’t scale.
For now, our answer to scalability is a weighted combination of crowdsourced research, direct vendor data and a small group of analysts. This is a simple, scalable and cost effective approach — the way to go!