This essay by Paul Graham is one of the best on startups. In short, he makes the point that startups (and investors) covet growth. Without it, nothing else really matters.

I’m not the only one who thinks this was essential reading. With that in mind, more and more startups include their growth metrics either when pitching or updating investors (e.g., “X% growth week over week”; “Y% growth month over month”). Some of the stats are easy to interpret - you don’t need to do the math to understand 50% monthly growth over a sustained period is pretty good. But for other statements of growth, it’s less obvious to me. That is, I can’t immediately intuit if 10% monthly growth is ‘good’ and what does that mean. And obviously this depends on industry - 10% monthly growth for a mobile app may be not so good whereas for an enterprise software company, it might be healthy and borderline explosive.

It’s disgraceful that I can’t synthesize these numbers in my head. I was a Math major and even went to graduate school (for 3 weeks) in Applied Math.

Because of my now limited math abilities, I use the “Rule of 72” to think about growth for back-of-the-envelope purposes. I won’t go into the math here, which is pretty basic for a lot of folks. But the idea is that if you are growing X% over one period and you sustain that growth, then you will double your size in 72/X time periods. As an example, let’s say a startup says they’re growing 12% month over month in users. Using the Rule of 72, they will double in users in 6 months (i.e., 72/12). And quadruple users in 12 months. And so on. This works for any time period. So if an early startup says they’re growing at 12% weekly, then they should be doubling their growth in 6 weeks. And this obviously works for any metric - revenue, users, traffic, etc.

Like all back-of-the envelope calculations, this is a rough approximation and has many caveats. Here are a few:

- This is just a rough estimate - that is, in my examples above I think the real answer is 6.12 months.
- This estimate becomes less accurate as the growth rate gets higher. For example, 72% growth in one month doesn’t mean you will double your growth in one month. You’d need to grow 100%.
- This also assumes that the growth rate will be constant over time. This is the biggest caveat since it doesn’t reflect probably about 99% of real-world examples.