Pinterest – a case study in capital efficiency

Pinterest is an awe inspiring example of how much can be achieved with minimal investment in tech and people these days. As most of you will know, the enablers are cloud computing and open source software. What is interesting in this case is the power of those enablers.

I just read the following numbers on the High Scalability blog. Pinterest now supports 18m users with growth running at 50% per month with the following:

  • Amazon costs are around $39k for S3 and $30k for EC2 (I assume per month)
  • Traffic costs them $52 an hour during peak times, down to $15 during the night
  • They have 31 employees, up from 12 in December.

The geeks amongst you will be interested in the following details:

  • Pinterest has 80 million objects stored in S3 with 410 terabytes of data
  • They have 275 EC2 instances
  • 70 master databases
  • Written in Python and Django

This is an amazing story, but it isn’t unique. Instagram had similar efficiencies, and we will see more and more startups aspire to and beat these benchmarks and try to emulate Instagram’s $1bn exit or Pinterest’s $1.5bn valuation.

  • smcllns


    It’s certainly a good example of how traffic is affordable to scale. Certainly that engineering is by no means trivial, and it’s an impressive feat to manage that growth.

    But is there no concern in the back of your mind that this influences the funding environment to focus on growth metrics, rather than future long-term value? I really mean this to be a constructive issue, not a wet-blanket whine:Investors appear to rally for investments in the uber-growth co’s with Pinterest like growth curves. Now in order to show that sort of growth, you basically need to be doing something with sharing photos. In reality, it’s more likely to involve something at the heart of consumerism + social media, as opposed to difficult technological advances. Steve Blank argues this point well here:

    I agree with Blank, but I’d love to hear your opinion — whether you have that nagging voice in the back of your mind somewhere — or how my perspective needs tweaking.

  • Companies like Pinterest, Twitter and Instagram that achieve massive user traction (read tens of millions and growing fast) are likely to be able to find revenue streams and profits and hence become valuable companies. Because they get their quickly they create value at an unprecedented pace, and that is attractive to investors. Nothing wrong with that. Conversely, if you are in a space with low monetisation per user unless you get to those kind of user numbers you won’t be valuable enough to be interesting to VCs. Hence for these companies the growth metric is appropriate.
    There aren’t many companies like this though.

    Fortunately, there are tonnes of businesses that make money with less exciting, but less risky models. The successful ones will sell in the $100-500m range, with the odd one breaking out higher. Our recent investments in companies like Sports Pursuit, StrikeAd, Conversocial and MoviePilot fall into this category, and this is where most of our bets are placed.
    Does that answer your question?

  • smcllns

    It does, thanks.

  • Encouraging and useful insight. For me this is most helpful. Thanks, Roger

  • Pingback: Kernel column: Falling share prices at leading tech companies don’t mean the end of the world « « The Equity KickerThe Equity Kicker()

  • Pingback: The changing dynamics of consumer internet investing « The Equity KickerThe Equity Kicker()