Be Very Very Afraid of the Public Cloud

Bain Capital won’t let us use Google Drive. It’s not “secure”.

This reminds me of a classic post that Dan Brinkmann wrote, which I paraphrase below.

We should all be very, very afraid of storing data in the public cloud. Data on-prem is much much safer:

  • Data on-prem is behind my firewall (which I run old code on), and cloud solutions don’t use firewalls
  • On-prem data is stored on infrastructure monitored by a solution that sends me lots of alerts, most of which I ignore
  • On-prem data isn’t monitored by an IDS, but that’s okay because it’s on the inside “safe” part of my network
    • And my users aren’t malicious (but my public cloud vendors are)
  • My company is in the data security business; it’s what our executives focus on, and in the DNA of everyone we hire
    • And the engineers at Amazon are dumb
  • Nothing is encrypted in the cloud
  • The government can’t access my data as long as it stays in my datacenter

Which reminds me. A CEO at a cloud monitoring company told me he has hundreds of customers, each who spend more than $1M / year on Amazon. And these customers are, on average, growing 6% month over month.

Yep, it’s definitely not secure, and it’s definitely not getting adoption by major companies.

Why software CEOs are building manufacturing companies

Steve Allan, head of SVB Analytics, presented at SVB’s security summit in Boston this week. He had a great slide, which I surreptitiously photographed:


What are we looking at? Steve has built an amazing dataset of private company income statements. This metric is S&M to R&D expense ratio. The cross over point is around $1M in revenue.

Think about what this means: when you reach $1M in revenue, you start spending more on sales & marketing than you do on engineering. You go from being a product company to a sales company. You shift the organization’s focus from developing software to selling software.

And here’s the kicker – you never go back. Successful companies just spend more and more on S&M, fueled by profitable revenue they can reinvest in the business. Successful software companies are manufacturing companies. They manufacture revenue.

For what it’s worth, Steve made a slightly different point. He says the big guys ($100M+ revenue) are the commercialization engines, and are buying up early technologies and pushing them through mature sales teams. But if you can manufacture revenue for yourself (i.e., stay independent), you’ll accrete much more equity value.