Risk and Open Source adoption
The popular notion that early adopters are risk tolerant is wrong.
Nothing like a definitive statement to start an post, eh? Anyway, let me explain what I mean.
First, an analogy. Suppose you need $20000 for a life saving operation. Ok, maybe not in this country. You can save $1000 a month. If you keep the money in the bank, it is safe, but interest rates are low so after a year you have perhaps $13,000, and you die. Alternately you can invest the money in options, which have a chance to return enough growth to accumulate $20,000 in the 12 months. So, for you, what is risky: saving money in the bank, or betting it on options?
The point is, risk has to take needs into account. The 'safest' approach that cannot meet your needs is more risky than alternatives, however unlikely to succeed, that can meet them.
I got thinking about this the other day as I was having a conversation about open source adoption in the enterprise. I was asked whether those who were using open source were early adoptors and thus more risk tolerant. All in one breath, like it was part of the definition. Well yes, I responded, they're certainly early adopters in that they have adopted and other haven't. But as for risk ... well, no, they just have goals that weren't being met so in fact open source adoption was less risky.
In fact, open source was the way they managed their risk.
Here's a true data point that I heard direct from the mouth of an IT manager, all identifying characteristics removed: a certain large company is using a certain open source package in their external infrastructure, and has millions of dollars a day in revenue flowing through it. There are a number of commercial alternatives to this package, one of which they already have a site license for. Or they could use a productized version of the precise open source package available as a binary from a large established vendor. This company insists on using the open source version. They have 20 people in-house productizing, maintaining and supporting it.
They're spending $3M a year on free software.
Why? Because if it goes down, they can chain their staff to their desks until it is fixed. Their words, not mine. Because they can't buy an adequate support service guarantee from their existing vendor or from the vendor who has the productized version to meet this critical support need.
The reality of risk is that people trade risk for return. People don't bet in Las Vegas if the house collects all the money from the table at the end of every game. People didn't jump out of a certain collapsing office tower because this was a fun thing to do, they did it to have a chance to survive, no matter how unlikely, instead of certain death if they stayed where they were.
Instead of focusing on early adopters as risk tolerant, we should be considering what their needs are and therefore how they measure risk. Then the question of whether they're early adopters or an outlier can be answered through determination of whether others share the same needs, and perhaps just don't feel them as strongly yet, and are just waiting for the cost to drop or their pain to grow.