Manufacturing Serendipity

Posted on May 8, 2013

A phrase that I’m hearing crop up these days is “manufacturing serendipity”. At the moment I mainly hear this from Rand Fishkin and Will Critchlow, although knowing these two the whole world will be using it within the next two year.

Rand describes it in a post on his personal blog as:

This manufacturing serendipity business breaks down pretty much like this:

  1. Go to places that are not your office (conferences, events, meetups, trains, etc)
  2. Participate in things, learn things, and be generally game for new experiences
  3. Meet interesting people in the process
  4. Build relationships
  5. Be generally awesome by helping the people you’ve met and doing good deeds with no expectation of a return
  6. Repeat 1-5 hundreds of times

To me this process seems unquantifiable and high risk. Any method for success is difficult to evaluate because of survivor bias where only those who have used a method successfully but “manufacturing serendipity” seems particularly dodgy to me.

Another word for serendipity is “luck”. The process of manufacturing serendipity involves working very hard for a number of year until you get lucky. There is nothing really very new to this but it does raise the following questions:

  1. What activities can I undertake to maximise my luck for the same amount of work?
  2. How do I know if this is working? How can I be sure that in five years time I will be reaping the benefits of what I’m putting in now?
  3. Is there something else I can do that will give a better or faster payoff for the same (or less) work?

Similar questions can be asked about anything, but because of the long and uncertain payoff schedule for manufacturing serendipity these issues are particularly pertinent.

By creating thousands of opportunities, each with a tiny chance of success it is very hard (even with hindsight) to evaluate the effectiveness of actions.

To quantify this and evaluate the strategy against others we will have to estimate a per interaction chance of success within a time period. Lets suppose that each time you attempt to manufacture serendipity you have a 1% chance of receiving a payoff in the next five years (personally, I think this is high).

If you can manage 100 times a year (i.e. twice(ish) a week) then you can expect one of these to payoff within the next five years. Remember that the value of this payoff should be discounted because it is better to have the money now rather than later.

Now you can evaluate the expected payoff from manufacturing serendipity against alternatives.

Trust to luck? I’d rather think about the numbers first.