# How Long Has Your Company Got?

One of the problems I encountered when trying to estimate how much a business continuity or disaster recovery plan was worth, was estimating the remaining lifetime of a company. (The value of a disaster recovery plan depends on the length of time the company is expected to survive, as well as its profitability).

Nobody likes to answer the question “how long is my company likely to last?“, and attempting to answer the question can be perceived as a lack of trust in current management and thus be a career-limiting move. Yet we all know that companies do not last for ever.

I was therefore intrigued when I came across a possible method or predicting the life of a company in the excellent little book Mind Performance Hacks by Ron Hale-Evans. It draws on the article A Grim Reckoning by J. Richard Gott.

It uses just one simple assumption: you are estimating the lifetime of the company at a random point in its lifetime. Making just this assumption, we can actually estimate the company’s lifetime with (forthis example) a 60% confidence level.

Consider this timeline of the life of the company from beginning to end:

20% | 60% You are Somewhere Here |
20% |

If you are making the calculation at a random point, there is a 60% chance that you are in the middle 60%. That means you are somewhere between:

20% | Here | 20% |

and:

20% | Here | 20% |

Now if you know when the company started (t_{start)} and the current date (t_{now}), you can calculate the limits on how much longer the company has got. With the example 60% confidence limit, the limits conveniently simplify to the range:

_{now}– t

_{start})/4 .. (t

_{now}– t

_{start})*4

Using the same calculation (in March 2009) gives the expected death of Google (founded 1998) as between September 2011 and March 2053. For Microsoft (founded 1975) the range is June 2017 to March 2145.

Of course, these calculations are based on only one assumption – that we make the calculation about the company at a random point in its existence. The resulting range is therefore broad.

For companies, the assumption itself is questionable: we are unlikely to become aware of (and thus make the calculation) at a random point in the company’s existence. We also know far more about the statistical lifetime of companies so a more refined estimate is possible.

We can, however, make this calculation about galaxies, unfamiliar objects, and the lifetime of the human race. J. Richard Gott’s article has some good examples. It’s well worth a read.

Michael Z. Bell

March, 2009