Risks exist in every business. We know this, therefore we manage them. Every company has made risk assessments and has consequently introduced risk reducing measures. After these measures, we accept the fact that there is a residual risk which is deemed acceptable, or “ALARP”. In a high hazard industry, a company may have hundreds of identified risks, and assumingly, all of them have been managed well by applying risk reducing measures. Every single one of them is thus “acceptably low”.
The fact is, there are remaining risks. Each one of them is acceptable on an individual level. But have we ever looked at the chance of the cumulative risk? What is the chance that a single one of those residual risks ever becomes manifest?
As a risk is defined as chance x effect, we can either take measures to limit the effect/damage that a manifest hazard does or we can take measures that limit the chance of an occurrence. Especially in the latter, there is a latent problem. If all of our managed risks have a very small chance of occurrence they are deemed acceptable, even if the possible effect is still high. The cumulative chance that a random one of that list ever becomes manifest over time, regardless which one, is a lot bigger than the individual chances. The longer the list of our managed risks, the higher the chance that eventually a random one out of that list finds its way through the Swiss cheese model.
Fig1. Hitting Green 0
It’s comparable to playing roulette. Black is winning (operations as normal), red is also OK-ish (near miss, or incident without consequence) but the single green “0” pocket is big bad news. In roulette, the chance of the ball ending on the green 0 is 1:37 (2,7%). For this comparison, let’s assume that 1:37 is an acceptably low risk. However, if our list of managed risks grows longer, we are actually adding roulette tables. If we play on two tables our chance that the ball chooses a green pocket on at least one table is considerably bigger. If we have an entire casino full of roulette tables and we’re going to play all night long, what is the chance that one (or more?) of those tables shows 0 somewhere during the evening?
Let’s assume we have a casino of 20 roulette tables, and we’re going to play 10 bets on every table, the chance that we’ll see one or more green 0’s during the night is 99,6%
Fig.2 The red and orange areas are the obvious ones. Usually well managed. The green ones at the bottom, especially the further right you go, that’s where the cumulative nastiness is hidden. Green suggests safe, but that is only true for individual cases. 50 times unlikely or rare is more likely than you may desire…
I understand that in high risk industries all risks have been managed to far below the 1:37 mark, and that the vast majority of the risks are not nearly catastrophic, so the above calculation is not very realistic. But in most companies the list of risks is far longer than our 20 roulette tables. And we perform our business far more frequent than our 10 bets…
Update from just this last week:
• A helicopter crash occurred offshore Brazil.
• A collision between a supplier and a FPSO happened offshore Brazil, in the same field.
• A riser was dropped creating an oil stain in the sea, again in the same field.
• A severe crane incident happened offshore Norway, where a 17-tonne load + a part of the crane crashed on the deck.
• A fire broke out offshore Congo, injuring 4 and killing 1 person.
• An offshore supply boat runs aground offshore Louisiana
The balls seem to actually hit “0” every now and then. I bet all these occurrences have been estimated as “extremely rare” in a matrix…
The bottom line is, that if we continue to operate our business long enough with a long list of risks, even if they are very well-managed, eventually something will hit the fan. No company is invulnerable no matter what your safety record says. If you play long enough, and especially if you play multiple games simultaneously, eventually a 0 will knock on your door. If this is a given certainty, then why is your dedicated crisis room  full of cardboard boxes with products that didn’t make it through the Q control and why has your crisis plan  been getting dusty somewhere on a shelf for years?
Rien ne va plus…
 ALARP: As Low As Reasonably Practicable
 Do you have one?
 You do have one, don’t you!?