We analysts have one devil we can never fully defeat but we
can fight it with some success every day of our working lives: assumptions. I
had assumed Ms D would document her analysis in a particular way and look what
that had cost me. Lord Todd had made all kinds of assumptions when he first met
me… For that matter, even before he met me, I suspect he probably assumed I was
a man! Then again, you may have now assumed that (because I pointed out Lord
Todd’s possible assumption) that I am a woman...but what objective evidence do you have to indicate if I am
a man or a woman?
That’s the trouble with assumptions: they adopt excellent
camouflage and become part of your scenery without you even noticing them. You
may even happily assume you have identified and resolved all of them and all
the while they are tricking you in to thoughts and actions that are just a
waste of time and energy. They may cause you to waste so much that in the end
your endeavour – whatever it is – becomes futile…and they have won.
The first battle in the unending war against assumptions is
to see them. Lose this battle and the war is lost – you never even engage with
the enemy as they remain unseen. So, how to see them?
In summary:
1. Trust no-one.
2. Believe nothing.
In more detail:
1. Trust
no-one – not even yourself. Especially not yourself. Most people may well not
be trying to hinder you (most are doing just the opposite – being as helpful as
they can and they are often the most dangerous!). But what people do (you
included) is hide assumptions in the guises of “in my experience [insert any assumption here]” and “as
far as I am aware….”, “it seems to me…” and that killer “look, it’s obvious,
isn’t it, that….”. So, when someone starts coming out with these kind of
statements first off document the statement and immediately name and shame it
as an assumption. Now you can start
to plan to resolve the assumption (this is analyst speak for kill it).
2. Believe
nothing – especially that a method or approach will identify and deal with all
assumptions! So what exactly do I mean by “believe”? Well, it’s so obvious that
you probably don’t really need to be told: everyone knows what a belief is. The
previous statement is exactly what I mean by “belief”: it could have been
rephrased as “I have no objective evidence to support this but I choose to
believe that you know what belief is and don’t need to be told.” It is at its
most dangerous when you are being told to believe something by someone higher
up the pecking order, and they make an outlandish claim we don’t dare challenge
and then draw conclusions from it! For example “Given that magical things are
performed by small winged people then we know there are fairies – fact!” or
“Look, we need to follow industry best practice and that means classifying
people as pigs or chickens”. The best real life (probably apocryphal) story I
have heard of is one quoted by Matt Ridley (see his book “The Red Queen: Sex
and the Evolution of Human Nature). During an early meeting of the Royal
Academy of Science attended by the king there was an earnest discussion about
why a bowl of water weighs the same with a goldfish in as without. The debate
got heated until the king said “I doubt your premise” – the same as “I don’t believe you”. So a bowl of water and
goldfish was found, the experiment was done and guess what? Yes, shock, horror:
the weight of the bowl increased by one goldfish when one was added! End of
heated debate. Start of embarrassment for some….
The next battle, once you have seen assumptions, is how do
you kill them? Annoyingly, they cannot be completely killed but they can be
incapacitated to the point where they are mostly harmless.
In summary:
1. Prove everything to your satisfaction.
2. Proceed at risk.
In more detail:
1. Prove
everything – and the key here is “to your satisfaction”. It is not possible to
establish absolute proof of anything so don’t get lost in semantic shenanigans:
do what you need to do to prove whatever you need to prove to your
satisfaction. Personally, I would want to get the bowl of water and the
goldfish, but if you want to trust all those eminent members of the Royal
Academy of Science then that is – ultimately – up to you. It is your
professional judgement call.
2. Proceed
at risk. Frankly, this is all we can ever do. The trick is to minimise the risk
(can’t remove as we can’t ultimately prove anything 100%). So how do we do
that?
Step 1. What is the implication
of the assumption being false? Do not assume
that a false assumption is always catastrophic – it may have little or no material impact on
what you are doing (for example, does it actually matter if I am a man or a woman for the purposes of the
investigation?). You see how devious assumptions can be? They get in
everywhere, infect everything.
Step 2. If you need to make sure
the assumption is not false you need to prove it one way or the other – to your
satisfaction. So do that and kill the assumption by turning it in to a fact (as
far as possible).
Step 3. If you can’t prove (or disprove) it to
your satisfaction, assess the likelihood and impact of the assumption being
false, plan what you would do should that situation occur and proceed…at risk.
No comments:
Post a Comment