REPLY Definition of root cause (SD6788)

[SDMAIL Kim Warren]

Posted by  "Kim Warren" <Kim at strategydynamics.com>

Sorry jack - My reply was not as clear as it could have been ... and
maybe I'm not as clear about 'root-cause analysis [RCA]' as I should be
either ...

I am assuming we are trying to understand the deep, underlying reasons
why something is not happening as we wish it to - a problem is
persisting or some performance we want is not improving as we would
like. I am also assuming that since we are an SD interest-group, the
indicator of concern is changing over time, rather than just being a
static question. 

The causal tracing I tried to summarise before follows as rigorous an
arithmetical logic as possible, e.g. "profit = revenue minus costs",
"delivery performance = on-time-deliveries/month divided by
total-deliveries/month" and so on. After repeating this question of what
causes each of these next elements, and continuing a few times, you seem
always to get back to one or more stocks, one or more decisions, and one
or more exogenous factors. For an airline example I've worked on:
- "revenue/year = passenger-journeys/year * fare-per-journey" [fare
being a decision]
- "passenger-journeys/year = customers * journeys/year-per-customer"
[customers being the stock of people who routinely use the airline, and
dominate its sales]
- ... and our "journeys/year per customer" depends on our fare [a
decision] vs. rivals' fares [an exogenous factor]

A *static* analysis of why our sales revenue is less than we want would
therefore identify these 3 items as the root causes. But that's not very
insightful or useful, because we want a *dynamic* answer rather than a
static one - i.e. why are revenues not growing as fast as we want? 
Only a part of this is likely to be explained by changes to
journeys/year-per-customer. A larger part will reflect changes to the
stock of customers, and this can only be changed by flow-rates - if no
customers are won or lost, then customer numbers remain the same and,
ceteris paribus, revenue will not change. This means that to continue
the quest for root-causes we have to explain what causes the rate of
customer wins and losses. ... and if we keep asking what-causes-what
behind those flow-rates, it seems we get back to the same three
categories of items, e.g.:
- "customers won per year = some function of the potential customers on
the routes we serve" - [a current stock level]
- "customers lost per year = some function of load factor, reflecting
customers and aircraft " - [another stock-level]
- ... and some function of our price [a decision] and or rivals' prices
[an exogenous factor].
If this is still not clear, here's a book chapter that explains in more
detail - www.strategydynamics.com/c4. It doesn't constitute any grand
'theory', nor is it any different than how SD has always depicted
situations. But the principles do seem to hit the main criteria needed
to solve dynamic problems, being as far as I can see general, useful and
true. 
As I say, RCA is not my field, but what I see on Wiki seems *very* close
to what we do with SD, and implies that root-causes will ultimately be
found in the three classes of factor above - stock-levels, decisions, an
exogenous factors. I suppose the last and critical extension is to
identify what it is about the decision[s] - i.e. the policy-rule that
guides those decisions - that lead to problems or suboptimal
performance. So perhaps the true root-cause is always a dysfunctional
policy? [unless there are resource-shortages or exogenous factors that
just make good outcomes impossible, no matter how well we make
decisions!]

Kim
Posted by  "Kim Warren" <Kim at strategydynamics.com>
posting date  Wed, 5 Mar 2008 17:50:45 -0000