Andrew Moreno asks, Replying to LO4277 --
On 16 Dec 1995, Rol Fessenden wrote:
> As time goes on, we mix in successes and failures, keeping the whole mix
> in front of everyone. This creates an environment in which management can
> easily see that the successes more than pay for the failures, and
> therefore, they condone ongoing experimentation.
Andrew asks, "Is this implicit or do you explicitly state this?"
Rol responds,
No hidden agendas. We state everything explicitly. We tell people we
have a number of experiments, we describe them and their probability of
'success', and we request funding. We describe the benefits as well.
After the experiments are substantially complete, we 'account' for the
money spent, attempting to quantify -- often not successfully -- the
return. Qualitatively a positive return is generally agreed by everyone.
My original quote
> Some people may view this as manipulation. It is not. It is absolutely
> essential in a competitive environment that we achieve enough ongoing
> success to pay for the failures and still make a little money. The
> process described above forces a discipline on the department to meet a
> certain positive payback ratio. Generally, the 'failures' provide learning
> that eventually is converted into a success further down the road.
> Keeping the process visible to all is merely responsible behavior.
Andrew says,
"This can get very tricky. The department has given you a certain level of
agreement to manage their interests. What you are doing could be
misconstrued as operating in a way that the department hasn't agreed to.
"The goal? To get their agreement with explicit understanding on their
part. Maybe the way to do this is through a technique used in sales
letters, the "damaging admission." State that there will be failures but
then really load up on the advantages - failures provide learning being
one of them."
My response:
Actually, it's very untricky. No hidden agendas. What I say is what I
mean. If I can't convince people to give me some slack, then I guess I
don't get to experiment. What I do is correctly intentioned. In the
execution, sometimes things change, we learn, and we end up taking a
different direction than planned. Next time we go back to management for
an update, we tell them what happened and why we changed. It's all very
clear.
We do not particularly focus attention on possible failures in advance,
but we account for them afterwards in a straightforward and factual
fashion. Mostly, that is not a problem. Sometimes failure can be a
problem, but that's life.
I know what you mean when you say it can be tricky. Generally, it is not
tricky here.
There are other situations where failure is more problematic, for example
in operational areas. If the goods are not in-stock, that is a major
problem, and I think it should be. If an experiment or a project has
_fundamentally_ changed direction without a thorough discussion, that is a
problem. Again, I think it should be.
I think the key to dealing with failed experiments successfully -- at
least here -- is to not dwell overly on it, but at the same time, not to
minimize it. "It failed, here is what we hoped, it was a reasonable hope,
but it didn't work. Here is what we learned. By the way, here is a new
experiment based on what we learned, and here are the potential benefits
of this experiment."
I thought this whole issue was a minor one that would be a bit
interesting. It appears to have been much more than that.
-- Rol Fessenden LL Bean, Inc 76234.3636@compuserve.com