Monday, July 11, 2011

Satisficing

Theory of Constraint talks a lot about avoiding local optima, about maximizing the system and not its parts. This leads to road runner ethics, to making sure we are not trying to be more accurate than the noise in the system and so forth. The Theory of Constraints solutions tell us what to do at a good enough granularity. The rules are simple. Don't protect every step, protect the system by protecting the constraint, maintain flow. The tools are there. DBR and sDBR, CCPM, Logistics. So what's the problem?

The problem is there is still a void, although it seems to be just a verbal one. If we are not optimizing, if we are not "doing the best we can", than what are we doing? We can't just brush this off as being "just a verbal thing, just definitions", definitions are the starting point of clear thinking, as Dr Goldtratt taught us repeatedly.

There is another reason that makes this important. While the human logic can accept fuzziness, our subconsciousness can't, and that's where it counts the most, since most of our decision making is actually done there, at the most instinctive level. There are no voids in the subconscious mind and there are no gray areas. Like Master Yoda said in one of the Star War movies: "Do, or do not, there is not try". That's your subconscious mind for you.

So, for our subconscious mind not "doing the best we can" is, actually "not doing the best we can" or, if we'll put English grammar aside for a second, "doing NOT the best we can". Which sounds bad, doesn't it?
So, we really need a good name for not optimizing. That name, to me, should be "satisficing", a term I first heard during my MBA from Prof. Boaz Ronen, a TOC disciple in the academic world.
Satisficing means being good enough, not perfect. It has lots of merits. It reduces "analysis paralysis" and stress levels, allows for better flow through faster outputs  (I demonstrated this while discussing the hiring process here) and it sounds good.

What do you think?

Edited to add:
Following the replies I got here and on LinkedIn, I guess satisficing needs a bit more elaboration. Here's an easy example - say I have to prepare a birthday bash for a 6 years old and my goal is to make that 6 years old as happy as possible. I know I'll make a chocolate cake, cover it with chocolate cream and decorate it. Once the cake is ready I can make this cake:   decorated with M&Ms and a sugar paste "Happy Birthday". Time invested is about 1 hr and child is very happy.


On the other hand, I can make this cake: covered with sugar paste and decorated with sugar paste designs. Time invested 4-5 hrs (at least) and child is very happy, perhaps a bit more than with the first cake but not significantly so.

Both cakes are eaten up at about the same rate.


(yep, I made both of these cakes, and other "optimizer" cakes you can see here.There is room for hobbies in this wold, IMHO)

What we see is that I can go for a satisfactory cake or an optimal cake. The result, in goal units of child happiness, will not vary substantially, the only difference is the effort I put in. The first cake demonstrates satisficing - doing the necessary, the sufficient and nothing over that. The second cake demonstrates optimizing - doing the necessary, the sufficient and everything else, trying to be perfect.

One last thought (Thanks Henry for bringing it up) - whatever you do, it has to move you in the right direction. It is not enough to "do it right", you have to "do the right thing right"

I'd love to hear what you think about this.

7 comments:

Bob Sproull said...

Hi Orion....great blog! I think maybe a better word might be subsatisficing since we are accepting suboptimal thinking and performance. I firmly believe that we owe the focus on local optimal to our Cost Accounting friends who continue to push metrics like efficiency and utilization as performance metrics. Metrics drive behaviors or as Eli so eloquently put it, "Show me how you measure me and I'll show you how I behave." I think it's that simple.

Bob

Unknown said...

Bob,Oh, I totally agree to the fact that metrics derive behaviours. The thing is you see organizations take on TOC and then digress back. I have a hunch that the self measurement process is at fault there, that even after top management stops pushing in the wrong direction there is a self inclination to optimize, because it must be good. That is why I would disagree with subsatisficing, it makes one feel in the wrong direction. There is such a thing as the power of the positive. In "The Choice" Dr Goldratt wrote something to the likes of "you can think you can or you can think you can't, either way - you'll be right". I did not use that quote in the post since I did not have the book at hand to make sure I quote correctly. So I wholeheartedly, completely believe that we must make sure the right direction is always positive, positive sounding, positive looking and so forth.Just before starting the seminar I did at the Goldratt House, I had to select a new password for something, as a memory aid I choose a sentence that said "this is a new start". It's been some 5 months now, and it was.Orion

Francois said...

Orion, even though you appear to be jsut looking for correct terminology, the contents of the blog still educate - thank you!

With tongue-in-cheek, maybe a neologism should be used here - I mean, we all are familiar with the terms "kaizen" and "poka-yoke" from the Japanese production environment, so why not simply talking about "yoding", derived from the Yoda quote in Star Wars?

The background to your blog, that of lcoal optima, again reminded me of my personal search - for the ideal way of "managing (employee) performance" as the current version is really focused on lcoal optima - but time for that later on.

I enjoy your blogs - thanks a lot!

Etienne said...

Satifycing = Necessary, but not sufficient.

Unknown said...
This comment has been removed by the author.
Unknown said...

Etienne,

Why do you say that Satisfycing = necessary but not sufficient? if that is so than we should optimize.
The way I have Satisfycing defined is "the necessary, the sufficient and nothing else"

Orion

Etienne said...

Hi Orion, I'm having difficulty understanding the link between 'local optima' and "Satisfycing". I see the point of reaching "Statisfycing" somewhere before the goal is reached where at that point one knows there is still a way to travel to reach the goal but at the same time one knows that what has been achieved up to that point, is good enough to actually satisfy enough of the Desirable Effects (DE's) that cause the goal to be reached.

In other words, to take your example of the cake, the goal comes from an expectation for a cake which has certain attributes. The point of "Satisfycing" is reached during the building (preparing, mixing, baking, decorating etc.) of the cake where at some point one thinks it is good enough. That point is not the goal but it is close enough to satisfy the DE's.

If Satisfycing, as it is defined, is the "good enough" solution = necessary, sufficient and nothing more, doesn't the "sufficient" mean the goal has been reached?