Monday, January 14, 2008

Sunday Afternoon Reading: Mistaken premises in Proust

Context: Marcel learns that Albertine is dead, then sometime later, he gets a telegram that he thinks is from Albertine saying that she is still alive. Later, another set of events leads him to the conclusion that this second telegram wasn't from Albertine at all, a realization which causes him to expound on the fact that we are all always making these kinds of mistakes, misreading everything with absolute confidence in judgment:

"We guess as we read, we create; everything starts from an initial error; those that follow (and this applies not only to the reading of letters and telegrams, not only to all reading), extraordinary as they may appear to a person who has not begun at the same starting-point, are all quite natural. A large part of what we believe to be true (and this applies even to our final conclusions) with an obstinacy equalled only by our good faith, springs from an original mistake in our premises.
--[my italics] The Fugitive, trans, Moncrieff, p. 671.

This passage made me think of a lot of things in our biz--broadly defined--from the challenges of having a debate/discussion with anyone who doesn't start with the same premise (E.g., See post below on the difficulty of articulating a distinction between two kinds of online experience) to a larger point about research companies who, because they are in the business of prediction, refuse to admit when they made a mistake (based on wrong or out-of-date premises) until they absolutely have to. I'm thinking of course of our pollsters in N.H. who are still scrambling to explain their failures. (race, the alphabet, weeping women, those contrarian NH'ers) Wouldn't the much more interesting and useful action be to admit that our old premises are no longer accurate and start working on adjusting the model accordingly?

I think I've posted before on this but maybe it's repeating: one of the questions I always ask researchers/testing companies/trend-spotters/insight-mavens is "Tell about a time when you were wrong and how you adjusted your model/premises to adjust to the new data?"

I haven't receive a good answer yet. Really, not one. It's one of the most annoying (and counter-productive) features of the business culture. You can never admit to being wrong without putting your livelihood at risk. Which means, of course, it's quite difficult to shed those old premises.


Beecham said...

I wonder whether thinking about this sort of question in terms of mistaken models or premises is misleading? The model may be right, but it doesn't apply to the situation, for example (I suppose you would say that that implies a mistaken premise?). Dividing our thought process into these steps--premises, models, conclusions--is appealing and suggests that indeed the challenge is to recognize when the first step was mistaken, allowing us to revise and can reach better and more accurate conclusions.

My thought is that what makes this such a hard proposition is not simply that none of us like to give up our premises, but that the challenge is more fundamental. What may be at fault is our basic instinct to find narratives.

Apparently, once we have three pieces of information the mind uses them to form an explanatory narrative. It is then cognitively hard to budge from that narrative: additional pieces of information are fit into that explanatory story. There's a study that a colleague mentioned to me once of the way in which this simple aspect of our cognition has quite disturbing effects on the ability of jurors to evaluate cases, for example.

But really, what we're all really wanting to know is, do you have an example of a time when you realized your premise was wrong and you had to go back and adjust the model?

(Maybe I'm revealing my distance from the business world to say that I would find a smart, good answer to that question extremely compelling. That may be my interest in the counter-intuitive and the conventional. Surely this is valued in the business world, too. The ability not just to overturn others' conventional thinking but even one's own would give me perhaps greater confidence even in that person or group . . . or blog.)

sk said...

Oh the number of times I've been wrong are legion: I've been wrong about target motivations (professional photographers, insurance salesmen, teens) and wrong about methodology (trying to generate emotional reactions to stimulus in a quant survey) and wrong about client expectations (assuming they wanted statistical validation when they just wanted inspiring quotes). Each mistake has altered my assumptions about how to do my job. Which I'll be sure to detail in a later post. I love being wrong, actually, because it means there is a need for the job in the first place.

I remember and miss the pleasure of pursuing an answer simply because it was interesting or elegant or counter-intuitive. But the goal of business, as you suggest is different. The answer or argument is not the end, but a means to an end, that end being profit.

The solutions people like me propose will be applied and those applications will be tested in the market. Counter-intuitive arguments can certainly be powerful (the keyword in marketing is "disruption") but they can also be totally wrong and ineffective. In fact they tend to be high risk strategies, either hitting big or missing entirely.

As one new client recently said to me after I presented my argument for a counter-intuitive approach to marketing his service, "Sounds great but six months from now, it doesn't matter what you say."

And of course he's right.