Puzzles and Mysteries

I was rummaging through old blog posts and came upon this one from almost 13 years ago. Since no one read my blog 13 years ago, and I liked this post, I thought I'd re-post it.

An article by Malcolm Gladwell in The New Yorker leads me to believe that advertising people can learn something from spies about solving business problems.

Gladwell tells us about a national security expert, Gregory Treverton, who distinguished between two kinds of problems: puzzles and mysteries. His distinctions have great value for us.

Puzzles, he wrote, are problems for which there is not enough information. An example of a puzzle: Where is Jimmy Hoffa buried? If we had more information, we would know the answer. If someone told us “Jimmy Hoffa is buried in New Jersey,” we’d know a little more than we know now. If they said,“He’s buried in northern New Jersey,” we’d know even more. If they said,“He’s buried in the Meadowlands,” we’d have an answer to our puzzle.

On the other hand, there are mysteries. Mysteries are problems for which we have plenty of information, but no accurate analysis. An example of a mystery: Why do inner-city schools do such a crappy job of educating kids? There are thousands of studies. Every education department of every university in America has done a study on this; every committee of Congress has done a report on it; every editorial writer has a theory about it, and every pundit has an opinion. And yet, we have no definitive answer. More studies and more information are not likely to yield an answer. What is needed is an accurate analysis of the voluminous information that already exists.

Gladwell gives a wonderful example of the importance of distinguishing between puzzles and mysteries. In 1943, during World War II, the Allies were concerned about Germany’s boast of having developed a “superweapon.” Did they really have a superweapon, or was it just propaganda? There were two ways to deal with this problem. It could be dealt with as a puzzle, and spies could be sent out to gather more information. Or it could be dealt with as a mystery, and information the Allies already had could be analyzed.

Fortunately, American intelligence had a bunch of brainiacs who were known as “the screwball division.” They were “slightly batty geniuses” who were brilliant at analyzing information that was readily available to anyone. So instead of sending out spies disguised in mustaches to infiltrate German munition factories, the intelligence community gave the brainiacs German newspapers and radio broadcasts. These guys pored over the information that already existed and accurately predicted that the Germans had, in fact, developed a new weapon, the V-1 rocket. And they also accurately predicted both that it had been stalled in development and when it would be ready.

As a matter of fact, in retrospective analysis, these guys had been correct an amazing 81% of the time in their analyses during the war. A success rate many times higher than the spies.

How does this apply to us? We ad people are almost always tasked by our clients with solving some variation of the following problem: How do we sell more stuff? Sometimes it is a subset of that question like -- who is our target customer? or, what should our primary ad medium be? or, which of these campaigns should we go with? or, what should our brand position be?

One hundred percent of the time these problems are dealt with as puzzles, not mysteries. We always assume that one more study will yield the magic answer. Research is commissioned. Spies are sent out to live with customers, or to interview them, or hold group discussions with them.

The results of these endeavors are all too frequently disappointing. The methodologies are usually dressed up to appear scientifically bullet-proof. But the initial enthusiasm for the process often recedes when the answers aren't conclusive or unimpressive real-world results start rolling in.

On the other hand, most companies have stacks of research and reams of data about their customers and about their industry. This mountain of existing information is almost never consulted. In my entire career, I have never seen or heard a marketing problem treated as a mystery. I have never once heard a marketing officer say,“You know, we have all this research we’ve done over the years and all this data from the industry. Before we do more research, I want someone to go through this stuff and tell me what it means.”

Why do we prefer to deal with business problems as puzzles rather than mysteries? Because it's easier. Going through thousands of pages and discerning patterns is hard work -- the important patterns are not usually visible on the surface. It requires a special intelligence to be able to examine imprecise, unfiltered, inconclusive, often contradictory information and come up with a correct analysis.

It is much more comforting to send a researcher or planner out with a fake mustache and give her six weeks to come back with an answer. Unfortunately, as we all have experienced, an alarming amount of the research we do yields little of actionable value. The following year the report becomes part of the dusty heap of poorly analyzed information sitting around adding to the mystery.

So what are we to do?

The key is to break down the tyranny of titles and find out who our "slightly batty geniuses" are. Simply because an individual has the title of “account planner” or “research director” or "global strategy guru" or “CMO” doesn’t make him an expert detective. In my experience, planners and researchers tend to be puzzle doers, not mystery solvers.

However, there are people who are just naturally good at solving mysteries. They may come from the accounting department or the creative department or the media department or the sales department. We need to identify these people in our organizations. 

Before spies are sent out, these people need to be exposed to all the information that exists and allowed to weigh in on the questions we’re all trying to solve.

The hard part of solving marketing problems is not getting more information, it’s figuring out what the information we already have means.