Last week, the Wall Street Journal reported on the exaggerations and assumptions required to make the claim that Michael Jackson sold 750 million albums. He sold a lot of albums, but probably nowhere near 750 million, and not just because of those meddling kids with their cassette dubs and their file sharing. ItÁ¢€â„¢s simply that fuzzy numbers are everywhere. They’re a function of good-faith estimation, spotty data, and outright lies.

Numbers frighten people, but they really shouldnÁ¢€â„¢t — they’re just as subjective as anything else. They look like they’re the result of precise calculation, but often they’re invented as a way to put an aura of precision on someoneÁ¢€â„¢s opinion.

Sometimes a number comes from deduction rather than data collection, with a result that’s reasonable, logical — and completely made up. ThereÁ¢€â„¢s a type of job-interview question commonly used by consulting firms called the case interview, where the interviewer throws the candidate a question to see how he or she approaches the problem. A simple example might be something like Á¢€Å“How many pencils are purchased in the United States each year?Á¢€ The candidate is expected to deduce an answer: Á¢€Å“Well, there are 330 million people, and letÁ¢€â„¢s assume that everyone uses at least one pencil a year. But some people go through more. Elementary school children only use pencils, and letÁ¢€â„¢s say they make up 10 percent of the population and use 12 pencils a year. Some adults prefer to use only pencils, too — letÁ¢€â„¢s make that another 10 percent. Then weÁ¢€â„¢ll assume that 5 percent of the population either bowls or plays golf and uses those little half pencils, so weÁ¢€â„¢ll say another 12 per year. So thatÁ¢€â„¢s 25 percent of 330 million using 12 per year, and 75 percent using only 1 per year, so the answer is: 990,000,000 + 247,500,000 = 1,237,500,000 pencils.”

Simple, yes?

But then you get cut from the next round of interviews because you forgot that the golfers and bowlers will also be using pencils outside of their sport and must be accounted for somehow.

The deductive-reasoning approach is common when trying to assign numbers. ItÁ¢€â„¢s one thing to say that Michael Jackson sold a heck of a lot of records, but how many is that? More than Elvis? Less than the Beatles? Are we counting the Jackson 5, or not? Sales to record stores or to consumers? As you define the question, you get into the different parts that can affect the final result because they affect your ability to collect the data in the first place.

IÁ¢€â„¢m not saying numbers should be dismissed — not at all. Good data help us make better decisions. However, not all data are good, and all decisions are, in part, emotional. ThatÁ¢€â„¢s how we can be manipulated. Think the current U.S. health care system is just dandy? Compare prostate-cancer death rates from different time periods with no context about diagnosis rates. Want to make the case that Michael Jackson is the biggest pop star ever in the history of the world? Claim he sold 750 million records.

In the case of books and records, at least in the olden days when Jackson was most active, the retailers tended to be small independent shops that had rights to return unsold inventory. The record labels and publishing companies dealt with distributors that handled the merchandise, and discontinued inventory might be resold at discount prices in special cutout or remainder bins. The result? Slippage between what was shipped and what was sold — at what price, to whom, and when. And thatÁ¢€â„¢s just in the U.S.

Maybe I feel a special pain because I’m an author, and I can never tell how my books are selling. I receive a statement from the publisher twice a year with sales figures; it arrives about three months after the close of the period. Amazon.com publishes sales ranking numbers, but those are easily manipulated in the short term. Several book promoters suggest that authors ask everyone they know to buy a copy of their book at an odd time, say 2 AM on a given Wednesday, to catapult the book into the Amazon top-ten list, even if just for a very short time. You can then take a screen shot and tell the world of your success, even if you never sell another copy.

Jackson sold a lot of records. Eating vegetables is good for you. The climate is changing. But how many records did he sell? How exactly are they good for you? And how fast is it changing? Who knows …