SELECT * FROM Celko – September 2010

One of the major concerns in the fashion and consumer industries is predicting trends. There are companies that do nothing but track trends for record companies by looking at play lists, sales figures, radio requests and downloads. The youth-oriented clothing industry has people who go out on the street, to malls or to school campuses where they can see what the kids are actually wearing. Other industries use market research, focus groups and surveys to try figure out what they should be doing.

Trend spotting is a lot of guess work and intuition. And a lot of this expert opinion is wrong. Many years ago PBS did a FRONTLINE show entitled “The Merchants of Cool” about predicting the teenage market. In the documentary, the “mook” and the “midriff” were the stereotyped teen consumers. The “midriff” was a highly mature and sexual teenage girl. The “mook” was a teenage boy locked in infantile, boorish adolescent behavior.

The kids are not like these shallow stereotypes. But they are an estimated $150 billion a year market. So we listen to the witch doctors. If their advice works, they take credit and a large chunk of cash. If their advice fails, they take a large chunk of cash and talk about “the next big thing” that is coming up.

In the fashion industry, the phrase “[insert color here] is the new black!” is one of the stock phrases. In IT, we talk about “new paradigms,” but we have a better idea about technical trends than guessing about upcoming color choices. Moore’s Law is a prediction by Gordon Moore of Intel about the number of transistors in a computer chip. Some people first quoted it as the number of transistors will double every 24 months and other people use 18 months, but Moore never gave a timetable in his original article. David House, another former Intel executive, actually gave us the 18-month figure. The actual figure when you graph it out is every 20 months, but House was close enough. Consumer marketing companies would give anything to have something like that.

Moore’s Law was not predicting a trend; it followed and measured the trend once the trend showed itself. Just having technical knowledge will not guarantee successful predictions. In the mid-1960s, Sylvania ran an institutional ad showing what their research people were doing. It featured four products. The only successful product was the Blue Dot Flashcube. This was a small plastic cube that held four flash bulbs in a socket on a camera and rotated them as they were used. The other three inventions were much flashier (pun intended) than the flash cube, but they did not make it. For example, there was a vacuum tube filled with an inert gas mixture that glowed blue-green instead of red-orange and would never burn out. It was perfected just as transistors became dirt cheap.

If Google and the other giant search engines had existed in the 1950s, they would not have been able to predict mini-skirts, bell bottoms and Nehru jackets in the late 1960s and early 1970s. Asking “What is the next big thing?” is a job for some super advanced artificial intelligence engine.

Let me postulate a super advanced artificial intelligence engine. It has taste and good fashion sense. But where does it get its data?

One of the marvels of this age is the ability to get all kinds of information in all kinds of digital formats. We expect that we can find scholarly publications and classics of literature in some kind of searchable format (paper, then microfiche, then digital, then who knows what). But we forget about non-scholarly publications and the pop-culture things that quickly disappear.

I used the word “marvel” deliberately because last weekend I browsed through a collection of old 1950’s CAPTAIN MARVEL comics on a DVD. It is one of many scans of old comic books from The whole four disk set is under $50 and includes comic book reader software on the disk. Yes, there is a standardized comic archive format! This delights my “standards uber alles” mind-set.

When I have been a really good boy, I will treat myself to a weekend of PLASTIC MAN and POLICE comics. Now to my point: no single issue of THE MARVEL FAMILY comic book would have been that cheap. Yet I could easily, casually, do a “mouse-click” on my computer one night while I was thinking about my early teenaged infatuation with Mary Marvel.

Data is now cheap. Really cheap. Not just commercial data in the enterprise where it sits on archival tapes or other media, but every bit of data. And it is abundant. I can get MAD, NATIONAL LAMPOON, NATIONAL GEOGRAPHIC magazines, all the major newspapers and a multitude of other publications. The Sears and J. C. Penney catalogs are a better look at what people actually wore than VOGUE or any of the high end fashion magazines. And I have not touched archives for movies and television programs, popular music and old catalogs.

There is a whole cultural history out there, rich in information. Blogs, tweets and other social websites can replace sending observers to the streets.

Sometimes you are lucky and can point to the start of a trend. It is the generally accepted wisdom that the haberdashery business in the United States was destroyed when President John F. Kennedy stopped wearing hats.

But more often, it is a converging of many factors. High taxes on adult clothes in the UK led to girls buying children’s tax-free clothes and the mini-skirts were born. British music dominated American pop music during the “British Invasion” period. This was made possible by transistor radios and lower priced consumer electronics. Suddenly, American kids were all trying to be “mod” and we entered one of the weirdest periods in popular fashions.

But how would you have asked the question? How would you have reacted if the super advanced artificial intelligence engine in my fantasy had told you that people will be wearing neon colors, paisley prints and platform shoes?


submit to reddit

About Joe Celko

Joe is an Independent SQL and RDBMS Expert. He joined the ANSI X3H2 Database Standards Committee in 1987 and helped write the ANSI/ISO SQL-89 and SQL-92 standards. He is one of the top SQL experts in the world, writing over 700 articles primarily on SQL and database topics in the computer trade and academic press. The author of six books on databases and SQL, Joe also contributes his time as a speaker and instructor at universities, trade conferences and local user groups. Joe is now an independent contractor based in the Austin, TX area.