Visit Your Local PBS Station PBS Home PBS Home Programs A-Z TV Schedules Watch Video Donate Shop PBS Search PBS
I, Cringely - The Survival of the Nerdiest with Robert X. Cringely
Search I,Cringely:

The Pulpit
The Pulpit

<< [ A hard(ware) man is good to find ]   |  When I, Cringely income will exceed that of IBM  |   [ Welcome to the World Wide Web, now go home ] >>

Weekly Column

When I, Cringely income will exceed that of IBM: Why market research estimates of Internet ad revenue just don't compute

Status: [CLOSED]
By Robert X. Cringely
bob@cringely.com

Market research is the collection of information about the sales of goods and services. It includes everything from focus groups that look at new types of cat food to A.C. Neilsen's television ratings. In the world of high technology, a lot of money is spent by vendors and customers on market research coming from places like International Data Corp. (IDC), the Gartner Group, Forrester Research, PC Data, and others. Market research companies seem to come and go, but the ones I have just mentioned have been around for quite awhile.

An important thing to consider when examining any market research is whether the research is looking back in time or forward. Research that looks back generally consists of counting how many people watched a show, bought a widget or walked past a showroom. Market research that looks forward attempts to guess how many people will watch a show, buy a widget or walk past a showroom.

Research that looks back is often (though not always) reliable. Research that looks forward is a total crapshoot. Read it, but don't trust it.

Part of the problem with forward-looking research is methodology: Just how did they come up with this estimate? Looking deep into the souls of the odd market "analysts" will often show they have no methodology at all other than the random application of an s-shaped demand curve to markets that they expect to grow.

Take Internet commerce, for example. How big will this market be in the year 2000? Goldman Sachs says $73 billion. Forrester Research says $45 billion. IDC says $189 billion. That's a four-to-one variation and much bigger than any expected standard deviation. Clearly, somebody is really, really wrong, yet all three companies stand by their numbers.

What I say is that Internet commerce will be bigger in 2000 than it is in 1997. That's all I can glean reliably from the numbers above, yet even this has value. At least it is a growing market.

Sometimes the variation -- and even the direction -- of projected growth has more to do with the expectations of some customer than it does with anything else. Remember businesses are PAYING for this, er, "information." Back in 1987, there were a number of market research firms -- some of which are no longer with us -- that projected IBM's OS/2 operating system would surpass MS-DOS and Windows in unit sales by 1990. With perhaps one million OS/2 users today that aren't automatic teller machines, and around 200+ million DOS and Windows users, these estimates could not have been further off. Yet they made sense if you take into account the original question posed by IBM: What year will unit sales of OS/2 exceed those of DOS and Windows? The bias was built into the question and the question came from the customer -- IBM. What analyst was going to tell Big Blue back in 1987 that their cause was hopeless?

I would have, but I've never been known for my savvy office politics.

Lately, I have seen similar figures projecting massive growth in Internet advertising in the next 2-3 years. As a guy who writes this stuff for a living and whose Web site carries advertising (look up at the top of the page -- that banner you've been ignoring is my paycheck), there is nothing I would like more than for the $180 billion U.S. advertising industry to dump 5 percent to 10 percent of that amount into Internet ads. But I don't think growth is going to happen that quickly.

Back in 1979, I wrote a book using an IBM 370/168 mainframe computer instead of a typewriter. I wrote using a line editor that had a separate number for every line. To change a word, you first had to specify the line number and then the text string you wanted replaced, which meant working with two printouts -- one with line numbers and one without. It was a difficult experiment that included one moment when I inadvertently and quite permanently deleted 8000 lines (200 pages!!!), but somehow I finished the book then attempted to deliver it to the publisher (Random House) on magnetic tape.

Today, you send a floppy disk to the book publisher, but in early 1980, no book had ever been delivered to a U.S. publisher on any magnetic medium. And my book wasn't delivered that way, either. Despite my attempts to put the file into the right form to drive the typesetter, Random House chose to retype the entire manuscript, citing concerns about labor problems with the typesetters' union.

Jump again to 1997 and the Internet advertising market, where I think something similar is happening to what happened to me delivering that manuscript back in 1980. The technology for Internet advertising is well developed, but the people in the advertising industry who can make effective use of that technology aren't. The ad agencies aren't ready for this brave new world, and so will only fund it reluctantly.

Consider the people who buy ad space for major advertisers like Coca-Cola. Who are these ad buyers? You'd think (I did) that since they decide where the money will be spent that they are among the most senior people in their companies. Nope. Ad buyers don't usually work for the advertisers at all, but for their advertising agencies. They are often straight out of college, and while this may mean they have more exposure to high technology than their bosses, it doesn't immediately make them more willing to take risks. So these young ad buyers put a little money into Internet advertising (about $400-500 million this year from that $180 BILLION budget), and that's only because they have been told to do so. Next year they will put in a bit more, and a bit more the year after that, probably in keeping with the 40 percent to 50 percent growth rates we've seen for the Net overall.

This does not mean that Internet ad revenue won't suddenly break out and show some spectacular gains. It just means that those gains can't be timed by you or me, and won't follow from past ad buyer behavior. They require a catalyst, and that catalyst has not yet appeared. Let's call it the killer advertiser.

Internet ads offer customer information never before imagined in the ad business. The terminology is already there -- ad clicks, click rates, ad views, gross exposures, visit lengths, and unique IP addresses. But who uses this technology? As far as I can figure, nobody. The ad buyers not only don't use this kind of information, they are openly hostile to it.

It's the same way that a magazine advertiser wouldn't want to know if a reader actually read the ad they placed in a given magazine. The ad buyers want to know the reader received the magazine, but not whether he/she read the ad. Why not? Because knowing whether the ad was actually read would hopelessly complicate the job of the ad buyer, who would then have to fine-tune ad placement to get maximum exposure (and do so in direct competition with every other fine-tuning ad buyer). It would be a nightmare. And Internet ad placement would be even worse because there is so much more data available. So everyone ignores the problem.

Back to the killer advertiser. This is the first person or company to embrace the complexity of fine-tuning an Internet ad campaign based on the kind of data presently ignored AND (this is important) kicking some major ass as a result. If General Motors or Eddie Bauer or American Express or Condom Express throws a lot of money into an Internet ad campaign, targets the ads scientifically, and produces spectacular results, then Madison Avenue will be changed forever. And until this kind of thing happens, Madison Avenue won't be changed by the Internet at all.

It WILL happen, but WHEN will it happen is the question nobody can answer. What's amazing to me is that I have talked to a lot of market researchers in this area who don't even know the question. Maybe they should be ad buyers. Or better still, they could be Larry Tate, king of the three martini lunch.

Comments from the Tribe

Status: [CLOSED] read all comments (0)