Even in mainstream filmmaking, producers and directors must operate primarily on instinct, though big studios rolling stockholders' dice at a hundred million bucks a throw certainly do everything they can to shorten the odds for a hit.

The principal methods by which they seek to better their luck are market research and marketing research.

First, they try to figure out something about the viewing public in general and what it wants to see. Then, in relation to specific projects, they try by various means to get a sense of what a potential audience will or won't like about that particular product.

In mainstream, they rely on focus groups, response cards from preview showings, consumer surveys, consultant analyses of the market performance of previous releases, demographic studies of the preferences of potential consumers, and painstaking number-crunching of the results of various sales strategies employed to flog earlier projects.

In X-rated productions, we do none of the above to any significant degree. Instead, we rely on the "throw everything possible at the wall and see what sticks" approach, which lowers the cost of releasing a given title but inflates the expense of doing business overall to a sometimes ruinous degree.

In our own defense, it could fairly be said that the kind of research tools available to giant media conglomerates are both pricey and difficult to apply to our peculiar situation as adult video manufacturers. For one thing, our audience is largely invisible and wants to stay that way. Though we know that huge numbers of people watch porn, most of them don't care to admit it, much less discuss it in detail with pollsters.

Nor can we learn much from observing the audience share that shows up for industry events. These folks are known as fans, and while fans are certainly helpful barometers to some degree and vital to the kind of viral marketing that makes certain titles, performers and directors bankable, they also tend to skew young and reckless or old and obsessive, making them too atypical for effective study.

Precisely because we rely so heavily on fan input in certain genres, as producers, we're often much too vulnerable to small but active pressure.

Having spent much of my career making fetish-oriented products, I'm woefully familiar with this effect.

How many porn buyers are really into watching girls smoke? I doubt it's a whole lot of them, but they sure are vocal about it. True, these few, outspoken consumers sometimes turn out to be predictors of larger trends.

I had one guy who wrote to me for years, back in the day when younger was universally believed to equal better, hectoring me to use more older performers, starting way before the whole MILF phenomenon emerged. Likewise, what we now call "pegging" and see commonly even in feature porn, was distinctly the province of a tiny but determined minority 10 years ago.

It's all well and good to pay close attention to the feedback of fans and niche consumers, and potentially a make-or-break necessity for smaller companies that depend on the repeat business of a narrow client base, but for larger producers faced with ever-escalating production costs and needing a larger, more conventional audience faced with thousands of new titles every year, simply cruising the chat boards or listening in on the casual conversations of autograph collectors at fan conventions isn't particularly helpful.

All we find out that way is TMI about too few consumers.

Even sales figures are misleading. Simply copying approaches that have made money for other producers guarantees nothing.

The competition may be better at whatever it is they do, or they may have hit on a formula that works for an audience they've already got locked up, or may be a nonrecurring phenomenon that made the racks at precisely the right moment.

If we're going to right our industry from its current wobbly course, we'll need better navigational information, and the good news (notice how I always try to work some of that in here) is that research tools may already exist for our specific needs that don't entail vast additional expense.

Most large companies now have websites they use to promote their products. That pipeline can work both ways. Anonymous polling and consumer reviews (carefully edited before public posting, preferably) can give a producer important clues as to what its existing customers like and don't like about the pictures it's making.

For physical packaging, survey inserts offering discounts on future purchases or other premiums can induce wary watchers to scratch a few hash marks on a card and drop it in the mail. Banner advertising on a variety of adult sites with coded response URLs for each can tell us something about where our target markets surf.

And for bigger releases, I wouldn't rule out some of the tactics employed in mainstream.

Screenings and promo parties not only generate buzz, they allow marketing personnel to interact with a public specifically interested in a given company's particular type of program, as opposed to porn in general or porn stars in particular.

But all these measures just chip away at the great, unanswered question that haunts every producer: Who watches porn in the first place?

Because of what we assume from the early days of theatrical exhibition and arcade traffic, we tend to pitch our products toward sexually frustrated guys in the 25-to-45 age bracket.

Of course, by doing so, we attract a lot of those guys. But we also make a lot of products that don't attract, or even repel, large blocks of other prospective buyers.

Anecdotally, we suspect that a lot of women watch gay porn because the guys are better looking. Since big feature companies started paying more attention to the attractiveness of straight male performers, they're undoubtedly pulling in more women and couples.

Likewise, the assumption that the women-and-couples market prefers plot and dialogue to wall-to-wall sex may be an anachronism from a time when that market was principally cable-oriented, as opposed to buying physical products or downloads via the web. Wall-towall sex with high-production values and a minimum of stunt-sex may be just what that market, which has grown bolder over time, now craves.

While it may be a bit early to call in the consultants just yet, we simply can't afford the luxury of trial-and-error marketing under the current challenging conditions.

What we don't know can't just hurt us, it can bury us, and as of right now when it comes to who buys our stuff and what they really want, Goldman's axiom applies with a vengeance. We don't know anything, and that's got to change.