Ken Levine’s new post “Your life is not a sitcom,” about people who think that their wacky workplace and co-workers would make a great show, touches on a theme I consider very important: the important thing about a TV show is not the setting or even the individual characters, it’s the relationships between the characters:
Here’s what nobody ever pitches me: a show about a relationship. THE OFFICE is funny because of the relationship between Michael and his employees. It is funnier still because of the relationships among the employees. What they actually manufacture is completely unimportant.
Start with the characters first.
What about the dynamics between them are interesting, fresh, and could sustain stories week after week? And then, what is the best setting to put them in? One that hasn’t been seen before is a plus but not imperative. How many shows and plays and radio series have been set in bars?
That said, there are some areas that are tougher sales than others. Madcap terrorist cells probably won’t fly.
Every TV character is to some extent a known quantity — there are only so many character types to choose from. And the setting, too, is not really that special, no matter how unique and high-concept it might seem. What gives a show its individuality is the way characters interact with each other, how they relate to their environment (the setting) and how they deal with the situations that come their way. A setting and a character, described on their own, are always going to be a lot like every other setting and character; the relationships are where the new stuff happens. The things that make a character “wacky” tend to be the things that make him exactly like every other TV character you’ve seen, but when he makes a choice that another TV character might not have made, or has a relationship with his spouse that isn’t exactly like the relationship of [fill in names of another couple on another show], he becomes something resembling a person.
This is one reason why most successful or semi-successful shows tend to drift away from their original settings, abandoning the high-concept premise for a generic one. 30 Rock started out as a show about a particular kind of workplace, and ended its first season as a show about an almost completely generic workplace; it was pitched as a show about making late-night TV, but it wound up being any workplace where the boss has to deal with inefficient underlings and meddling corporate management. Tina Fey did what most people just dream about: she made a show about the place where she used to work. But that idea couldn’t sustain a show for more than a few episodes, because once you’ve used up all the wacky real-life workplace stories, it’s time to move on to a more generalized setting and more specific relationships. It’s almost like TV is better when the setting and concept are more generic.
Successful TV producers and executives tend to understand that while a high-concept premise or unusual setting might be a good selling point for advertisers or network executives, the concept itself is the least important part. One of my favourite TV executive quotes is from Grant Tinker, who said essentially that he didn’t care about the premise of a show:
Ideas in themselves are never interesting to me. It’s all a matter of execution.
And Bob Boyett, not a quality TV producer but a successful one, said something similar:
Basically, the concept of a show is merely a vehicle to get it launched. What keeps it going is the ability to present characters people want to follow.
In other words, the fact that a show takes place in the craziest office/home/military base/terrorist cell ever is not going to keep it on the air.
Update: I shouldn’t make it sound like there’s absolutely no difference between one premise/setting and another (though I came close to making it sound exactly like that). Different settings do produce different shows, and the setting itself can be a character in the show, like Sunnydale or Mayberry. But when a show depends too heavily on the idea that the uniqueness of the setting is enough to make the concept viable, then it runs into trouble.