Why some new shows get canned after one episode

WASHINGTON — Remember "Emily's Reasons Why Not"? Of course you don't. So few people watched the ABC sitcom's debut last season, despite heavy promotion and Heather Graham's sex appeal, that the show achieved a distinction shared by only a dozen or so series in network history:

It was canceled after one episode.

Although a program can't get the ax sooner than that — there's no record, after all, of a series being canceled during its first broadcast — "Emily's" instant disappearance put it at the bleeding edge of a network trend. A decade or so ago, a new network series might get a dozen episodes to find an audience. Now the end tends to come quicker — often much quicker. (During the 1995-96 season, 39 series were canceled before they aired more than 10 times; last season, 61 were spiked before they attained that modest milestone.)

If recent history is any guide, about half of the 26 new shows that will appear for the first time this fall on the major broadcast networks will appear for the last time this fall. They will be vaporized just like last season's short-lived losers, such as "Emily," "Head Cases" (which Fox ran twice before pulling), NBC's "Inconceivable" (two episodes) or CBS' "Love Monkey" (three on CBS, before the show expired on VH1).

Network executives say that each cancellation represents a small failure — because each new show is a multimillion-dollar investment of production and promotion — but they also say that excessive patience is frequently futile. Given the number of programs available to viewers, they say, a show that isn't attracting viewers its first few times out faces longer odds of ever being "discovered" several more weeks into its run.

"There are so many choices out there now," says Kelly Kahl, CBS' head of scheduling. "You just can't afford to stick with something that's not working for too long."

Increasingly, he says, the debut of a network series is like opening night on Broadway or the first weekend of a feature film's release: "Frankly, there will be some shows that just won't open," he says.

Part of broadcast TV's shortening attention span is a story of technology. These days, programmers have ever more sophisticated tools to detect what's working and what isn't.

The networks have long known how many viewers a show has, but advertisers have long wanted to know who those people are. In 1999, Nielsen, the ratings company, introduced a software system called Npower that enables the networks to analyze the income, education, race or ethnicity, and location of sample audiences. The system also shows where a target audience (say, women age 18-34) is going after one show ends and another begins — a process known as "audience flow."

Armed with such data, a network can determine rapidly when to schedule a program relative to its other shows (or relative to another network's programs) to increase the desired flow. That kind of analysis used to take weeks, as viewers settled into predictable patterns; now it takes a few days after an episode airs, according to Nielsen.

The software "will allow [a programmer] to look, minute-by-minute, at how a viewer watched television that night with a full analysis of the demographics of that viewer," says Karen Watson, a Nielsen spokeswoman.

Then, as shows that aren't attracting the desired viewer get pink-slipped quickly, the holes often are plugged with reality shows, which are cheaper than dramas or sitcoms and can be produced quickly.

As a result, the very idea of a "new TV season" each fall is becoming increasingly quaint. The networks still debut dozens of new shows in September, just as they have since the late 1940s, but the weeding of the schedule now goes on year-round. Many long-running shows made their debut at midseason in January; for the past five or six years, a big launch period has been the historically drowsy summer months, when programs such as "Who Wants to Be a Millionaire," "Survivor," "Dancing With the Stars" and "Deal or No Deal" debuted.

All this weeding and reseeding has diminished the chances that a slow-starting series will build momentum and develop into a hit. Some of the most important and popular shows ever owe their survival to network executives who were willing to endure low ratings early on — shows such as "Hill Street Blues," "All in the Family," "M*A*S*H," "Cheers" and "Seinfeld."

Such slow-developing hits still happen — NBC's "Scrubs" is an example, as is "Close to Home" on CBS — but Kahl notes that it takes an intensely loyal fan base or tremendous critical support to keep an underachiever on the schedule when the numbers don't add up. Fox famously stuck with the ratings-deficient sitcom "Arrested Development" for several seasons, collecting several Emmys, before finally giving up on it last season.

John Rash, a senior vice president at the Campbell Mithun ad agency in Minneapolis, predicts that the 2006-07 season could see a return to some stability on the networks' schedules, albeit modestly. Rash bases his prediction on two factors: the increased number of serialized dramas — the new "Kidnapped," "The Nine" and "Six Degrees," for instance — and the sorry state of the sitcom after such '90s shows as "Seinfeld."

The serial dramas — which are much in the style of "24," "Desperate Housewives" and "Lost" — feature weekly cliffhangers, enhancing the prospect that viewers will keep coming back, Rash says. "If viewers repeatedly invest in these shows only to be left hanging (by an abrupt cancellation), it will be very difficult for the networks to sell the next serialized drama," he says.

As for sitcoms, the recent failure of so many suggests the need for more patience; Rash points out that NBC kept "The Office" on its schedule last year and eventually was rewarded with a modest ratings success and an Emmy for best comedy series.

"The industry is realizing that their search for the next generation-defining (sitcom) may still be a ways off," he says. "They don't have a lot of choice" but to stand pat.

Which would mean that the newest change about network TV this fall would be to make not much change at all.