Broadcast stations used to be largely autonomous. Some still are. Managers of local stations can easily make local decisions on programming, advertising, personnel and all other questions that must be reviewed daily to keep a transmitter on the air. But the economics of broadcasting have dramatically and irreversibly changed. The combined effects of falling revenue and rising costs have tipped the scale permanently towards less autonomy for all stations, whether they are part of a group or not. Ad revenues have fallen steadily for an uncomfortably long time. Networks are considering reverse compensation to stem their own altered economics. Unfortunately, all this has happened in an era when the FCC has granted the broadcasters' wish to have an extra channel on which to broadcast HDTV (or other services as finally authorized).
Every month, boardroom conversations consider ways to improve economics for broadcasters. Some broadcasters postulate that centralization, commonly termed centralcasting, may be the only way to save local broadcasting from imploding in the jaws of the economic vise it is caught in today.
Centralcasting, or more generically, centralized operations, can increase cash flow and help balance the balance sheet. But it involves making careful, sometimes difficult, choices. Experts who have studied centralized operations liken these choices to picking fruit from a tree. The low-hanging fruit represents the easiest changes to implement, the higher fruit more difficult — and some fruit may be out of reach. When broadcasters think of centralcasting, their concept is often limited to creating a common master-control operation for stations with common parents. But there is other, lower fruit to consider first.
One candidate for centralized operations is the traffic department. Before today's highly computerized operations, a station's traffic department worked hard to output a log from paper records. Now, the staff is decidedly smaller and the work highly automated. By centralizing traffic, a group owner can marginally reduce the number of people required to keep multiple stations operating, and the relatively low bandwidth required to connect a local sales operation to central traffic makes it quite economical.
Only slightly higher up on the fruit tree is the promotions department. Often, the promotions department bargains hard to get the personnel and capital resources needed to do a good job of promoting the station. Computer non-linear editing and production systems have made sophisticated spot creation less expensive, but the creative individual is no easier to find or cheaper now than in the past. Centralizing promotions also comes with a little penalty in distribution cost. If the central promotions department is created as part of a broader interconnection, the distribution may come nearly for free.
Above the low-hanging fruit lies the potential for large savings over the long haul. The true redundancy among broadcast stations varies with factors that are principally tied to the overlap of the individual programming grids. If all of the stations to be centralized have the same network affiliation, it seems logical that redundant processes can be centralized.
One of the approaches first used was based simply on remotely controlling the station from a central site, but not moving any of the video hardware or content to a central repository. Using software like PCAnywhere, or a remote-desktop control built into Microsoft's free NetMeeting conferencing package, it is easy to see how you could simply remote the display from the automation system to a central site and eliminate the local person watching the screen for errors. When coupled with a return path for low-bandwidth video and control/status on the transmitter, this approach seems to be the least capital-intensive topology possible.
Of course, with low cost comes tradeoffs some stations will find unacceptable. If the connection to the remote PC is lost due to a power failure, the process of restoring control, perhaps even programming, might not be quick. With few, if any, eyeballs watching the store at the local station, minor problems that might be solved in a couple of minutes by a skilled operator become much more complicated.
At the other end of the spectrum lies a fully centralized facility in which a local station without news programming might consist of nothing more than a connection to the transmitter and the return status that is required in all interconnection topologies. This allows the local station to avoid any renovation as equipment wears out. And, if the interconnection bandwidth is high enough, the full ATSC DTV signal can be sent from the centralcasting site. If the DTV stream contains a copy of the NTSC programming, it is possible to decode that stream from the multiplex and convert it to NTSC at the transmitter site, reducing the bandwidth to no more than 19.3 Mbits downstream, and perhaps 1.5 Mbits upstream. The central site can provide a level of N+1 redundancy that the local station might not be able to afford, since spares are spread across a larger capital base.
Full centralcasting works best when the stations all are of one network affiliation, since one program stream could be concatenated and only interstitials would vary. But that's a pretty idealized case, to which any owner of multiple stations will attest. Thus, the probability is that a unique program stream is assembled for every station and some of the redundancy the centralcasting site could enjoy is lost, requiring more hardware.
Centralcasting has its own Achilles' heel. The cost of interconnection is the combination of local access cost to a high-bandwidth network (local loop), and the long-haul interconnection circuit (often inter-LADA). WAN circuits with the appropriate quality-of-service guarantees and sufficiently robust disaster-recovery provisions can be expensive. High-quality video can be transmitted in as little as 6 Mbits/s, but full-time 6 Mbit circuits are hard to book unless one uses DS3 service (nominally 45 Mbits/s). The excess bits cost the same as the fully occupied ones, wasting money.
Packetized transport, like ATM, offers potentially flexible WAN bandwidth. Some argue that ATM is inappropriate for dedicated video service. While it might suffice to note that FOX Sports and others use ATM on a permanent and full-time basis, the engineering arguments should not be ignored. Confusion about the suitability of ATM begins with its ability to deliver isochronous video, i.e., with deterministic delivery. However, ATM adds other features that are potentially useful in centralized operations, including multicasting the same packets to multiple locations, a wide variety of standard computer-industry interfaces, and interfaces designed specifically for converting video data types to ATM.
Another variation on the centralized operations theme stands with one foot in both camps. Distributed broadcasting, a technique that requires careful evaluation by competent personnel, holds the promise of reducing labor to the same extent as centralcasting, but does not necessarily play all programming from the central site. The probability is high that the station has equipment in place to receive and air network programming without receiving it from the hub site. Taking advantage of that existing hardware may reduce the bandwidth requirements in the WAN. Such an approach requires that some of the programming plays from the distributed broadcasting site and some plays from the station. While complex, this approach allows distributed servers to move media efficiently between sites in the background using FTP or other techniques under automation and asset-management control.
If all television was recorded and played under automation control, all of these strategies would work well. But what seems simple becomes potentially difficult when considering live events. In the centralcasting-hub case, it may be necessary to have multiple operators on duty during peak sports times. Networks have a habit of regionalizing networks, and football teams don't coordinate their plays to fit into a standardized model for inserting commercials. A hub that sends out signals to FOX, NBC, ABC and CBS stations might find a situation quite impossible for one operator to handle, though one operator could arguably control most normal programming on over 10 stations.
As the number of combined stations rises, the potential total capital-cost reduction rises, but not exactly at the same rate. As the number of streams at risk rises, the level of redundancy needs to rise to avoid the catastrophic case of many stations going off the air due to a single failure. It seems unlikely that many broadcasters would have the capital to take such an approach; at least until operations have proven successful for a while and net savings have accumulated. But the savings in capital cost have to be looked at over years. With depreciation on an aggressive five-year basis, a 20 percent savings nets only 4 percent of that amount to the bottom line per year. While the total is important, the boardroom will not see that 4 percent as worth much risk. The real savings are in human resources, traded against the cost of interconnection. This is where the rubber truly meets the road in centralized operations. If the approach you take can save sufficient personnel to overcome the cost of interconnection by a wide enough margin, you will have a persuasive case.
It is worthwhile to note that centralized operations is a fundamental shift in the business operations of the company as well. Broadcasting essentially has been a local business since radio began over 70 years ago. In the last few years, radio has moved aggressively to centralized operations, and the once-marginal local-radio business has returned to profitability. A change at the root of television broadcasting may well achieve the same result. But it will take years to fully explore the quirks in the factory operations required by centralized operations, years to minimize risks and years to see the full benefit of centralized operations come to fruition.
John Luff is vice president of business development at AZCAR.
