Commentary originally published in Current, May 13, 2002 By Mark Fuerst Since Internet services emerged as a small and costly part of its service repertoire, public radio has viewed streaming with equal parts optimism and revulsion. Optimism because streaming offered a way to achieve one of public radio’s most cherished goals, namely universal distribution of multiple program streams. And revulsion for two reasons: cost and potential competition.
Streaming—at least until now—has meant scandalously high bandwidth costs to reach a relatively small audience. And if streaming ever took hold, every station would face boundless new competition, from other stations around the world, from new services without transmitters, and—horror of horrors—from public radio’s own networks.
Recently, revulsion seemed to be in the ascendancy. The costs were too high, the audience and return on investment too low.
Still, in this atmosphere of ambivalence, there are optimists trying to make streaming work and there’s new evidence that streaming can cost less than expected and pull in new revenues to support those costs.
I work with one group now tilting toward optimism, the M*STAR Project, a CPB-funded multistation effort to figure out a strategy for public radio music streaming.
M*STAR brings together the Public Radio Internet Station Alliance (PRISA) and a set of "core stations," including Internet pioneer WKSU in Kent, Ohio; the country’s leading Native station, KNBA in Anchorage; and two of the system’s heaviest streamers, WXPN in Philadelphia and KPLU in Tacoma. Earlier this year, these stations were joined by four other music-oriented stations for at least part of the research—WFUV in New York; KEXP in Seattle; WPFK in Louisville; and WWOZ in New Orleans.
We looked at streaming volumes, costs and revenues and asked eight major streaming providers for quotes on bandwidth—a comparative shopping exercise that produced an almost unbelievable range of prices. Most recently, the eight stations launched a survey, asking their streamies, as we call these little-known new listeners, about how they use the streams, how they found them and whether they like the service enough to help pay for it.
Our findings are supported by information from other sources, including the stream-monitoring service MeasureCast and the Edison/Arbitron Research Reports on webcasting.
The Edison/Arbitron reports were an important inspiration for developing the M*STAR project. It was Edison principal Larry Rosin, one of the first analysts to place audio streaming at the center of radio’s web strategy, who championed the idea of streaming "side-channels"—niche-oriented extensions of a station’s main format. In fall 2000, Rosin surveyed 14,700 radio website visitors at 33 station sites and concluded: "For stations that want to build a legitimate web presence, streaming is essential. . . . 64 percent of the radio station website visitors surveyed said that they were very interested in listening to radio stations via the Internet."
Our M*STAR surveys so far have collected almost 9,000 responses for analysis in a preliminary report to CPB this month. The findings reflect the experience of only a few stations over a short period, but they nevertheless point to opportunity: a growing audience, the prospect of lower costs and a new stream of audience-based revenues.
Growth: At the largest stations in the project, online listenership increased steadily over the last 18 months. KPLU, a streaming leader with relatively large volume for several years, saw its streaming double (up 104 percent) since Jan. 1, 2001. In the same period, streaming at WFUV more than tripled (up 230 percent), and streaming at WXPN multiplied by a factor of five (up 429 percent).
While the streaming audience is still comparatively small by terrestrial broadcast standards, it is not insignificant. The maximum number of simultaneous connections—a kind of peak average quarter-hour audience—hovers around 2,000 for both WXPN and KPLU and around 1,500 for WFUV. So these streams are now attracting the audience of a small-city station, and they appear to be growing well beyond that size.
Streaming costs: The cost of serving these listeners has been painfully high. In early 2002, WFUV was paying about $6,700 a month; KPLU shelled out almost $10,000 a month; WXPN ran up bills of more than $12,000 a month. (Ouch!) The costs are different because the stations have not only different volumes but different rates in their contracts with bandwidth providers. Stations agree in advance to purchase a specific volume of streaming for a set fee each month. If they exceed this target, they pay "overage" charges, which can often be usurious.
However excited stations may have been about the early growth of their streams, the ever-rising costs have been equally discouraging, to say the least.
Revenue: Only one thing brightened the picture recently: with encouragement from the M*STAR project, they began to analyze membership records, and they discovered that streams were beginning to attract membership revenue.
Revenue varied widely among the participants, but the two largest stations—KPLU and WXPN—each reported more than $100,000 in stream-related member revenue in the six months between Sept. 1, 2001 and Feb. 28, 2002. In the same period, WFUV reported $28,000 in stream-related revenue. KPLU and WFUV had been tracking stream revenue in their databases, using a question on their pledge pages and making notes during drives.
Notably, these revenues are almost certainly yielding a surplus above the stations’ direct costs of streaming.
Sources of this membership revenue were completely different for KPLU and WXPN— an encouraging finding because it may indicate diverse revenue possibilities. Ninety-three percent of WXPN’s streaming revenue came from outside the Philadelphia coverage area. For KPLU, 92 percent came from inside the Seattle market. We will analyze these patterns of audience reach.
Whatever the reasons for the difference, each station has reason to be optimistic. Each has developed an online service that can generate a new revenue stream that is completely compatible with its core values and staff expertise.
Once M*STAR discovered this link between streaming and revenues, project stations unanimously endorsed continuing the research. Only this time they want to do more work on methods of converting streamies into members.
The discovery of significant stream-related membership revenue—which was completely unexpected—would have been cause enough for celebration. But, as we continued with the second phase of the project, we uncovered a second piece of financial information that may have even more benefit for more stations.
Shopping for price What we found in our comparison shopping early this year was that streaming bandwidth may be purchased at surprisingly low costs. In January and February we solicited bids for a two-year package of four streaming channels, with expected growth rates.
The range of the resulting bids was shocking: from as high as $721,580 (using Activate) to as low as $47,443 (using StreamGuys).
Naturally, we were skeptical about the quality and reliability of the low-price provider (actually, there were several low-priced bids but we focused on one, StreamGuys). How could someone offer streaming service for so much less? This had to be a mistake.
Fortunately, we had a chance to find out if the rate was really too good to be true. Encouraged by our research and their killer streaming fees, WXPN decided to move its entire streaming operation from VitalStream to StreamGuys in early March. Within a month, costs dropped from $12,000 a month to $1,800. After two months of "beta testing," WXPN saw no increase in reported streaming problems and no appreciable degradation in service. Assuming the relationship continues, the switch to StreamGuys will save WXPN $150,000 in the next 12 months, compared with their previous contract with VitalStream (assuming that streaming continues to expand at its current rate of growth).
KNBA followed suit, moving its stream service from Public Interactive to StreamGuys, lowering its monthly stream expense from $750 to $250.
Our aim in this part of the work was not to discredit one source and promote another. It was to increase competition and awareness of alternatives. As stations become aware of their options, we suspect that many others will want to revisit their streaming choices.
Of course, cost is not the only factor stations need to consider in choosing a streaming service. So in Phase 2 of M*STAR, we propose to monitor cost, quality, reliability and performance across a range of ordinary streaming providers (one stream per listener), and we propose to expand our research into "peer-to-peer" technology (one stream to multiple listeners), as offered by ChainCast, Abacast and Blue Falcon, which may offer additional cost savings.
Will wider research confirm the optimism? None of these findings is conclusive—we might even call them field observations. Both the comparative cost study and the revenue research need replication, confirmation and expansion. To that end, we have asked CPB to support additional research into both costs and revenues using more stations with a wider range of formats, including classical, jazz and news.
But even if these observations are limited, they are not isolated. Other information coming from a diverse range of sources points in the same basic direction: A streaming economy is developing with sustainable costs and an audience willing to pay for unique service.
For example, on the question of comparative costs, Virgin Free Radio, the world’s largest streaming audio provider, recently reported that it reduced costs by 75 percent when it switched to BlueFalcon, which uses peer-to-peer technology.
While the growth of XPN.org seems remarkable in isolation, MeasureCast, one of the two primary sources for Internet radio data, recently reported that overall listening to several hundred Internet radio stations increased by an average 83 percent in the first four months of 2002 and 510 percent when compared to January 2001. This comes very close to the growth rates reported by XPN.org.
Our stream-revenue numbers may seem to come out of nowhere, but they are consistent with the most recent findings in the Edison/Arbitron series reported in February: "Those who have ever listened to Internet audio (audio streamies) show a willingness to pay subscription fees for the right mix of offerings. Four in ten audio streamies say they would be willing to pay a small fee for commercial-free content, high-quality audio or content they can’t find anywhere else."
With CPB’s agreement to cover the recording industry’s copyright fees for station-based services, this combination of observations and research, insight and experience suggests that the most painful period of webcasting may be over, at least for public radio stations that are poised to gain a revenue foothold using their traditional business model of voluntary memberships.
All of this positive news may surprise some readers, for it challenges a widely circulated view of streaming as an interesting but somewhat meaningless and expensive sideshow, compared with broadcast programming. Just a few months ago, even some of the managers of M*STAR stations viewed streaming as a high-tech obligation.
Obligation or opportunity? This sense of streaming’s "unimportance" has been reinforced by the ongoing tracking study conducted by researcher George Bailey that found—as reported in Current—"Internet listeners are virtually nonexistent . . . [and] only 5 percent of public radio listeners listen online every week."
So, who’s right? Bailey, who dismisses the value of streaming, or Rosin, who sees streaming as "essential" to a radio website?
To unravel this apparent contradiction, we need to look more carefully at what our research colleagues are actually saying. The Edison/Arbitron findings are, at the core, very close to the basic facts uncovered by Bailey’s tracking study. Where Bailey found 5 percent of public radio listeners listening online each week, Edison found that 4 percent of "all Americans" listen to radio stations online at about the same time.
If the basic numbers are so similar, why does Rosin see opportunity where Bailey finds "stagnation?"
In large part, it’s a classic difference in perspective. Bailey sees a nearly empty radio audience, while Rosin finds it partially filled. The optimists can point to a gradual increase in people who have listened to online radio—up from 11 percent in 2000 and 18 percent in 2001 to 25 percent this year.
Bailey is warning broadcasters to avoid distraction from their main business.
Considering the size of web audiences, he feels there is much more to be gained by continuing to improve broadcast radio. Rosin, on the other hand, sees the Internet as a battleground on which webcasters will compete with broadcasters to control a new medium. In Rosin’s view: "If the radio industry does not do what it needs to do to co-opt this new medium, it leaves Internet audio open to the Ted Turners of the world."
Each of these perspectives has merit; neither should be considered the final word—especially in a field as new and as volatile as Internet streaming.
The best strategy, in my view, is to invest in moderate-sized pilot projects that can explore the streaming opportunities that now seem to be more promising than ever. The downside is minimal.
If we contain costs and build revenues, we can reasonably expect a good return on investment. If M*STAR only stimulates competition among bandwidth providers, the system will save hundreds of thousands of dollars annually. In this debate, we should not lose site of the reason why many stations embraced streaming in the first place: to expand their program service. If the costs come down and membership support holds up, the entire system moves closer to program "diversity through streaming." There is just no need to choose: We can acknowledge Bailey's caution and still follow Rosin’s advice, which can be found in the most recent Edison/Arbitron report.
"Consumers are using more streaming media than ever before," Rosin writes. "Our prior studies have shown that consumers would be upset if their favorite online audio content disappeared due to digital rights controversies, and they would search for other types of Internet audio to listen to in its place. Therefore, webcasters with the most compelling content and strongest brands should maintain their streaming efforts because they will be most capable of weathering the short-term obstacles and be best positioned for success when the market matures."
Maintaining and even expanding streaming efforts may be the surest way to prepare for whatever comes next in the roller-coaster world of Internet exploration.
Mark Fuerst, a public radio management consultant based in Rhinebeck, N.Y., is coordinator of the Public Radio Internet Station Alliance (PRISA) and project director of M*STAR. He previously managed WXPN in Philadelphia.
Read it in the Press: Current Magazine