Printer Friendly

How early can video revenue be accurately predicted?

Americans love videos. Last year, consumers spent $17.4 billion on videos, renting 3 billion and buying 700 million (VSDA, 2000). Predicting video revenue is critical, because it accounts for 55 percent of gross studio revenue, more than box-office, pay-per-view, and television revenue combined (VSDA, 1998). How early can video revenue be accurately predicted? Several early indicators are tested: first and second weeks' theatrical revenues, fall-off, opening screens, advertising, genre, and critics' ratings. Two models are developed to predict rental and sell-through video revenue with 86 percent accuracy on average by the second week of a movie's release.

**********

AMERICANS LOVE VIDEOS. Consumers rate video viewing as their number one favorite leisure activity (VSDA, 1998). Every week at least 30 percent of all VCR households rent a video, and nearly 70 percent rent at least once a month. The average household with children owns several videos, and many households collect them for their "video libraries." In 2000, consumers rented more than 3 billion videos and purchased more than 700 million, with continued growth expected (VSDA, 2000).

VIDEO INDUSTRY BACKGROUND

The video industry has even surpassed the movie industry. Consumer spending on videos is more than double that of consumer spending on theater tickets; consumers spent $17.4 billion on videos compared to $7.5 billion on theater tickets in 2000 (VSDA, 2000). In fact, videos accounted for 55 percent of studios' gross movie revenue, more than that from theatrical, pay-per-view, and broadcast television revenue combined (VSDA, 1998).

To satisfy consumers' seemingly insatiable appetite for videos, home-video managers at every studio, rental retailer, and discount store scramble to attempt what some say is impossible: accurately predict video revenue, sometimes up to six months in advance of release. They need to fore cast video demand in order to produce a sufficient quantity, place orders, manage inventory, and plan promotions. But such forecasting is difficult because so many different factors affect video revenue, and there are breakout hits that surprise everyone each year.

What makes predicting video revenue even more complex is that there are two different video markets: rental and sell-through. In the rental market, studios sell the videos for about $70 per tape to rental stores, who then rent to consumers for an average of $2.64 per day (VSDA, 2000). In the sell-through market, studios sell videos to discount stores such as Wal-Mart for about $12 per video. The discount stores then sell them to consumers for about $20 per video. Although 100 percent of theatrical releases become rental releases, studios choose only about 10 percent of all theatrical releases for sell-through release due to higher marketing costs and lower per-tape revenue. The VSDA (1999) estimates that studios would have to sell up to five times more videos to make a sell-through release profitable. Therefore, early forecasting of video revenue under each scenario is critical.

Attempts to forecast rental and sell-through revenue must occur several months before the planned video introduction at retail. It takes months just to complete production, packaging, and movement of the videos through the distribution channels. With the average window (time between theatrical and video release) decreasing from six months to as short as four months, there is increasing pressure on managers to predict rental and sell-through revenue as early as possible, sometimes even before cumulative theatrical revenue is known. Most videos are only "hot" for a limited time, so accurately predicting demand as early as possible is important.

Ask managers what the best indicator of video success is, and they will answer "the theatrical performance." If the movie is a hit in theaters, it is assumed that the video will also be a hit. If the movie is a dud in theaters, it is likewise assumed that video revenue will be low. During personal interviews, home video managers at several major studios and rental retailers said that they look at the following early indicators to forecast rental and sell-through revenue: (1) the opening week of theatrical revenue; (2) the "fall-off" (percentage decline in revenue between the first and second weeks of theatrical revenue); (3) the opening number of screens on which the movie is shown (a kind of "vote of confidence" from exhibitors); (4) advertising expenditures; and (5) genre.

RELATED RESEARCH

Several academic studies concur that the above early indicators can accurately predict theatrical revenues, but no study yet has tested whether these same early indicators can accurately predict video revenues, which may accrue six months to a year later. For example, one study showed that the first few weeks of theatrical revenue can accurately forecast cumulative theatrical revenue (Sawhney and Eliashberg, 1996). A second study correlated advertising expenditures to theatrical revenue (Zufryden, 1996). A third study found website activity to be significantly related to theatrical revenue (Zufryden, 2000). A fourth study showed that critics' ratings influenced theatrical revenue (Eliashberg and Shugan, 1997). However, no study to date has tried to answer how early rental and sell-through video revenue can be predicted using theatrical indicators. (For a study analyzing the optimal window for a video, see Lehmann and Weinberg, 2000.)

This paper attempts to answer the following research question: How early can video revenue be accurately predicted? Specifically, can early theatrical indicators such as opening revenue, fall-off, screens, advertising, and genre forecast rental and sell-through revenue? Null hypotheses correlating each early indicator mentioned by managers to both rental and sell-through revenue are proposed. Then, each hypothesis is empirically tested, and a comparison of each early indicator's effect on rental versus sell-through revenue is made. It is suspected that certain early indicators will be useful in predicting rental revenue, while other indicators will be useful in predicting sell-through revenue.

EARLY INDICATORS OF RENTAL VS. SELL-THROUGH REVENUE

Opening theatrical revenue

Much of the media hype surrounding a new movie focuses on its opening week. Mainstream newspapers and television shows mention opening theatrical revenue as if it conclusively determines the ultimate success or failure of the movie. Home-video executives breathe a sigh of relief if it opens well or rethink sell-through plans if it does not. Is this reliance on opening theatrical revenue warranted? Can it predict rental and sell-through video revenue several months later? These questions are restated in the following null hypothesis:

H1: The higher the opening theatrical revenue:

a) the higher the rental revenue and

b) the higher the sell-through revenue

Fall-off

The "fall-off" (also known as "drop off") is the percent decline in theatrical revenue from the first week to the second week. In the majority of cases, the fall-off is negative. A large fall-off indicates that cumulative theatrical revenue will not be high and the movie will not have longevity in theaters. If a movie opens big and has a low fall-off, managers hold their breath-a blockbuster could be in the making. A blockbuster opens big and stays big, and sometimes ends up making more than $100 million. Only about 10 movies a year accomplish this. Blockbusters are sure hits at rental and sell-through. On the other hand, if the fall-off is large, theatrical revenue is expected to be lower, and the video revenue is likewise expected to be less. Consequently, the second null hypothesis is:

H2: The higher the fall-off:

a) the lower the rental revenue and

b) the lower the sell-through revenue

Opening screens

Exhibitors such as Cineplex Odeon and Mann Theaters decide how wide to screen each film. Some movies are launched with a "wide release" (in almost every theater nationwide), while others, such as independent films, are launched with a "platform release" (perhaps initially only in New York and Los Angeles). A wide release, on the nation's more than 3,000 screens, is considered a good omen that the movie will be a hit. It also shows exhibitors' faith in the movie. The null hypothesis is:

H3: The higher the opening screens: a) the higher the rental revenue and

b) the higher the sell-through revenue

Advertising

High advertising expenditure stimulates consumer awareness and intent to see the movie. Usually, advertising starts two or three weeks before theatrical launch, so the planned advertising expenditure is known well before the movie's opening. A high advertising investment indicates a studio's commitment to the movie. The greater the advertising, the more people who become aware of the movie, and the more people will want to see it. Since not all people can get to the movie while it is still in theaters, many rent or buy it later, resulting in high rental and sell-through revenue. The null hypothesis is then:

H4: The higher the theatrical advertising:

a) the higher the rental revenue and

b) the higher the sell-through revenue

Genre

A final early indicator in predicting rental and sell-through revenue is genre. Genre is known well before the movie's opening, so revenue forecasts can be made very early. Since six out of the top-ten selling videos in 2000 were animated features aimed at children, many managers believe that genre is a good indicator of sell-through revenue. In 2000, the top-selling videos were (in order): Tarzan, Star Wars: Episode I, Toy Story 2, Stuart Little, Little Mermaid: Return to Sea, Pokemon: The First Movie, The Tigger Movie, Chicken Run, XMen, and Erin Brockovich. Tarzan alone sold 11.2 million units, representing $222 million in sell-through revenue for Buena Vista/Disney. Children's movies are expected to be big sellers because children often delight in seeing the same movie over and over, so buying rather than renting the video makes sense. In fact, 80 percent of VCR households with children have bought videos (VSDA, 2000).

In contrast, none of the top rental videos were animated features during this same period. Instead, the top 10 included a broad array of genres, including action-adventure films, comedies, and dramas. In 2000, the top-10 rental videos were (in order): The Sixth Sense, Double Jeopardy, The Green Mile, The Bone Collector, Runaway Bride, American Beauty, Three Kings, Erin Brockovich, Blue Streak, and Stir of Echoes.

Are certain genres more strongly correlated with rental versus sell-through revenue? The null hypotheses can be broken down as follows:

H5a: Children's movies are more strongly correlated with sell-through revenue than rental revenue.

H5b: Action-Adventure movies are more strongly correlated with rental revenue than sell-through revenue.

H5c: Comedies are more strongly correlated with rental than sell-through revenue.

H5d: Dramas are more strongly correlated with rental revenue than sell-through revenue.

RESEARCH METHOD

The method of analysis in this research will consist of three steps. First, in order to test the above hypotheses, the correlation matrix of each of these variables will be analyzed and a comparison of the magnitude, sign, and significance of each factor with rental versus sell-through revenue will be made. Next, stepwise regression will be employed to determine the relative importance of the aforementioned factors in addition to other factors such as critics' ratings and second week of theatrical revenue. This will help determine which early indicators are best at predicting rental revenue and which ones are best at predicting sell-through revenue. As a benchmark, each early indicator will also be examined to identify which ones are best at predicting theatrical revenue. Third, a comparison of the current industry model and the proposed model will be made to see which has higher forecasting accuracy.

Data

The data set includes all 60 movies that had a theatrical release followed by simultaneous rental and sell-through releases from 1994 through 1997. Starting in 1998, some rental and sell-through releases became staggered, so this period of simultaneous rental and sell-through release was chosen. Weekly theatrical revenue, screens, and genre were obtained from CYNESIS (NRG); weekly rental and sell-through revenues were obtained from Alexander and Associates; box-office advertising expenditures were obtained from LNA (Leading National Advertising); and critics' reviews were obtained from the "Ebert and Roeper" website (http://tvplex.go.com/buenavista/ebertandroeper/today.html). Movies that received two "thumbs up" were coded as 1; two "thumbs down" as 0, and one "thumb up" and one "thumb down" as .5.

Descriptive statistics

For the 60 movies in the sample, the mean theatrical revenue was $81.3 million, the mean rental revenue was $43.5, and the mean sell-through revenue was $86.2 million. The biggest moneymaker in theaters was Forrest Gump (1994), at $329.5 million, making it the number one movie that year. The biggest rental was also Forrest Gump, at $194.2 million, but the biggest sell-through title was Toy Story (1995) at $424.7 million. In terms of cumulative revenue over the four years of data, sell-through generated the most revenue ($5.2 billion) versus box-office ($4.9 billion) and rental ($2.6 billion). For a summary of descriptive statistics, please see Table 1.

RESULTS

In order to complete the first step of the method of analysis, the hypotheses were tested by determining the correlation between each of the early theatrical indicators and rental and sell-through revenue. Magnitude of correlation, sign, and significance of each early indicator were tested. See Table 2 for a summary of the correlations.

H1 suggested that the higher the opening theatrical revenue, the higher the rental and sell-through revenue. It was anticipated that the higher the opening box-office receipts, the higher the chances of the movie being a hit. If the movie was a hit, then consumers would be more interested in renting or buying the video. In fact, this correlation was deemed true, with a .49 correlation for rental and .55 correlation for sell-through. Therefore, the null hypothesis was accepted. Note that the correlation was more strongly positive for sell-through than rental. This means that opening theatrical revenue was a slightly better early indicator of sell-through revenue than rental revenue, though it can be used to predict both. The second step of our analysis will utilize stepwise regression to determine whether opening week is indeed the best predictor of video revenue.

H2 proposed that the greater the fall-off percent between the first and second week of theatrical revenue, the lower the rental and sell-through revenue. It was hypothesized that a dramatic decline of a movie's box-office receipts in the second week mirrors a waning of consumer interest, perhaps due to negative word-of-mouth. Therefore, it was expected that consumers would be less likely to rent or buy the movie. Surprisingly, the correlation between fall-off and rental revenue was not significant at only .10. Interestingly, the correlation between fall-off and sell-through revenue was also weak at -.03. Although it was not significant, a negative sign indicates that the higher the fail-off, the higher the sell-through revenue, the opposite of what was expected according to H2. This was the only early indicator that resulted in different signs for rental and for sell-through revenue. Different signs suggest that the same early indicator may affect rental and sell-through revenue in opposite ways. Since neither correlation was significant, H2 was rejected for both rental and sell-through revenue. This means fall-off was not a good early indicator of either rental or sell-through revenue.

H3 posited that the higher the number of opening screens, the higher the rental and sell-through revenue. It was expected that the wider the movie opened, the higher the likelihood it would be a hit and the more consumers would rent or buy the video. This turned out to be true. The correlation between opening screens and rental revenue was .32 (significant at p < .01), and the correlation between opening screens and sell-through revenue was slightly stronger at .40 (significant at p < .001). Therefore, H3 was accepted.

H4 proposed that the higher the advertising in support of the theatrical release, the higher the rental and sell-through revenue. It was expected that the higher the advertising, the greater the consumer awareness and subsequent demand for the product. The null hypothesis was accepted. The correlation between advertising and rental revenue was .58 and the correlation between advertising and sell-through revenue was almost the same, at .57. Both were significant at the highest level (p < .0001). This indicates that the millions of dollars in advertising spent by studios to spread awareness among consumers about the movie was found to be a very good early indicator of both rental and sell-through revenue. This is the only variable that is in the direct control of studios; advertising budgets are usually set weeks before launch. Studios cannot directly control the number of opening screens or opening theatrical revenue, but they can control how much they spend to support the movie. It is gratifying to get confirmation that investing in advertising is correlated with higher revenues. This finding leads to the next question: Are there different effects of advertising on theatrical, rental, and sell-through revenues?

The strong correlation between advertising and video revenue proven in H4 can be misleading if advertising is considered in a vacuum. Marketing managers often plead with senior studio management for a higher advertising budget, promising that the money will reap dividends not only from the theatrical release but also from the subsequent video releases. In effect, they reason that they can get "three bangs for the buck"; an expensive advertising campaign that raises consumer awareness is expected to increase theatrical, rental, and sell-through revenue.

Two further analyses were conducted to investigate the relationship between advertising in support of the initial theatrical movie release and its subsequent effect on rental versus sell-through video revenue. First, regression analysis was employed for three dependent variables: theatrical, rental, and sell-through revenues and only one independent variable, advertising. The results showed that while advertising had a strong effect on theatrical revenue, it had a more diluted effect on video revenue. The R-square for theatrical revenue was .54, whereas the R-squares for rental revenue and sell-through revenue were only .34 and .33, respectively. Note that advertising had about the same effect for both rental and sell-through revenue.

The second analysis showed that not all videos were alike. Advertising had a differential effect on rental versus sell-through revenue when genre is controlled. Stepwise regression including dummy variables for each genre was performed. The analysis showed that the partial R-square for advertising fell from .34 to .09 for rental revenue. The children's genre had the strongest partial R-square at .45. This means advertising explained only 9 percent of the variance in rental revenue, whereas genre explained 45 percent of the variance.

Surprisingly, the same reduction was not true for sell-through. The partial R-square for advertising remained the same for sell-through at .33, whether genre was included in the model or not. This means that advertising alone explained 33 percent of the variance in sell-through revenue. Interestingly, the children's genre was not significant for predicting sell-through revenue. The comedic genre was significant, with a partial R-square of .04.

These findings caution marketing managers from counting on the "three [equal] bangs per buck." The study results can be extrapolated to infer that introductory advertising has the strongest immediate effect on theatrical revenues, a medium effect on sell-through revenues, and a weak effect on rental revenues. An interpretation of the findings suggests a domino effect, which begins with a strong advertising push that largely affects theatrical revenues, that impacts sell-through revenues, and that to a lesser extent influences rental video performance. A marketing manager with $10 million to support all three releases would be best advised to spend the bulk of the budget to promote the first two weeks of theatrical launch rather than to apportion the budget equally among the three releases.

H5 addressed the four main genres: movies aimed at children, action-adventure films, comedies, and dramas. First, the correlations were analyzed within each genre. H5a suggested that children's movies would be more strongly correlated with sell-through revenue than rental revenue. Surprisingly, this turned out not to be supported. Children's movies were more strongly correlated with rental revenue (-.67) than sell-through revenue (insignificant at -.24). Note that the negative signs for both rental and sell-through indicates that a children's movie is expected to have both lower rental and lower sell-through revenue. Therefore, H5a was rejected.

H5b stated that action-adventure films are more strongly correlated with rental revenue than sell-through revenue. This was supported. Action-adventure films were significantly correlated with rental revenue (.31) but not significantly correlated with sell-through revenue (.19). H5c showed the same pattern. Comedies were significantly correlated with rental (.40) but not sell-through revenue (.004). Therefore, H5c was accepted. H5d also exhibited the same direction. Dramas were significantly correlated with rental (.30) but not sell-through revenue (.14). Consequently, H5d was accepted as well.

Next, the correlation between each genre was analyzed. Of all the genres, children's movies had the strongest correlation with rental revenue. None of the genres had significant correlation with sell-through revenue. So, managers' reliance on genre alone to forecast sell-through revenue may be unwarranted.

Other early indicators

If not genre alone, then what other early indicators could be used to predict rental and especially sell-through revenue? Two other early indicators may be used: the second week of theatrical revenue and critics' ratings. Opening week of theatrical revenue was a good early indicator of video revenue, but fall-off was not. Perhaps the second week of theatrical revenue would be a better indicator than fall-off, because it reveals actual magnitude of revenue rather than just percent decline. To illustrate, two movies can have the same fall-off, but one can be a blockbuster and the other can be a dud. For instance, both could have a 20 percent fall-off, but the blockbuster could open at a successful $50 million then fall to $40 million, and the dud could open at $5 million then fall to $4 million, probably still in the red. Furthermore, previous research has shown that the first three weeks of theatrical revenue are critical in forecasting cumulative theatrical revenue (Sawhney and Eliashberg, 1996).

Critics' ratings may also be good early indicators. Previous research (Eliashberg and Shugan, 1997) has found that critics' ratings are beneficial in forecasting cumulative theatrical revenue. If a movie has good critics' ratings, it would be expected to do better both at rental and at sell-through. In particular, since people like to own only good movies, strong critics' ratings are expected to be especially helpful for predicting sell-through revenue. An advantage of employing critics' ratings is that they are known well before the movie's launch.

STEPWISE REGRESSION

In order to understand the relative strength of each early indicator, stepwise regression was conducted to address the second step of the method of analysis. There were three dependent variables: theatrical, rental, and sell-through revenue. The early indicators analyzed included all those mentioned so far: first and second weeks of theatrical revenues, fall-off, opening screens, advertising, genre, and critics' ratings. Stepwise regression is a statistical analysis that determines how much each early indicator can explain in predicting theatrical, rental, and sell-through revenue. The percentage of explained variance is called "partial R-square." This procedure prioritizes all the early indicators in terms of highest partial R-square. Discussion of each dependent variable will follow. A summary of the stepwise regression results can be found in Table 3.

Theatrical revenue

From the beginning, it was suspected that different early indicators would be good at predicting different types of revenue. In other words, the same early indicators were not expected to be equally good at predicting theatrical, rental, and sell-through revenue. The stepwise regression results bore this out. The top four early indicators for forecasting theatrical revenue were the following (with partial R-squares reported in parentheses): the second week of theatrical revenue (.88), advertising (.01), critics' ratings (.005), and comedy genre (.004). The total explained variance was 90 percent. This means that 90 percent of all variance in theatrical revenue was explained by these four factors. in fact, 88 percent of the variance was explained by the second week of theatrical revenue alone. Note that opening theatrical revenue was not significant and was dropped from the model. As suggested in the hypothetical example above, the second week of theatrical revenue was even more important than opening week of theatrical revenue. This is truly surprising given the media's fascination with opening box-office receipts.

Rental revenue

The top four early indicators of rental revenue were: the second week of theatrical revenue (.47), opening week of theatrical revenue (.11), critics' ratings (.07), and children's genre (.05). The total explained variance of the four early indicators was 70 percent. Once again, the second week of theatrical revenue was the most important and alone explained 47 percent of the variance in rental revenue. Note, however, that this was only about half the explained variance of theatrical revenue. Also, opening week of theatrical revenue was significant in predicting rental revenue but not theatrical revenue. As suspected, different factors were found to be important in predicting rental revenue than in predicting theatrical revenue. Consequently, the same model for predicting theatrical revenue should not be used for predicting rental revenue.

Sell-through revenue

The top four early indicators of sell-through revenue were: the second week of theatrical revenue (.40), the action-adventure genre (.10), the comedy genre (.11), and critics' ratings (.04). The total explained variance was 65 percent. Similar to forecasting rental revenue, the second week of theatrical revenue was the most important early indicator in forecasting sell-through revenue. Also similar to rental, critics' ratings were a significant early indicator of sell-through revenue. However, that is where the similarities ended. While the children's genre was the only significant one for rental, both action-adventure and comedy genres were significant for sell-through. Therefore, the study concludes that the same models cannot be used for predicting rental and sell-through revenue, even though both are videos. Certain factors drive rental, and other factors drive sell-through.

Comparison of models

The third and final step of the method of analysis used in this study involved conducting a comparison of models. The industry model based upon early indicators gleaned from managerial interviews was compared to the proposed model using the early indicators determined from the stepwise regression outlined above. The industry model utilized the same four early indicators as suggested in the null hypotheses to predict both rental and sell-through revenue: opening theatrical revenue, opening screens, fall-off, and advertising. The proposed rental model utilized the second week of theatrical revenue, opening theatrical revenue, critics' ratings, and the children's genre. The proposed sell-through model utilized the second week of theatrical revenue, the action-adventure genre, the comedy genre, and critics' ratings. The proposed models follow the suggestion by previous academic research to log-transform the dependent variable when forecasting nonlinear revenue data. Both models were compared in terms of adjusted R-square and MAPE (mean average percent error). See Table 4 for a summary of the comparison of models.

Industry model

The industry model based upon the early indicators that managers reported using to predict video revenue yielded only .42 adjusted R-square for rental and only .38 adjusted R-square for sell-through revenue. This means that only 42 percent of the variance in rental revenue, and only 38 percent of the variance in sell-through revenue, were explained by these four early indicators. Even worse, the MAPE for rental was .85 and .81 for sell-through, which means forecasts using the industry model have only a 15 percent accuracy rate for rental and a 19 percent accuracy rate for sell-through. There is room for substantial improvement.

The proposed rental model yielded a .60 adjusted R-square for rental and a .62 adjusted R-square for sell-through. This means that 60 percent of the variance in rental revenue was explained by the four early indicators and 62 percent of the variance in sell-through revenue was explained by the other four early indicators identified in the stepwise regression. After completing the log transformation, the proposed models had a MAPE of .14 for rental and .14 for sell-through. In other words, the proposed models had an 86 percent accuracy rate. This is a substantial improvement and a very high accuracy rate given that the prediction can be made as soon as the second week of theatrical release, up to six months in advance of the video introduction.

SUMMARY

From the comparison of the industry model and the proposed models, results conclude that one model cannot predict both rental and sell-through revenue; certain early indicators are better at forecasting rental versus sell-through revenue. Specifically, reliance upon opening week of theatrical revenue is not warranted in forecasting video revenue; the second week of theatrical revenue is a better indicator. Finally, advertising, while correlated with both rental and sell-through revenue, is not necessarily as good an indicator as genre, if managers use different models for predicting rental and sell-through revenue, accurate forecasts can be made as early as two weeks after the theatrical launch and up to six months before the video release.

LIMITATIONS AND FUTURE RESEARCH

This study, as all studies, suffers from limitations. These limitations may seem more salient because the movie industry is rapidly changing. Due to increased competition and shaved margins, studios, rental retailers, and discount stores are constantly experimenting with new techniques designed to increase revenue and profits. In just a few years, new strategies involving sequential video releases, revenue sharing, bargain-bin buying, and new technologies have emerged. These trends are topics ripe for future research.

First of all, the data set collected in this study included only those movies that had a simultaneous rental and sell-through video release. During the time period of the data set, this was the norm. However, recently, studios have begun to stagger video releases, first selling to the rental market, then a few months later selling to the sell-through market. In this way, they can charge rental retailers a higher price (about $70) while demand is high, then sell to mass merchants at a lower price (about $12) after rental demand fades. Therefore, a topic for future research would be to update the data set and utilize both early theatrical and early rental indicators to forecast ultimate sell-through revenue. Is cumulative revenue maximized with a sequential or simultaneous rental and sell-through release? What are the optimal windows between theatrical and rental release followed by sell-through release?

A second limitation is that the data do not address revenue sharing. Once again, during the time of data collection, revenue sharing was only an experiment. Today, however, it is the norm for large rental retailers such as Blockbuster. In revenue sharing, studios lease videos to rental retailers at a reduced price then split the profits as the video "turns" (rents out to consumers). Revenue-sharing incents retailers with "bonus copies" if they purchase a minimum number of videos (VSDA, 2000). Future research could analyze the profitability of this new arrangement. Does revenue sharing increase profits from both the studio's and rental retailer's perspectives?

A third limitation is that the study does not address "bargain-bin buying." In bargain-bin buying, rental retailers sell "used" (previously rented) videos for about $5 (Arnold, 1999). This may cannibalize sell-through revenue, as used rentals are snatched up by consumers directly from rental retailers. During the time of data collection, this was not a common strategy. Bargain-bin buying has been a recent result of revenue sharing, as rental retailers have begun to buy deep and as numerous bonus copies take up inventory and shelf space. In the past year, revenue of previously viewed tapes has increased. Now, 6 percent of the average rental retailer's revenue comes from selling previously viewed tapes (VSDA, 2000). Future research investigating the impact of the used market on video revenue would be fruitful.

New technologies often can change the face of an industry. The exponential growth of the internet has facilitated online buying of videos to such an extent that both discount stores and rental retailers selling used tapes are threatened. Consumers can purchase videos over the internet from a host of websites, such as: Amazon.com, BarnesandNoble.com, Reel.com, eBay, etc. Furthermore, sales of DVDs (digital videodiscs) are expected to more than double this year, to about 200 million discs. About 10 percent of both the rental and sell-through markets are now DVDs (VSDA, 2000). In addition, VOD (video-on-demand) and DBS (direct broadcast satellite) offer consumers new formats for entertainment. How to manage these multiple formats and capitalize on new technology are ripe areas for future research.

The author would like to thank Gerard Tellis, Ph.D., for his useful insights and encouragement.

REFERENCES

ARNOLD, THOMAS K. "Video Buyers, Prepare To Rent." Los Angeles Times, May 19, 1999.

ELIASHBERG, JEHOSHUA, and STEVEN M. SHUGAN. "Film Critics: Influencers or Indicators?" Journal of Marketing 61, 2 (1997): 68-78.

LEHMANN, DONALD R., and CHARLES B. WEINBERG. "Revenue Through Sequential Distribution Channels: An Application to Movies and Videos." Journal of Marketing 64, 3 (2000): 13-33.

SAWHNEY, MOHANBIR S., and JEHOSHUA ELIASHBERG. "A Parsimonious Model for Forecasting Gross Box-Office Revenues of Motion Pictures." Marketing Science 15, 2 (1996): 113-31.

VIDEO SOFTWARE DEALERS ASSOCIATION (VSDA). "An Annual Report on the Home Video Market." Encino, CA, 1998.

--. "An Annual Report on the Home Video Market." Encino, CA. 1999.

--. "An Annual Report on the Home Video Market." Encino, CA, 2000.

ZUFRYDEN, FRED S., "Linking Advertising to Box-Office Performance of New Film Releases--A Marketing Planning Model" Journal of Advertising Research 36, 4 (1996): 29-41.

--. "New Film Website Promotion and Box-Office Performance" Journal of Advertising," Research 40, 1 (2000): 55-64.

ELISE K, PROSSER is assistant professor of marketing in the School of Business at the University of San Diego. She earned a Ph.D. in marketing from the University of Southern California. Her research interests include entertainment marketing, brand strategy, and new product development. Her work has been published in the Journal of the Academy of Marketing Science and has been presented at the Entertainment Marketing and American Marketing Association conferences. She has served as consultant to various marketing firms.

ELISE K. PROSSER

University of San Diego

eprosser@sandiego.edu
COPYRIGHT 2002 World Advertising Research Center Ltd.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2002 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Prosser, Elise K.
Publication:Journal of Advertising Research
Geographic Code:1USA
Date:Mar 1, 2002
Words:5609
Previous Article:Attitude toward the site II: new information.
Next Article:Reaching the male consumer by way of daytime TV soap operas.
Topics:

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters