1 Introduction

The motion-picture industry represents an area of entertainment marketing and cultural consumption in which commonsense expectations often lose ground to problematic data and unexpected empirical findings. One such case in point concerns the fretted efforts by numerous researchers to establish a meaningful relationship between artistic excellence as measured by various evaluative judgments and popular appeal as gauged by various aspects of commercial acceptance. Most such studies have reported at best significant-but-weak associations between the favorability of journalistic reviews or other evaluative judgments assumed to reflect artistic quality and theatrical box-office receipts or other indicators of popular appeal.Footnote 1 Thus, recent reviews of these and other studies have concluded that the relationship between evaluative judgments of excellence and aspects of popular appeal in the case of motion pictures tends to be significant-but-weak, typically accounting for less than 10% of the variance in popularity or market performance (Holbrook 2005; Holbrook and Addis 2007). Indeed, after reviewing the role of critics in the cultural industry, Cameron (1995) calls for “more regression studies of the influence of review ratings on sales” (p. 330).

Counterposed against such empirical realities, we find the impressionistic intuitions of many educated observers who attend and appreciate films or other cultural events regarded as excellent in their aesthetic merit and artistic integrity. Such enlightened spectators may naturally assume that members of the mass audience also tend to pursue artistic quality in their consumption choices. Indeed, scholars sometimes react with impatience or indignation to anyone who would claim otherwise—no matter how insistently the empirical evidence questions their assumption.

Unfortunately, theory—whether drawn from cultural studies, psychology, sociology, aesthetics, the philosophy of art, marketing research, or any other branch of the social sciences or humanities—offers little promise of resolving the puzzles that have emerged from the confrontation between empirical observations and commonsense intuitions. Indeed, theoretical contributions to this area of inquiry often devolve into heated debates—the printed equivalent of shouting matches—between those who argue that ordinary consumers tend to like what’s good (thereby evincing an essentially subjective canon of aesthetic standards) and those who assert that artistic excellence inhibits the chances for popularity in the mass market (premised on the exigencies of a commercialistic appeal to the lowest common denominator). The former brand the latter as elitist; the latter reciprocate by accusing the former of selling out—all without producing testable theory, much less solid evidence, to support these competing claims.Footnote 2

In our view, one reason for this unsatisfactory situation stems from multiple answers to the question of what constitutes success in cultural industries in general or in the movie business in particular. Some concepts of success hinge on artistic merit or industry acclaim (say, in the form of Oscars or other awards) while others hinge on commercial acceptance (say, in the form of box-office or video-rental revenues)—often with inadequate attention to the differences between the two. Indeed, in the case of motion pictures, we shall argue that these aspects of success are separable phenomena that reflect two independent dimensions of industry recognition and market performance.

1.1 A new approach

Toward the goal of clarifying such issues as they apply with special relevance to the case of films, the present article offers a conceptualization of the problematic relationship between judged cinematic excellence and popular appeal that we refer to as a Two-Path Model of Motion-Picture Success. Previous studies have explained the weak relationship between expert judgment (evaluations of excellence by reviewers) and popular appeal (word-of-mouth or market performance) by the mediating role of ordinary evaluation (ratings of excellence by mass-market consumers) as an intervening variable that weakens the overall correlation between expert judgment and popular appeal (Holbrook 2005; Holbrook and Addis 2007; Holbrook et al. 2006). By contrast, the present approach regards evaluation and buzz as essentially independent phenomena. Thus, the hallmark of the new approach is that it distinguishes between two separable aspects of a film’s success—namely, critical and popular evaluation (i.e., the appraisal of excellence by film reviewers and/or by ordinary consumers themselves) versus critical and popular buzz (i.e., the tendency of reviewers and/or audience members to praise a film they like by recommending it to others).

In other words, for consumers or reviewers, thinking that a film is good (favorable evaluation of its cinematic excellence) and enjoying it to the point where one recommends it to others (buzz via word of mouth, click of mouse, or other forms of communication) are not the same thing. Specifically, the present study focuses on how a film’s market-directed effort or marketing clout (production budget, number of opening screens, opening box office) influences different aspects of (1) critical and popular evaluation leading to industry recognition and (2) critical and popular buzz leading to market performance. With support from the data at hand, we shall argue that favorable evaluation of a movie’s quality versus buzz via interpersonal communication work as separate routes toward the film’s success as gauged by industry recognition versus market performance, respectively.

In this spirit, we shall propose a tentative two-path model and shall then use this general viewpoint as the basis for examining data representing the success of the 190 motion pictures released in America during the year 2003. To anticipate our key findings, this approach will support a Two-Path Model of Motion-Picture Success in which (1) Industry Recognition depends on Critical and Popular Evaluation with an additional contribution from Marketing Clout and in which (2) Market Performance depends on Critical and Popular Buzz with an additional contribution from Marketing Clout.

1.2 The Two-Path Model of Motion-Picture Success

The proposed Two-Path Model of Motion-Picture Success embraces the following key variables: Marketing Clout or MktgClout (production budget, early marketing effort, and initial market success); Critical and Popular Evaluation or CritPopEval (evaluations of excellence by film reviewers and ordinary consumers); Industry Recognition or IndRecog (Oscars and other awards from the motion-picture industry); Critical and Popular Buzz or CritPopBuzz (praise or recommendations of a film by reviewers and audience members); Market Performance or MktPerf (overall market earnings of a film).

The relevant two-path model is estimated by ordinary least-squares regression (OLS), as represented by the following equations.

$$ \begin{aligned}{} {\text{Path}}\,1:\, & {\text{CritPopEval}} = \beta _{{{\text{CPE,MC}}}} {\text{MktgClout}} + {\text{Error}} \\ & {\text{IndRecog}} = \beta _{{{\text{IR,CPE}}{\text{.MC}}}} {\text{CritPopEval}} + \beta _{{{\text{IR,MC}}{\text{.CPE}}}} {\text{MktgClout}} + {\text{Error}} \\ \end{aligned} $$
$$ \begin{aligned}{} {\text{Path}}\,2:\, & {\text{CritPopBuzz}} = \beta _{{{\text{CPB,MC}}}} {\text{MktgClout}} + {\text{Error}} \\ & {\text{MktPerf}} = \beta _{{{\text{MP,CPB}}{\text{.MC}}}} {\text{CritPopBuzz}} + \beta _{{{\text{MP,MC}}{\text{.CPB}}}} {\text{MktgClout}} + {\text{Error}} \\ \end{aligned} $$

Here, each construct is represented by a factor score derived from principal components analysis of several relevant variables (described later). By definition, such factor scores are standard normal deviates (mean = 0.0, SD = 1.0) so that OLS regression constants are all identically zero with all regression coefficients equivalent to beta weights (β).

In general—as an overview with key details to be filled in later when discussing our method and measures—the conceptual rationale for the model proceeds as follows.

  1. (1)

    We assume that—as a reflection of various design characteristics (e.g., drama or comedy versus action or adventure genre; shorter versus longer length; English versus foreign language), key participants (unknown versus famous actors, obscure versus celebrated directors), noteworthy features (routine versus distinguished script, drab versus lavish costumes, familiar versus exotic scenery, ordinary versus brilliant cinematography, mundane versus spectacular special effects), MPAA ratings (G or PG versus PG-13 or R), and comparable variables not explicitly modeled in the present analysis but often considered in the literature on motion-picture studiesFootnote 3—the marketing strategy for a motion-picture embodies a higher or lower level of Marketing Clout (MktgClout). That is, the film’s degree of marketing effort—as manifested, for example, by the size of its production budget and by the intensity of its theatrical distribution on the opening weekend—represents its positioning as a mass-marketed blockbuster (large budget, big opening on numerous screens) as opposed to an art-, cult-, or niche-market film (small budget, restricted opening on few screens). In recognition of this distinction between wide-release and slow-build movies, the inclusion of such marketing-clout variables as control measures is standard practice among motion-picture researchers.Footnote 4

  2. (2)

    We further assume that the film’s Marketing Clout (MktgClout) contributes to its Critical and Popular Evaluation (CritPopEval) in the sense that ratings of excellence by reviewers and consumers are likely to reflect a negative influence of blockbuster-related marketing effort. In other words, reviewers and audience members are likely to associate intense marketing effort with an absence of artistic quality so that β CPE,MC < 0.0.

  3. (3)

    We also assume that Industry Recognition (IndRecog) in the form of Oscar nominations, other awards, and related honors reflects Critical and Popular Evaluation (CritPopEval) with a further positive contribution from Marketing Clout (MktgClout). Here, extending the previous literature on motion-picture studies, we refer not only to Academy Awards—that is, Oscar nominations and winsFootnote 5—but also to awards from other organizations and sources.

  4. (4)

    In contrast to previous studies that have treated evaluation and buzz as covarying interdependent aspects of success (Holbrook 1999, 2005; Holbrook and Addis 2007) but in accord with the central premise of our two-path conceptualization, we assume that—by taking advantage of multiple indicators (described later)—Critical and Popular Evaluation (CritPopEval) along with Critical and Popular Buzz (CritPopBuzz) can be measured as factor scores in a manner that preserves their statistical independence as the basis for our two-path perspective.

  5. (5)

    We then assume that—by contrast with Critical and Popular Evaluation (CritPopEval)—Critical and Popular Buzz (CritPopBuzz) reflects a positive influence of Marketing Clout (MktgClout) so that β CPB,MC > 0.0.

  6. (6)

    Finally—consistent with the studies cited heretofore–we assume that Critical and Popular Buzz (CritPopBuzz) and Marketing Clout (MktgClout) combine to stimulate Market Performance (MktPerf) in the form of gross box office and video rentals. In this connection, previous research has examined the relationship between opening-week box office and opening-week video rentals per store (r = .65, Lehmann and Weinberg 2000) or between DVD sales and various film-related predictors (Luan and Sudhir 2005). However—to the best of our knowledge, though they represent a key aspect of market performance—video rentals, in the context of buzz-related determinants, have not previously attracted the attention of motion-picture researchers.

1.3 Preview

Using terminology consistent with the present construct definitions, previous studies of evaluative judgment and popular appeal have generally taken a form in which Popular Evaluation or PopEval (ratings of excellence by ordinary consumers) intervenes between Critical Evaluation or CritEval (ratings of excellence by reviewers) and Popular Buzz or PopBuzz (appeal to consumers or market performance), as follows (Holbrook 2005; Holbrook and Addis 2007; Holbrook et al. 2006):

$$ {\text{PopEval}} = \beta _{{{\text{PE,CE}}}} {\text{CritEval}} + {\text{Error}} $$
$$ {\text{PopBuzz}} = \beta _{{{\text{PB,PE}}{\text{.CE}}}} {\text{PopEval}} + \beta _{{{\text{PB,CE}}{\text{.PE}}}} {\text{CritEval}} + {\text{Error}} $$

By contrast, using the constructs developed earlier, the present approach defines and measures Evaluation and Buzz as conceptually distinct and statistically uncorrelated phenomena that work along separate paths to influence similarly independent Industry Recognition and Market Performance, respectively, as follows:

$$ \begin{aligned}{} {\text{Path}}\,1:\, & {\text{CritPopEval}} = \beta _{{{\text{CPE,MC}}}} {\text{MktgClout}} + {\text{Error}} \\ & {\text{IndRecog}} = \beta _{{{\text{IR,CPE}}{\text{.MC}}}} {\text{CritPopEval}} + \beta _{{{\text{IR,MC}}{\text{.CPE}}}} {\text{MktgClout}} + {\text{Error}} \\ \end{aligned} $$
$$ \begin{aligned}{} {\text{Path}}\,2:\, & {\text{CritPopBuzz}} = \beta _{{{\text{CPB,MC}}}} {\text{MktgClout}} + {\text{Error}} \\ & {\text{MktPerf}} = \beta _{{{\text{MP,CPB}}{\text{.MC}}}} {\text{CritPopBuzz}} + \beta _{{{\text{MP,MC}}{\text{.CPB}}}} {\text{MktgClout}} + {\text{Error}} \\ \end{aligned} $$

The report that follows chronicles our efforts to develop such a Two-Path Model of Motion-Picture Success via an exploration of the data provided by several sources for the 190 films released in America during the year 2003, as listed by Worldwide Box Office at www.worldwideboxoffice.com. Other aspects of these films have been examined previously in the manner just indicated (Holbrook and Addis 2007), but the present study investigates a new set of constructs and measures in order to develop and test the new two-path conceptual model just described. To preview briefly, our analysis will suggest that reviewer-and-consumer evaluations and buzz respond differently to a film’s marketing clout and that these audience responses contribute independently to a film’s industry recognition and market performance along two separable paths.

Toward that end, we shall begin with an account of measures for the relevant variables of interest, including further comments on the rationale for our two-path model. We shall then turn to the results of an analysis that tests the two-path conceptualization. Finally, we shall end with a discussion of relevant conclusions, implications for the film industry, limitations, caveats, and directions for future research.

2 Method

As detailed in what follows, our measures of the constructs just described came from a number of different industry sources available on the World Wide Web. We shall define each measure and provide a brief rationale where appropriate.

2.1 Marketing Clout (MktgClout)

Preliminary analysis of the data indicated strong inter-correlations among three variables obtained from www.imdb.com and associated with the blockbuster-related aspects of Marketing Clout (MktgClout) early in the life of a film—namely, Production Budget (ProdBudget), Number of Opening Screens (NumOpenScreens), and Opening Box Office (OpenBoxOff). Because the distributions for these three variables were strongly skewed—with very small numbers of very high values—we applied logarithmic transformations to obtain three indicants of MktgClout: LnProdBudget, LnNumOpenScreens, and LnOpenBoxOff. Here, LnProdBudget for 48 missing values was estimated by regressing LnProdBudget on LnNumOpenScreens and LnOpenBoxOff (R 2 = .523, p < .001). Inter-correlations among the three items were quite strong (all with p < .001): r LnPB,LnNOS = .753; r LnPB,LnOBO = .793; and r LnNOS,LnOBO = .951. Exploratory factor analysis of the three marketing-clout variables showed that one principal component with an eigenvalue of 2.668 accounted for 88.924% of the variance and that loadings on this component all exceeded .90—namely, .896 (LnProdBudget), .959 (LnNumOpenScreens), and .973 (LnOpenBoxOff), respectively. Accordingly, as represented by factor scores on this principal component (mean = 0.0, SD = 1.0) and consistent with the contrast between a wide-release versus a slow-build strategy, MktgClout captures aspects of a motion picture that reflect its degree of blockbuster status as a big-budget movie with high opening-week box-office receipts on a large number of screens.

2.2 Critical and Popular Evaluation (CritPopEval) versus Critical and Popular Buzz (CritPopBuzz)

As defined in the present study to represent assessments of excellence by ordinary consumers and journalistic reviewers, Critical and Popular Evaluation (CritPopEval) has only recently begun to receive the attention we believe it deserves in research on motion pictures (Holbrook 1999, 2005; Holbrook and Addis 2007) and other cultural offerings (Holbrook et al. 2006). Meanwhile, the importance of Critical and Popular Buzz (CritPopBuzz)—with respect to the diffusion of innovations, the spread of ideas, and the evolution of cultural phenomena—has been discussed for decades under such headings as the adoption process (Rogers 1962), personal influence (Robertson 1971), crossing the chasm (Moore 1991), the tipping point (Gladwell 2000), aggregated buzz (Rosen 2000), the ideavirus (Godin 2001), memes (Dawkins 1989), and the memetics of memeplexes (Blackmore 1999). Here, we measure CritPopEval by the favorability of reviewer- and consumer-based excellence ratings; by contrast, we assess CritPopBuzz as the degree to which people communicate amongst themselves about a particular film via media coverage and click-of-mouse Internet messages. In this spirit, as indices of such communications-related phenomena, our measures of CritPopEval and CritPopBuzz come from www.imdb.com, www.movies.yahoo.com, and www.rottentomatoes.com.

Specifically, overall evaluations of each film are rated by imdb.com users on a ten-point scale of excellence from 1 (awful) to 10 (excellent). IMDb provides the average rating of each film (IMDbUsersRating) and the number of responses on which this average is based (NumIMDbUsers). For our sample of 190 films from 2003, NumIMDbUsers ranges from 162 to 76,740 with a mean of 6,655.43. Because—here and elsewhere in the present section—the distributions of the numbers of responses were highly skewed, we took the natural logarithms of these numbers as our key measures in all cases (LnNumIMDbUsers and so forth).

Similar scores for consumer evaluations of each film appear at movies.yahoo.com. Here, Yahoo! reports the mean overall grades provided by movies.yahoo.com users on a 13-point scale of excellence (YahUsersRating)—ranging from F (All-time worst! = 1) to A+ (Oscar-worthy = 13)—along with the number of responses on which these average grades are based in each case (NumYahUsers). For our sample, after coding two missing entries as one-half the lowest value, NumYahUsers ranges from 4.5 to 97,481 with a mean of 5,867.095. As previously mentioned, we worked with the logarithmic transformation, LnNumYahUsers. The two missing values for YahUsersRating were estimated by a regression on the IMDbUsersRating scores (R 2 = .597, p < .001).

Consistent with the observation that journalistic film reviewers cater to the tastes of their audiences by anticipating and thereby reflecting the preferences of ordinary consumers (Eliashberg and Shugan 1997), movies.yahoo.com also provides a mean score of evaluations by various film reviewers from over 16 news publications. The review of each film from each source is assigned a rating on the aforementioned 13-position scale from F (All-time worst! = 1) to A+ (Oscar-worthy = 13). Yahoo! reports an overall average reviewer rating (YahRevsRating) and the number of reviews on which this average is based (NumYahRevs). For six missing values, we set NumYahRevs at one-half its lowest value and estimated YahRevsRating by regressing these scores on those discussed next from rottentomatoes.com (R 2 = .825, p < .001). In our sample, NumYahRevs ranges from 2.5 to 15 with a mean of 11.984. Again, we worked with the logarithmic transformation, LnNumYahRevs.

Finally, from www.rottentomatoes.com, we compiled scores for the average reviewer rating on a scale of excellence from 1 to 10 (RotTomRevsRating) and for the number of reviewers involved (NumRotTomRevs). In our sample, NumRotTomRevs ranges from 4 to 228 with a mean of 128.184 and is represented by its logarithmic transformation, LnNumRotTomRevs.

Here—consistent with work such as that by Rosen (2000), Godin (2001), and Blackmore (1999)—it appears clear that in each case the number of responses represents degree of attention, word of mouth, or click of mouse in a manner that indicates the level of critical and popular buzz generated by a particular film among ordinary consumers and from journalistic reviewers who reflect audience tastes. This approach operationalizes critical and popular buzz in a way that may offer an improvement over other approaches that have appeared in the literature. For example, Elberse and Eliashberg (2003) measure word of mouth by a film’s revenues-per-screen. Though they report that this measure is common in the industry, it seems to require fairly strong assumptions about the link between revenues and buzz. Our IMDb-, Yahoo!-, and RottenTomatoes-based data actually appear to provide a more direct measure. Further, our timing perspective differs from that adopted by Elberse and Eliashberg (2003) or others. Specifically, we look at the whole year (via data assembled primarily during the first week of October 2004), whereas other authors focus on weekly changes in the market. Beyond that, as Godes and Mayzlin (2004) suggest, our research design considers both the volume and the ratings of reviewer and consumer responses. Indeed, CritPopBuzz refers to the “volume” of media coverage, word of mouth, or click of mouse; while CritPopEval corresponds to the “ratings” by media, mouth, or mouse. Hence, going beyond previous studies, we cover both these important aspects of the relevant phenomena.

An interesting and important issue of interpretation surrounds the sorts of measures just described. Specifically—as implicitly true for all word of mouth or other recommendations, but too seldom considered explicitly—a question arises as to whether the relevant buzz is necessarily favorable in direction. We assume that consumers or reviewers tend to communicate with their reference groups about films that they particularly enjoy or like. But it might also happen that they communicate extensively about movies to which they react in an especially negative manner—perhaps, to discourage others from seeing something that they have found displeasing (Rosen 2000, p. 40). We checked this possibility for all four of the measures just mentioned (IMDb users, Yahoo! users, Yahoo! reviewers, and RottenTomatoes reviewers) and found that—in each case—the number of consumers or reviewers rating a movie is lower rather than higher for films that are more unfavorably evaluated. Hence, in the present case, the logical possibility that buzz might reflect negative evaluations does not stand up to empirical scrutiny.

In general, as one would expect, the four correlations between the parallel measures of evaluation and buzz are quite strong (all with p < .001): r IUR,YUR = .773; r YRR,RTRR = .911; r LnNIU,LnNYU = .643; and r LnNYR,LnNRTR = .622. In addition, the eight correlations between parallel measures for reviewers and consumers are also generally quite substantial (all with p < .001): r IUR,YRR = .812; r IUR,RTRR = .848; r YUR,YRR = .573; r YUR,RTRR = .615; r LnNIU,LnNYR = .489; r LnNIU,LnNRTR = .669; r LnNYU,LnNYR = .431; and r LnNYU,LnNRTR = .440. By contrast, all 16 correlations between aspects of evaluation (the various ratings) and aspects of buzz (the logarithms of numbers of raters) are far weaker—namely, less than .400 in all cases, with a median at .179.

All this suggests that the two types of measures—mean overall ratings and numbers of responses—can be regarded as assessments of separable phenomena involving critical-and-popular evaluation and buzz. Again, we note that evaluating a film favorably (a positive assessment of excellence) is not the same thing as recommending it to others for its enjoyability (word of mouth or buzz). Put differently and recalling the art-versus-commerce theme, reviewers or consumers might think that a motion picture is excellent (high artistic merit consistent with industry recognition) without recommending it to others in search of something to enjoy (viewing pleasure consistent with market performance). To repeat, in adopting this viewpoint, we emphasize that journalistic film reviewers are commonly regarded as giving evaluations that reflect the popular tastes of their mass audiences (Eliashberg and Shugan 1997). That is, as reflected by the high reviewer-user correlations just listed, journalistic reviewers of the type represented by the Yahoo! and RottenTomatoes data appear to anticipate what films their readers will favor or disfavor and then to react accordingly so that, in essence, critical evaluations serve as proxies for popular evaluations while numbers of reviews mirror the level of popular buzz.

Given this pattern of inter-correlations—in which reviewer- and consumer-based scores are strongly related within ratings and within logarithms of numbers of ratings, but not between ratings and numbers—it made sense to construct statistically independent measures of CritPopEval and CritPopBuzz by computing the factor scores from a principal components analysis with varimax rotation. In this principal components analysis of the eight measures just described, two factors with eigenvalues of 3.805 and 2.228 accounted for 75.407% of the variance, with a clearly interpretable pattern of varimax-rotated loadings. Specifically, one factor represents CritPopEval with strong loadings for IMDbUsersRating (.938), YahUsersRating (.772), YahRevsRating (.930), and RotTomRevsRating (.940). The other factor represents CritPopBuzz with strong loadings for LnNumIMDbUsers (.853), LnNumYahUsers (.809), LnNumYahRevs (.762), and LnNumRotTomRevs (.795). The eight cross-loadings of ratings and Ln(numbers) on the opposite buzz and evaluation factors were less than .300 in all cases, with a median at .126.

Factor scores for these two factors (mean = 0.0, SD = 1.0) are, by definition, uncorrelated at r CPE,CPB = .000. The relevant factor scores are therefore taken as statistically independent measures for CritPopEval and CritPopBuzz along the two separable routes to success-related outcomes represented by our Two-Path Model of Motion-Picture Success.

2.3 Industry Recognition (IndRecog) and Market Performance (MktPerf)

As aspects of Industry Recognition (IndRecog), IMDb presents separate tallies for the numbers of Oscar awards won, Oscar nominations (not including Oscar winners), other awards won, and other nominations (not including other winners). Unfortunately, the latter two categories include wins and nominations in certain “competitions” that might be considered derogatory rather than laudatory in tone—that is, the Razzie awards for worst films or worst performances (in a variety of categories) and the Golden Trailer awards for the most tasteless or most misleading promotional previews (dubbed Trashiest and Golden Fleece, respectively). Because our purpose was to represent favorable rather than unfavorable industry recognition, we subtracted the negative awards from the relevant other-award categories and then summed the remaining numbers of Oscar-and-other wins-and-nominations to derive overall measures of nominations for Oscars and for Other Awards (OthAwards). Including additional credit for wins above and beyond nominations did not improve predictive fits (cf. Deuchert et al. 2005). In sum, as gauged by nominations for favorable industry recognition, Oscars and OthAwards reflect the extent to which members of the motion-picture industry and other interested parties appreciate the excellence of a film’s script, acting, direction, special effects, music, costumes, sound recording, cinematography, and other virtues. As would be expected, the correlation between these two measures is quite strong: r O,OA = .783 (p < .001).

Meanwhile, both www.imdb.com and www.worldboxoffice.com provide information on each film’s Gross Domestic Box Office (GrossBoxOff). Further, www.imdb.com also provides data on Opening Domestic Video Rentals (OpenVidRent) and Gross Domestic Video Rentals (GrossVidRent). These measures represent the video-rental revenues achieved by a motion picture’s videocassettes and DVDs in their first week of U.S. release (OpenVidRent) and at the time when the film drops from the top-fifty list (GrossVidRent). Thirteen missing values were coded as one-half the smallest values for OpenVidRent and GrossVidRent. As before, due to skewed distributions, all three market-performance variables were represented by logarithmic transformations. As would be expected, their inter-correlations were quite strong (all with p < .001): r LnGBO,LnOVR = .818; r LnGBO,LnGVR = .777; and r LnOVR,LnGVR = .976. (The six inter-correlations between industry-recognition and market-performance variables were all less than .250, with a median of .112.)

In order to construct Industry Recognition (IndRecog) versus Market Performance (MktPerf) as conceptually distinct and statistically independent measures of artistic versus commercial success, we performed principal components analysis on the five success-related variables just described. Two factors with eigenvalues of 2.830 and 1.684 explained 90.273% of the variance in the success variables. Factor scores for the two resulting varimax-rotated dimensions clearly represented Industry Recognition (IndRecog) with large loadings for Oscars (.939) and OthAwards (.943) versus Market Performance (MktPerf) with large loadings for LnGrossBoxOff (.893), LnOpenVidRent (.982), and LnGrossVidRent (.969). The five cross-loadings were all less than .180, with a median at .070.

3 Results

3.1 Descriptive statistics and inter-correlations

Descriptive statistics for all variables and constructs mentioned in the preceding discussion of our method appear in Table 1. Correlations among all constructs included in the two-path model appear in Table 2.

Table 1 Descriptive statistics for all variables involved in the Two-Path Model of Motion-Picture Success (all with N = 190)
Full size table
Table 2 correlations (p-levels) among constructs included in the Two-Path Model of Motion-Picture Success (all with N = 190)
Full size table

3.2 The Two-Path Model of Motion-Picture Success

Results for the OLS regressions to estimate the Two-Path Model of Motion-Picture Success appear in Table 3. Notice that, because all key variables are represented by factor scores (mean = 0.0, SD = 1.0), all regression coefficients are identical to beta weights obtained with standardized variables. For ease of interpretation, these results also appear in the diagram shown in Fig. 1.

Table 3 Beta weights (p-levels) in regressions estimating the Two-Path Model of Motion-Picture Success (all with N = 190)
Full size table
Fig. 1
figure 1

Two-Path Model of Motion-Picture Success

Full size image

3.2.1 Critical and Popular Evaluation (CritPopEval)

As anticipated, along the artistically related path, CritPopEval shows a negative relationship with MktgClout: β CPE,MC = −.497 (p < .001).

3.2.2 Industry Recognition (IndRecog)

Further along the path that represents artistic integrity, IndRecog reflects a positive contribution from CritPopEval (β IR,CPE.MC = .564, p < .001) with an additional direct positive contribution from MktgClout (β IR,MC.CPE = .185, p < .01).

3.2.3 Critical and Popular Buzz (CritPopBuzz)

Meanwhile, along the commerce-related path, CritPopBuzz responds positively to MktgClout (β CPB,MC = .608, p < .001).

3.2.4 Market Performance (MktPerf)

Finally, as the ultimate criterion along the path representing commerce, MktPerf shows a positive impact of CritPopBuzz (β MP,CPB.MC = .492, p < .001) with an additional direct positive effect of MktgClout (β MP,MC.CPB = .492, p < .001).

4 Discussion

4.1 Substantive conclusions

Subject to certain limitations and caveats noted later, our findings appear to support the Two-Path Model of Motion-Picture Success proposed at the outset. They thereby justify some key substantive conclusions of interest to those concerned with research on audience responses to films, entertainment, and other cultural offerings in general or concerned with the contrast between art and commerce in particular.

Specifically, along a path that appears to reflect aspects of artistic integrity, critical and popular evaluation actually responds in the negative direction to the impression made by a film’s marketing clout as represented by its budget, opening screens, and first-week box-office receipts (β = −.497). This implies that both reviewers and audience members appreciate the aesthetic value of small-budget, low-profile, arthouse-type films of the sort that open on small numbers of screens as part of a slow-build strategy and that feature fine acting in dramas or comedies, sometimes in a foreign language with English subtitles. In other words—while the artistically oriented aspects of acting, scripts, and direction appear to pay off in the aesthetic appreciation of a film by consumers and reviewers—all those celebrity superstars and spectacular special effects (burning buildings, massive explosions, tidal waves, earthquakes, sinking ships, and car chases) that typically characterize films with high levels of marketing clout appear to detract from evaluative judgments of excellence. Further, along this same artistically oriented path, critical and popular evaluation contributes positively to industry recognition (β = .564) with an additional direct positive contribution from marketing clout (β = .185)—implying that the members of the industry who vote for Oscars and other awards place a high value on films regarded by reviewers and consumers as high in cinematic excellence but that industry members also respond positively to films that, in addition to receiving favorable evaluations, are also strongly promoted by big-budget marketing efforts. The net result of these two contrasting effects is that, when mediated by CritPopEval as an intervening link, the indirect negative influence of MktgClout on IndRecog (−.280 = −.497 × .564) is mostly cancelled by the additional direct positive effect of MktgClout on IndRecog when controlling for CritPopEval (.185) so that the overall correlation between MktgClout and IndRecog becomes quite weak and drops from statistical significance (r IR,MC = −.280 + .185 = −.096, n.s.).

By contrast, along a path that appears to reflect aspects of commerce, critical and popular buzz—which, as measured here, is uncorrelated with critical and popular evaluation—responds positively to the impression made by a film’s marketing clout (β = .608). This suggests that both reviewers and consumers enjoy the blockbuster aspects of big-budget, high-profile, mass-market films of the sort that open on large numbers of screens as part of a wide-release strategy and that feature big-name stars in action or adventure scenarios with spectacular special effects, lavish costumes, and other expensive trappings. Apparently, marketing efforts of the expensive and spectacular variety add strongly to the critical and popular buzz surrounding a film by enhancing the extent to which consumers and reviewers engage in word of mouth, click of mouse, and other communications intended to recommend the movie to others. Further, along this essentially commercial path, critical and popular buzz contributes positively to market performance (β = .492) with an additional direct positive contribution from marketing clout (β = .492)—implying that the market rewards films that generate a high level of buzz and, beyond that, those that represent high levels of marketing effort. The net result of these combined effects is that, when mediated by CritPopBuzz as an intervening link, the indirect positive influence of MktgClout on MktPerf (.299 = .608 × .492) is supplemented by an additional direct positive effect of MktgClout on MktPerf when controlling for CritPopBuzz (.492) so that the overall correlation between MktgClout and MktPerf becomes strongly and significantly positive (r MP,MC = .299 + .492 = .791, p < .001).

Here, we find support for the general principle of a contrast between art and commerce—consistent with the two-path model—wherein evaluations of excellence by reviewers and consumers influence industry recognition in ways that are essentially independent from the manner in which critical and popular buzz contributes to market performance. Indeed, in the present study, evaluation versus buzz and recognition versus performance are represented by orthogonal factors that are, by definition, uncorrelated. Evaluations related to industry recognition respond negatively to marketing clout. By contrast, buzz reflects the sorts of blockbuster characteristics that ultimately result in lucrative market performance.

In sum, the main thrust of the Two-Path Model of Motion-Picture Success proposed and tested in the present study is that, as responses to marketing clout, two separable paths—the first corresponding to aspects of art involving critical and popular evaluation leading to industry recognition, the second corresponding to aspects of commerce involving critical and popular buzz leading to market performance—are essentially independent, conceptually distinct, and empirically unrelated. And never the twain shall meet.

4.2 Industry implications

Our findings also suggest some implications for members of the motion-picture industry in general and for cinema producers, film distributors, or movie actors in particular.

The results of our study pose something of a paradox concerning the positioning of a film as one that wins favorable evaluative judgment as opposed to one that prompts a high level of buzz. Because the two paths are uncorrelated (r CPE,CPB = .000), it follows that a film can in principle be strong in one, strong in the other, strong in both, or strong in neither. However, as a practical matter, aspects of marketing clout that movie producers can manipulate directly tend to operate on the two paths in opposite directions. For example, big-budget wide-release films with elevated levels of marketing clout tend to promote strong critical and popular buzz along the commerce-oriented path but to detract from evaluations of excellence by reviewers and consumers along the art-related path. Hence, actionable measures to increase buzz on the former path may damage evaluation on the latter. This means that, in practice, a film producer must choose between competing positioning strategies. One potential approach would seek to maximize evaluation for excellence via a sort of art-for-art’s-sake ethos. An alternative and potentially more lucrative approach would strive to build buzz in the pursuit of commercial success. (Maximizing both simultaneously, while possible in principle, would seem to hinge on design characteristics and marketing variables not included in the present study—such as, say, superlative acting skills, a powerful script, superb direction, a massive advertising campaign, or heroic publicity efforts by celebrity stars who make tours of the talk shows.)

Our results suggest that the evaluation-maximizing strategy requires only a comparatively modest financial investment in low-cost production supported by a limited or slow-build distribution strategy. Conversely, an attempt to encourage positive evaluations would be hurt by excessive spending on aspects of marketing clout that can work against the impression of excellence. Indeed, opening a low-budget film on a small number of screens appears actually to enhance the favorability of its evaluations. It seems that economically produced motion pictures of the type found in the scattered specialty houses that show art films are perceived as better in artistic quality.

By contrast, the buzz-generating approach requires a major financial investment in a large-budget production of the type that typically opens big on a large number of screens via a wide-release strategy so as to win elevated levels of attention via word of mouth, click of mouse, and other sorts of interpersonal recommendations. Here, it seems that blockbuster movies such as those widely shown at ubiquitous multiplex theaters in every shopping mall across the land—for example, the typical action-adventure extravaganzas—are perceived as more suitable targets for critical and popular buzz, which in turn promotes commercial success by encouraging enhanced market performance.

We might also infer that—because different approaches to motion-picture marketing are implicated in the two positioning strategies just outlined—implementation of the two approaches would involve contrasting communication strategies highlighting various core aspects of the opposing kinds of film productions. A movie positioned to win favorable reviewer-and-consumer evaluations would benefit from publicity emphasizing its artistic value. A grass-roots, fine-tuned, selectively targeted promotional approach—perhaps involving brochures, flyers, and visits to local community groups—might work best as a way of conveying this message to the inherently small-scale target segment of interested viewers. By contrast, a movie aimed at creating critical and popular buzz would benefit from publicizing its big-budget aspects—perhaps via splashy advertising and glitzy trailers of the type that convey its spectacular scenery, lavish costumes, or other expensive special effects and that provide material for word-of-mouth communications by those who wish to recommend the film to others.

Meanwhile, actors and actresses concerned with managing their images in the marketplace would tend to behave rather differently in directing their skills toward alternative branches of our two-path model of success. Those wishing to establish reputations based on their professional acting skills along the art-related path would do well to accept low-paying roles in small-budget slow-build films such as dramas, documentaries, and even foreign or R-rated movies. By contrast, those seeking star power associated with appearing in blockbuster showpieces along the commerce-oriented path should seek roles in big-budget wide-release epic-length action, adventure, or sci-fi movies with spectacular special effects and other expensive production values. Paradoxically, it appears that such aspects of marketing clout detract from the stars’ abilities to demonstrate their skills as actors in ways that would foster favorable reviewer-and-consumer evaluations but offer opportunities to tap into the star-creating critical and popular buzz that rewards a true high-budget big-opening eye-popping blockbuster.

Hence, in the end, movie actors face choices comparable to those that confront film producers, distributors, and audiences (as well as participants in other cultural arenas)—namely, the trade-offs associated with the contrast between art and commerce, as represented by our Two-Path Model of Motion-Picture Success. Regarded by consumers as excellent or blessed by popular appeal? Rewarded by favorable reviews or loudly trumpeted by a clamorous media buzz? Distinguished by awards that embody the admiring recognition of industry members who comprise a niche market of discriminating cognoscenti or indiscriminately and insatiably viewed and enjoyed by the masses?

4.3 Limitations, caveats, and directions for future research

Like virtually any study in the tradition from which the present research emerges, our findings are subject to various limitations and caveats.

For example, our results pertain to the 190 films released in America during the year 2003. Even in the case of motion pictures, they might well differ in other countries or at other times. Future research should address the generalizability of our findings across time periods and geographical settings.

In addition, though we would expect our two-path approach to apply to other areas of the arts and entertainment, our model will clearly require appropriate modifications if applied in future research to other kinds of cultural offerings such as musical recordings, television programs, live theater, or the performing arts.

Similarly, the present study addresses just two main sources of evaluation and buzz—consumers and reviewers on Websites such as IMDb, Yahoo!, and RottenTomatoes. Future research should examine the role of additional information sources such as those found on Internet blogs or in collections of critiques compiled by professional experts such as Cinebooks, Jim Craddock, Roger Ebert, Leonard Maltin, Mick Martin and Marsha Porter, John Walker, and others.

Further, our measure of critical and popular buzz depends on a rationale that—though justified by the absence of buzz at low levels of evaluation—requires a small leap of faith to the effect that people (consumers and the journalistic reviewers who represent their tastes) tend to communicate most vociferously (word of mouth, click of mouse, published reviews) about films that they wish to recommend to others (acquaintances, fellow Web users, readers) as worthy of attention (likable, pleasurable, enjoyable). Justification for this leap of faith stems from the nomological validity of buzz in the way that it responds to marketing clout and helps to explain market performance. Further support for this assertion awaits future research based on the collection of respondent-specific data examining the differences in word of mouth, click of mouse, or other recommendations between movies or other cultural offerings that are more or less enjoyed by consumers and reviewers.

Finally, some may feel that the strong relationship between the logarithms of opening video rentals and gross video rentals appears almost too good to be true, accounting for over 95% of the variance in the latter aspect of market performance. We, too, have had our suspicions that this relationship may involve some sort of artifact. However, our query addressed to those who maintain the IMDb Website evoked a response expressing surprise that we would find anything strange in the strength of this relationship. Isn’t it logical, the argument runs, that video rentals should start at a high level and then cumulate predictably as they inevitably peter out over time? Perhaps, for now, it is safest to content ourselves with the observation that if this strong relationship continues to appear in future research with demonstrably valid and independently obtained measures, it will represent one of the strongest associations that we have ever observed in correlations based on real-world data.

In sum, all the caveats and limitations just enumerated support the need for future investigations of the two-path model using data from multiple time periods, multiple countries, and multiple ways of operationalizing the key constructs of interest—perhaps including additional kinds of cultural offerings such as television programs or musical recordings. Refreshingly, we note, the availability of movie- and entertainment-related data on the Internet and elsewhere has grown by leaps and bounds in recent years. This progress augurs well for the potential feasibility of future studies aimed at addressing the issues just raised.

4.4 Finale

In that spirit, let us end with the potentially happy observation that—in accord with our Two-Path Model of Motion-Picture Success—movies endear themselves to members of our culture in part by virtue of their ability to offer something to everybody or, at least, to provide two very different sorts of value to two very different kinds of targets. Some look to films as part of their appreciation for the finer things in life—along the art-related path of reviewer-and-consumer evaluations leading toward industry recognition. Others—along the commerce-oriented path of critical and popular buzz leading toward market performance—relish the big-budget mass-marketed spine-tingling blockbuster-type special effects that thrill them on the big screen and send them from the movie theater with the desire to recommend this exciting cinematic experience to others. We should not make the mistake of looking for ways to combine these two essentially independent phenomena into just one measure or expression of excellence or appreciation. Nor should we forget that—while the former helps to win awards for industry recognition—the latter plays the major role in contributing to a motion picture’s market performance.