Is multichannel advertising worth it? A comparison between fast moving consumer goods and service product categories

(1)

Is multichannel advertising worth it?

A comparison between fast moving consumer goods and service product

1. Introduction

Online advertising is rising steadily in the last decade. Based on annual reports of the Interactive Advertising Bureau, 78% quarters have experienced positive growth of online advertising over the prior quarter since 2003 (IAB, 2014). As consequence, the rise of smartphone usage together with the popularity of social media in our lives changed drastically the way people exchange, share and consume information. While social media advertising demonstrates continues growth, mobile advertising reached its strongest growth with total revenue of $5.3 million in the first part of 2014 (IAB, 2014). In total, online advertising revenues increased 15.1% in the first part of 2014 compared to 2013, while internet advertising revenue reached a record high of $12.4 billion in the third quarter of 2014 (IAB, 2014).

More importantly, customers are turning away from traditional media channels such as radio and print advertising, and moving towards new interactive media channels such as mobile, internet and interactive television (Kumar, 2015). Customers, for the first time, are spending more time on interactive new media than on traditional ones (Kumar, 2015). Furthermore, not only time allocation between channels is changing, but also the way that consumers search and acquire information about a product or service. Customers are able to search for information through their smartphones while shopping offline and quickly spot additional information online about a product from a television or banner advertising. Apparently, customers prefer new media over traditional ones.

(5)

5 sales than online sales yielding a great deal of revenue benefit (Abraham, 2008). On the other hand, for the trend of gathering information from offline advertising to buy online, the main offline advertising comes from broadcast television. Even though based on the IAB report (IAB, 2014) that the internet advertising surpassed television by 7%, the 2013 broadcast TV advertising still garnered $40.1 billion in its revenue. Related study shows that television advertising has positive impact on online shopping, in which advertising content plays a key role in online traffic and sales (Liaukonyte et al., 2014). Those advertising contents, especially those with action focus, directly increase the internet traffic and online sales; but those contents with information focus and emotion focus reduce website traffic while still increase online sales. The study results indicate that there is a positive net effect on sales from television advertising for products of most brands included in the research (Liaukonyte et al., 2014).

As consequence, It is not a surprise that there is currently a trend from companies to explore the influence of the new media advertising as there is huge potential for profit in online advertising. Indeed, companies develop systems in contrast to traditional media, and results from online advertising can be measured more effectively due to the accountability of online advertising. As a result, marketing is becoming an integral part of the organization (Kumar, 2015). But a simple shift from traditional media to new media is not enough for companies to realize their situation in new trends of advertising because there are different channels in online advertising, and there are different ways of advertising for offline companies to build their brands and maintain their profit. One of the new concepts of current marketing is multichannel or cross-channel marketing, which studies across marketing channels such as traditional print publication, social media, mobile phone, email, blog, online news sites, search engine, and company website (Glasner, 2015). The multichannel advertising can also be divided as paid advertising and free advertising as so called “organic” advertising. Cross-channel marketing is the current trend of business marketing because it is the near future of marketing, and well handling and controlling multichannel marketing will decide if a company can survive in the market of the 21st century (Glasner, 2015). One of the key aspects of multichannel marketing is interacting marketing, which means a strategy that considering the interaction between different channels, to define a best way for marketing that targets a specific customer (Ratchford, 2015).

(6)

6 products of clothing finds that the cross-effect is large particularly crossing from online advertising to offline sales, which has different effect on customer counts and spend (Dinner, et al. 2011). When comparing online and traditional advertising, it is found that online advertising is effective in increasing offline customer count while traditional advertising is effective in increasing online customer spend (Dinner, et al. 2011). This indicates that each company needs to work on marketing research based on multichannel marketing strategy and interacting marketing to find out a marketing strategy that fits to their products.

In addition, the internet can also be used for offline companies selling mostly through offline channels, especially for their brand building, which is important for offline companies. Research finds that building brand through online channels is not just faster, but more importantly it is more effective compared to offline channel (Ibeh, et al., 2005). And online brand building should consider strategies such as creating a sustainable brand, establishing effective brand communication, and developing an international brand.

Accordingly, it is important for companies to be aware of the potential synergies and interaction between traditional marketing channels and new interactive marketing channels to be able to allocate advertising budget effectively, and take full advantage of new opportunities from cross-channel marketing. The goal of this study is to examine and compare advertising effectiveness and cross-channel advertising effect in two different product categories: (1) A fast moving consumer good product which is sold through offline channel; and (2) a service product which utilizes both offline and online channels to reach customers. Hence, the research question of this study is: What effect does internet advertising have on the sales of a product sold online or offline? And how does online–offline channels synergize together on product marketing. Moreover, the role of brand awareness in relation to marketing activities will be examined.

(7)

(8)

8

2. Literature review

Advertising is one of the most important elements of marketing which has impact directly on business profit and revenue. Current advertising, is facing challenges from multiple advertising channels that are available, and the handling of advertising budget is critical for business survival and development. Based on available advertising channels with their effectiveness, this section, as a literature review, discusses offline advertising effectiveness, online advertising effectiveness, and synergy effect between offline and online advertising.

2.1 Offline Advertising effectiveness

Offline advertising is currently not as efficient as it was decades ago, but it is still playing its important roles in products with certain categories such as food and non-food goods that are not ready to be sold through online advertising (Sethuraman et al., 2011). It is critical to understand the effectiveness of offline advertising which is often measured using advertising elasticity or say the increased percentage in sales for 1% increase in advertising (Assmus et al., 1984). Identifying elasticity and factors influencing elasticity will clarify advertising effectiveness and help firm to further decide its marketing strategies.

A recent meta-analysis study by Sethuraman et al. (2011) investigated 751 short term and 402 long term advertising from 56 studies on advertising elasticity published between 1960 and 2008. The study does not include any effect, of online advertising and thus it is only about the offline advertising category. It identifies several features of offline advertising elasticity, as the general elasticity is higher for durable goods compared to nondurable goods in the early stage compared to the later mature stage of the product cycle, for data from yearly compared to quarterly, and when measured in gross points compared to monetary terms. The findings of this study suggest that firms need to focus on advertising during early stage and on pricing during later stage of each product cycle for durable goods. As an additional finding, they show that advertising elasticity is generally higher in Europe compared to in North America, which may be the consequences of under-advertising in Europe and over-advertising in the United States. With regarding short term and long term effectiveness, it is found that the short term elasticity is of .12 on average, and the long term elasticity is of .24 on average, in which most of the short term elasticity results hold for the long term elasticity.

(9)

9 impact on sales. They found that, for short term advertising elasticity, systemic variability is related to several factors including product type, model specification, measurement, estimation and study setting. It shows that short term elasticity in models with carryover effect and those with exogenous variables are significantly lower, and long term elasticity is even lower. More importantly, it is found that short term elasticity is of .22 on average, and long term elasticity is of .47 on average. Compared to what Sethuraman et al. (2011) found, there is an obvious decrease in offline advertising effectiveness over the years as the short term elasticity is decreased from .22 to .12, and the long term elasticity from .47 to .24. Accordingly, Sethuraman et al. (2011) argue that the changes in advertising elasticity indicate a reduction in offline advertising since it does not decline due to recession. The major reasons for the changes are considered to be the rise of the internet as an additional channel and in general as a source of information, and consumers’ capability of opting-out television commercials through electronic devices play important roles to the decrease of offline advertising effectiveness (Sethuraman et al. 2011).

2.2 Online advertising effectiveness

As a consequence, a trend shows that more research is ongoing in scientific literature towards the understanding of the potential of internet advertising, and more speci fically, the influence of internet advertising on customers. Internet or online advertising can be divided into three categories which are paid advertising (banner, text and video advertisements), free advertisement (organic results from search engines such as Google), and social media advertisement or user generated content.

Paid Advertisement

(10)

10 Lots of previous studies of paid advertising are about display and banner advertising (Dijkstra et al., 2005; Manchanda et al., 2006), and those studies about the response rate of paid advertising search are rare. A study focused on estimating response rate of paid advertising search defines that, in terms of click through and conversion, customers who click but do not purchase immediately are considered as lost (Rutz and Bucklin, 2011). A way to reduce the lost may increase marketing effect, and it is confirmed that when customers are exposed to branded information, it ensures generic search and increases relevance awareness. But leaks out of the bucket are observed since the relevance awareness decays when time passes, which can be resolved by enhancing the relevance awareness to promote subsequent brand related searching (Rutz and Bucklin, 2011).

With regard to the relationship between online advertising and offline sales, Lewis and Reiley (2014) studied the effect of online advertising on offline sales and found that an increase in online advertising budget increases sales by 5%, in which 93% of the sales is in offline stores while only 20% of the increase is attributed to customers who click on the advertising. Further support from Rutz and Bucklin (2011) found the spillover effect of online advertising that incremental revenue of retailers are seven times more than the cost of advertisement indicating the practical profiting from paid advertising. In addition, online advertising enhances purchase probability and amount that amplify business profitability from paid online advertisement.

To compare marketing effect from the traditional and new advertising, Dijkstra et al. (2005) explored the effect of print, television and internet advertising on consumer responses to investigate their separate and synergy effect. The cognitive responses of participants are higher with television campaigns compared to multi-media campaigns, and print campaigns are found to have the same effect as multi-media ones. Their results indicate that offline advertising has its predominant feature compared to online and traditional ones, applying internet too much in marketing campaign may not be a good idea because it may lower customer attention, and internet may not be a substitute to television and traditional campaign but a complement instead. Dijkstra et al. (2005) mentioned that the limits of the study are similar to other studies as forced-exposure and short-time interval of questionnaire answering because participants were forced to watch advertising. Since participants had to decide whether to click on advertising pages on the internet, it may affect the internet advertising results leading to less responses.

(11)

11 using number of advertising, click volume, cost per click, and determination factor of competition including demand and supply factors affecting entry probability. Empirical analysis was used to analyze the results showing that the number of competing advertising has significant impact on click volume, cost per click and decay factors. In addition, counterfactual analysis was used to analyze profit implication of two policies as coupons providing to advertisers and consumer encouraging engaging in in-depth search or click-through. Their results show that paid search host can increase revenue by increasing decay factor which may intensify competition between search hosts, and that appropriate selection of coupons from 10 to 30% of average management cost increases profit of paid search host. When comparing to click volume decay factor, the coupon system is under more control and thus more practical to implement by paid search host.

Another parameter affecting competition in paid advertising is purchase conversion of keywords. Rutz et al. (2012) studied the position of a text advertisement combining a click-through and conversion model to estimate the purchase conversion of keywords. Their model enables estimation of keyword conversion and click-through rates when accounting endogenous position of text advertising. They found that the purchase conversion rates of keywords and click -through are improved when advertising position improved, in which about one-third of total conversions are derived from increased click-through and about two-third of total conversion are derived from increased conversion rates. But there is no effect of competition observed from this study. On the other hand, it is reported that position on conversion has both positive and negative effects. For example, Agarwal, Hosanagar, and Smith (2011) found that keywords helping increasing awareness and generating purchases do not success at a higher rate when advertising is placed in higher position. Ghose and Yang (2009) found that keywords with higher positions yielding higher conversion rates and click-through are unnecessary the most profiting factor because profits can be higher when keywords are at middle position. Different studies on keyword position indicate that advertisers need to be aware that chasing a higher keyword factor may have negative impact on profiting, and thus considering overall customers’ behaviors on competition is the right way to yield higher profit (Jerath et al., 2014; Rutz et al., 2012).

(12)

12 unobserved individual behaviors, and correlations between advertising impression, website visits and conversion are considered; and accumulation with decay of advertising effect, advertising wear-out as well as restoration effect are allowed in this study. Results show that website visits are up to 12.7% of increase and conversion number are up to 13.8% of increase, which indicate that the factors of multiple advertising creative and individual advertising impression history are important in paid advertising campaign. It is suggested that online campaigners can increase various creative contents based on visitors’ individual impression history to increase website visits and conversions.

To compare the difference between free and paid advertising, Jerath et al. (2014) studied the behavior of consumers on both free and paid links over 1.5 million user searches for multiple keywords and found that click activates are low after keyword search which is focusing on organic links with 95% of click numbers. They also found that there are significant differences between different composition of keywords, and keywords with popularity is an import indicator of click tendencies and should be kept in mind of advertisers. For example, as keyword with lower popularity, it is searched by consumers generating more clicks with more sponsored links indicating that less popular keywords are interesting to consumers to search for more information with higher tendency to purchase behaviors. Authors concluded that consumers who search more popular keywords are more likely to click on an organic link, whereas consumers who search for more complex and unique keywords are more likely to click on a paid link. In addition, it is concluded that consumers click on less popular keywords have a higher tendency to search and purchase who should be the targets of advertisers.

2.3 Social media advertisement

(13)

13 traffic. It is also revealed that the firm’s self and its rivals’ buzz and traffic have impact on market competition significantly. Advertisers can learn from this study to manage website design, consumer relationship, and the best way to react to competitors’ strategy changes.

Nelson-Field, Riebe, and Sharp (2002) conducted a study based on Andrew Ehrenberg’s negative binomial distribution theory, which considers the most effective advertising media are those reaching both light and heavy buyers of a brand. The study investigated two brands of fast moving consumer goods (FMCG) of buying concentration of Facebook fans compared to actual buying bases of the brands. Each of the two brands were distributed using negative binomial distribution for the buyer bases while using the opposite pattern as distributed towards the heaviest buyers for the Facebook fans to make it unappealing on purpose. Results show that the Facebook fans based buying distribution is the opposite to typical buyers, which shows that Facebook fan buyers are with a significantly higher incidence of heavy buyers. This indicates a deficiency of earned media and raises question about the value of Facebook as standalone medium for earned advertising research. As consequence, advertisers need to identify the value of each medium as earned advertising research platform; for instance, if Facebook is chosen as a platform, other media that can reach light buyers can be used as compensation, because a platform like Facebook can provide opportunities to listening to customers from the same and different companies, and offer potential of creating networks of active brand advocates to attract light buyers.

(14)

14 2.4 Synergy effect of cross-channel advertising

Cross-channel advertising is a relatively new concept and its studies are emerging recently. Study from Chan et al. (2011) stresses the importance of offline channel transactions for customers who were acquired through online text advertisement compared to traditional one, which further supports the importance of connectivity between online and offline channels and firms need to evaluate long term impact of cross-channel advertising investment. Most importantly, Dinner et al. (2014) reviewed literature and suggest that online activities have major impact on offline sales, and offline activities promote online sales. For example, Naik and Peters (2009), as the first to estimate the impact of multichannel to consider dependent variable for automobile choice by examining the impact of offline advertising, online advertising and direct mail on offline sales. They found that online advertising has impact on offline sales, and direct mail (traditional advertising) has impact on online sales, and the most effective advertising channel is the traditional advertising as using catalogs. Even though no actual purchase was measured, this is an important finding for cross-channel concept in synergistic effect of cross-channel marketing.

By measuring actual purchases in sales, Wiesel et al. (2011) found significant effects between channels by applying their approach in a B2B medium sized company as a more scientific approach for their marketing budget allocation. They found very high sales from online advertising and much lower sales from flyers and faxes advertising with sales elasticity of 4.35, .05 and .04 respectively based on a B2B company of major clothing retailer. The strong cross-channel impact is found as 73% of online advertising using Google is derived from offline sales and 20% of direct mail flyers is derived from online sales. This study makes a major progress towards cross-channel impact on B2B field indicating that Google search influences purcha se outcomes significantly. Similar to Chan et al. (2011) that stresses the importance of offline channel transactions for customers acquired through online text advertisement, results from Wiesel et al. (2011) showed that almost three quarters of Google AdWords profit is attributed to offline sales. Further supporting the importance of connectivity and synergistic effect between online and offline channels.

(15)

15 setting to promote branding and pricing; they also introduces an inexpensive model to assist firm advertisers to access relative effectiveness of multichannel advertising to determine budget allocation across media channels.

As one of the studies with thorough research of synergistic cross-channel advertising between online, offline and traditional advertising, Laroche et al. (2014) studied the long term and short term effect of online and offline advertising on brand searching with factors of adjustment period and brand search queries based on hypotheses formed by using the Elaboration Likelihood Model (ELM). They found that the factors affect brand searches are total marketing expenditures, total advertising impressions, and online display exposure, in which total advertising impression and total marketing spending have the greatest short term effect on brand queries. For example, the adjustment period lasts about three weeks, and the time with the greatest elasticity is the first three time periods, i.e., the first three weeks. Furthermore, the impact of television and online advertising on brand searches are similar and to a lesser extent when compared to the impact of total advertising impression and total marketing spending; and the total impact of advertiser efforts result in the increase of brand queries. When comparing different advertising channels, it seems that the online advertising channels are more effective because it allows brand search to occur more often compared to traditional channels; and it emphasize the importance of consistency in advertising across different media for synergistic marketing campaign because large advertising campaign yield higher traffic to company websites, which is important that the campaigns drive sustained traffic to company websites to yield purchase incentives. A key finding of the study is that firms may be able to maximize brand queries by marketing efforts to design media strategies to maximize and prolong the short-term brand search through adding additional online information.

(16)

16 significantly, which does not appear to influence offline sales and profits but can be considered for brand awareness and brand building.

3. Conceptual model

As shown by Figure 1, the conceptual model of the research is provided as an overview to be used in the next section.

Figure 1. Conceptual Model

4. Hypothesis formation

Based on information reviewed in section 2, it is shown that offline advertising, online advertising, and social media (web conversation) are different advertising channels, which can be chosen by firms according to their product and service categories and budgets. Each of the three advertising channels may be predominant in a specific product marketing effect depending on the campaign design and expenditure impact. As a new trend with its powerful marketing impact, cross-channel marketing belongs to a new strategy of marketing that cannot be ignored by most of firms making their endeavor effectively in product marketing.

(17)

17 offline channels are synergized together on product marketing?”, hypotheses are needed to be formed to enable related study design.

ELM theory, as mentioned in section 2, has been used widely in cross multichannel marketing research to help hypothesis formation and study design, which was applied in this study to form hypotheses (Allen et al., 1998; Chang and Thorson, 2000; Putrevu and Lord, 2003; Laroche et al. 2014). As Laroche et al. (2014) believe, the application of the ELM theory to explore the synergistic effect of multichannel advertising on communication of an advertising message of a specific product will result in linking of the motivations to scrutinize arguments to the likelihood of message elaboration. For example, when the likelihood of elaboration is low, it indicates that the individual customers are changing attitudes from the initial attitude along the peripheral route; and when the likelihood of elaboration is high, it indicates that the synergistic effect across multichannel is more effective when individual customers are changing attitudes along the central route. In addition, customers with higher elaboration likelihood are more motivated in scrutinizing arguments of understanding true merits of brands, and in making attempts of presenting from one medium to other media to yield cross-channel advertising effect.

In this study, firstly, even though offline advertising seems to decline as the consequence of the emerging of online advertising, it is confirmed that offline advertising is still affecting sales effectively (Sethuraman et al., 2011). As some of the important features, the offline advertising elasticity is higher particularly for durable goods, but it also works for nondurable goods; and firms need to focus their offline advertising on early stage in their product cycles. Moreover, Clark, Doraszelski and Draganska (2009) state that offline advertising may also influence brand awareness. Since the two categories of goods in this study are both nondurable goods, and some of them cannot be sold online, as a result, the following hypotheses is formed.

H1: Offline advertising expenditures have a positive effect on sales.

H2: The effect of Offline advertising expenditures on sales is mediated by brand awareness.

(18)

18 expected to be lower compare to FMCG because the service goods are nondurable and can be sold online and the effect of online advertising is considered as higher than offline advertising, as mentioned in the next hypothesis.

Secondly, based on the ELM theory (Laroche et al., 2014) and other reports (Rutz & Bucklin, 2011; Lewis & Reiley, 2014), internet advertising has its impact on sales of product sold online and offline, and online - offline channels are synergized together on product marketing to enhance expenditure. The internet advertising is considered as the major channel based on the literature review that customers are expected to have higher level of elaboration likelihood to be exposed when they are offered with multichannel advertising through media. In addition, there is a relationship between firm value, brand awareness and sales effect because online buzz increases brand awareness remarkably since web buzz influences the value of a firm, and web traffic is affecting positively brand awareness (Luo & Zhang, 2013); and brand awareness promotes marketing outcome (Huang & Sarigöllü, 2012). According to the ELM theory, total advertising impression and total marketing spending have the greatest short term effect on brand awareness (Laroche et al., 2014). Based on these results, hypotheses 3 and 4 are formed:

H3: Web conversation (online buzz) has a positive effect on sales.

H4: The effect of Web conversation on sales is mediated by brand awareness.

We expect that there will be higher online advertising and brand awareness effect on service sales especially because it is promoted by both online and offline advertising channels with bias of online channel; and sales with brand awareness through online advertising of FMCG will be effective partially since it is promoted mostly by offline advertising.

(19)

19 categories of goods (Clark et al., 2009). And thus online advertising should last certain period of time to allow brand awareness and identification of ideal identity and quantity of advertising effect of a specific category of goods. Based on these findings, hypotheses 5 and 6 are formed as follows.

H5: Online advertising (paid advertising) has a positive effect on sales.

H6: The effect of Online advertising on sales is mediated by brand awareness.

The expected effect from paid advertising with its resulting brand awareness will be higher for service when paid advertising last certain period of time because it relies on online advertising, but will be lower for FMCG because it relies on mostly offline advertising.

Last, based on the literature review, the effect of interaction between online advertising, offline advertising and cross-channel advertising will be studied with each of the two different goods with different product categories. As customers who are exposed to online text advertisement made higher offline transactions compared to traditional advertising confirming the importance of long term connectivity between online and offline advertising channels (Chan et al., 2011). On the other hand, it is also confirmed that offline activities promote and enhance online sales (Dinner et al., 2014). In this study, we expect the effect on sales to be stronger from both online and offline advertising, as well as web buzz to each of the goods of two different categories with specificities targeting each type of goods. Therefore the following hypotheses are formed.

H7: Offline exp.*Web conversation cross-effects have a positive impact on sales. H8: Offline exp. *Online adv. Cross-effects have a positive impact on sales H9: Web conv.*Online exp. Cross-effects have a positive impact on sales.

5. Methodology

5.1 Data Sample

(20)

20 included two different datasets. The first dataset included the FMCG product category and more specifically a type of yogurt, it included entries from the first month of 2009 to the end of 2010. In total, the dataset included 105 entries. Furthermore, the second dataset included the Service product category and more specifically an electricity provider in energy sector. It included entries from the 29th_{week of 2009 to 38}th_{week of 2011. In total, the dataset included 114 entries. Both} datasets were aggregated weekly. Before proceeding, both datasets were examined for duplicate entries and outliers which could affect the validity of our results. As a result, a duplicate entry for FMCG dataset was deleted and an outlier in service dataset. In the next section each variable is explained separately to provide better understanding of our research.

5.2 Variables

Dependent Variable

Sales: The dependent variable in our model is Sales. The sales variable does not distinguish

between online and offline sales but includes the sum of both as one entry. Therefore, we cannot distinguish between the two and draw further conclusions about the effect of independent variables. Nevertheless, we can safely assume that for the FMCG category the online sales would be very few, if there are any at all. For both datasets sales are measured weekly during a two year span. Even though there is an overlap of the time period the data was recorded, it is not the same for both products. A descriptive of the variable is provided in tables one and two.

Independent Variables

Offline Advertising expenditures: The first independent variable is offline advertising

expenditures. This variable includes the sum of money spend on offline advertisement channels such as T.V, radio, newspaper and flyers by each company. It is measured weekly as well.

Web conversation: The second independent variable is web conversation. Web

(21)

21

Online paid advertising: The last independent variable is online paid advertising. It

measures the expenditures through Google advertisement platform, which includes both targeted text advertisement and banner ads. In this study it is used as the derived cost for the advertising campaign for each product. It is measured weekly but we cannot compare the costs between the two products as Google advertisement platform when calculating the cost of each ad takes into account the popularity of keyword targeted and competition of each keyword. Therefore, it is safe to assume that for energy sector the cost of an online ad would be higher than the dairy product.

Mediator

Brand Awareness: The type of measurement which is included in the dataset is spontaneous

brand awareness. It is defined as “whether the brand is recalled or not without the use of any aid” (Agarwal and Rao, 1993). In this case, as the dataset for both products is aggregated weekly, it is interpreted as the percentage of consumers who recalled the brand (Agarwal and Rao, 1993). Brand awareness in this study is tested as a mediator.

Interaction Effects

Finally, to include the cross-effects on sales of online and offline channels and also the combined effect of both online variables, three new variables were calculated. After mean-centering offline advertising expenditures, web conversation and online paid advertising, (offline advertising expenditure*web conversation), (offline advertising expenditures*online paid advertising) and (online paid advertising*web conversation) were calculated to assess the interaction effects between the independent variables.

In the following tables a summary of the variables of this study is provided for each product category.

mean median St. deviation minimum maximum Sales 301756.06 290170 64503.83 205760 579412 Offline Adv. Exp. 18375.93 0 34267.97 0 115223.69

Web Conversation 13.95 3 17.63 0 89

Online paid adv. 50.82 44.37 49.92 0 151.67

Brand awareness 23.23 23.26 4.4 13 30.94

(22)

22 mean median St. deviation minimum maximum

Sales 2892.82 2824.5 953.76 202 4904

Off. Adv. Exp. 240334.75 138592 242985.31 1130 997357

Web Conversation 53.45 0 124.45 0 655

Online paid adv. 179.93 184.96 124.98 0 550.75

Brand awareness 53.45 53.5 5.96 40 73

Table 2: Variable summary for Service dataset

5.3 Method

Multiple regression Analysis

To test the hypotheses formed on section 4 a multiple regression analysis was used. The analysis was conducted by the software SPSS. As SPSS software does not include an option for mediation analysis the PROCCESS-macro add-on was installed to proceed with the analysis. The PROCESS-macro add-on was programmed by Dr. A. Hayes and utilizes the bootstrap resampling technique to calculate the indirect effect in a mediation study (Preacher and Hayes 2008). In the conceptual model of this study we hypothesized that the effect of all independent variables on sales is mediated by brand awareness. Therefore, we have multiple independent variables simultaneously having an effect on the dependent variable though the mediator. Hayes (2013) suggests to regress each variable separately, testing each time the remaining independent variables as covariates and finally interpret the coefficients together or separately. This method has been used successfully by other researchers facing a similar problem (e.g Von Hippel 2011, Simpson et al 2011). Before proceeding with the analysis a correlation test should be conducted. As it stressed by Hayes, when including multiple independent variables in a mediation analysis. High correlations between the independent variables will cancel the effect of each other (Hayes 2013). In our case, there are no correlations between the independent variables (Appendix A). Therefore we can proceed with the analysis. Last, the model of the estimation is provided below:

𝐵𝑅𝐴𝑁𝐷𝐴𝑊𝐴𝑅𝐸𝑁𝐸𝑆𝑆 = 𝑖1 + 𝛽1𝑂𝐹𝐹𝐴𝐷𝑉 + 𝛽2𝑊𝐸𝐵𝐶𝑂𝑁𝑉 + 𝛽3𝑂𝑁𝐿𝐴𝐷𝑉 + 𝑒𝑀

(23)

23

6. Results

6.1 Assessing the mediator effect

As already mentioned in methodology section to examine if the mediation effect is significant three methods of assessing mediator effects were applied (Baron and Kenny test, Sobel test, and Bootstrap method). According to Baron and Kenny test, in order to have a significant mediation effect in the model three conditions must apply. First, both the independent variable and the mediator should have a significant effect on the dependent variable. Second, the effect of the independent variable on the mediator should be significant. Third, the effect of the independent variable on the dependent variable when mediator is excluded should also be significant (Baron and Kenny 1986). Second, Sobel test, which directly estimate the indirect effect of the independent variable through the mediator on the dependent variable in a single step (Preacher and Hayes 2008). While Sobel test is preferred for larger samples (n>200), nevertheless, it is calculated for further insights. Last, a bootstrap method that generated 10000 sample with an estimate to the 95% confidence intervals was used for testing the mediation effect.

Baron and Kenny test

For FMCG product category the results were the following: The effect of Brand awareness on sales was found not significant (B=-2074.66, p-value=0.204). Furthermore, from the three independent variables only offline advertising expenditures was found significant (B=0.479, p-value=0.011) while both web conversation (B= -159.126, p-value=0.7) and online paid advertising (B=14.88, p-value=0.91) had no significant effect on sales. Since brand awareness had not significant effect on sales, it violates the first condition of Baron and Kenny test (Appendix A). For service product category the results were the following: As in FMCG product category the effect of brand awareness on sales was not found significant (B=3.104, p-value= 0.839). Moreover, from the three independent variables offline advertising expenditures (B=0.000, p-value=0.467) and online paid advertising (B= -0.05, p-value=0.944) were not found significant and only web conversation was significant (B= -2.590, p-value=0.00) with a negative effect on sales. Therefore, as well as in FMCG product category, in service product category there is not a mediation effect as the results violate the first condition of Baron and Kenny test (Appendix A). Sobel test

(24)

24 were estimated with the use of PROCESS-macro add-on (Preacher and Hayes 2008). In the tables below the results of the estimation for each product category are provided. As suggested by Dr. A. Hayes each indirect effect was calculated separately. The detailed calculations can be found on Appendix B of this paper. In the first column of each table, the independent variable that was used for calculating the indirect effect though the mediator on sales is provided. Followed by the effect, standard error, Z-test and p-value.

Independent Variable Effect Standard error Z-test p-value Offline advertising expenditures -0.0085 0.027 -0.3144 0.7532 Web conversation 248.884 251.5223 -0.9895 0.3224 Online paid advertising 34.7209 39.9559 0.869 0.3849 Table 3: Sobel test results for FMCG product category

Independent variable Effect Standard error Z-test p-value Offline advertising expenditures 0.0000 0.0001 0.1944 0.8459 Web conversation -0.0252 0.1412 -0.1787 0.8582 Online paid advertising -0.0072 0.0773 -0.0935 0.9255 Table 4: Sobel test results for Service product category

For both FMCG and service product categories the results of Sobel test indicate that there is not a significant mediation effect for any of the independent variables. More precisely, for FMCG, offline advertising expenditures test= -0.3114, p-value=0.7532), web conversation (Z-test= -0.9895, p-value=0.3224) and online paid advertising (Z-test = 0.869,p-value=0.3849). For service product category, offline advertising expenditures (Z-test=0.1944, p-value=0.845), web conversation (Z-test= -0.1787, value= 0.858) and online paid advertising (Z-test= -0.0935, p-value=0.925). The cut-of values for being significant are (Z-test > 1.96) and (p-value < 0.05). Therefore the indirect effects of the three independent variables for both product categories on sales through brand awareness are not significant, further supporting the results of Baron and Kenny test (Appendix B).

Bootstrap method

(25)

25 Independent variables Effect SE 2.5th_{percentile 97.5}th_percentile

Offline advertising expenditures -0.0085 0.0256 -0.066 0.0429 Web conversation 248.884 234.8487 -705.7777 235.8174 Online paid advertising 34.7209 35.8171 -30.7396 111.7684 Table 5: Bootstrap test results for FMCG dataset

Independent variables Effect SE 2.5th_{percentile 97.5}th_percentile

Offline advertising expenditures 0.0000 0.0001 -0.0002 0.0003

Web conversation -0.0252 0.1482 -0.3544 0.2616

Online paid advertising -0.0072 0.0789 -0.2044 0.1375 Table 6: Bootstrap test results for Service dataset

In order for mediation effect to be significant the value of zero has to not be between the 2.5th_{and 97.5}th_{percentile. In our case, for both product categories, the indirect effects of every} independent variable on sales through brand awareness are not significant since zero is between the 2.5th_{and 97.5}th_{percentile for every independent variable}_{(Appendix B). Therefore, bootstrap} method also confirms Baron and Kenny test and Sobel test.

Concluding, all three methods agree that there is not a significant mediation effect for any of the independent variables for both product categories. As a result Hypotheses 2, 4 and 6 are rejected for both product categories.

6.2 Multiple regression analysis

As mediation effects were not found significant for both datasets, to proceed with the analysis a multiple regression analysis was conducted to test the remaining hypotheses (H1, H3, H5, H7, H8 and H9) and cross-effects on sales. Three different models were tested for each dataset, the first one includes only the three independent variables. The second one includes the mediator as an independent variable and the last one includes also the independent variables for interaction effects.

(26)

26 2 adjusted R2_{is slightly lower than model 1. Model 3 had the highest explanatory power out of the} three models 14% and 9% based on adjusted R2_{(Model=1, R}2_{= 0.115, adjusted R}2_=0.091, p-value=0.004) (Model=2, R2_{=0.116, adjusted R}2_{=0.083, p-value=0.009) (Model=3, R}2_{= 0.146} adjusted R2_{=0.09, p-value=0.017). A summary of the models is provided on the tables below.}

model R-square R-square adj. F-test df1 df2 p-value 1 0.063 0.035 2.256 3 101 0.086 2 0.078 0.041 2.112 4 100 0.085

3 0.105 0.041 1.629 7 97 0.136

Table 7: Model summary for FMCG dataset

model R-square R-square adj. F-test df1 df2 p-value 1 0.115 0.091 4.782 3 110 0.004 2 0.116 0.083 3.566 4 109 0.009

3 0.146 0.09 2.589 7 106 0.017

Table 8: Model summary for Service dataset

(27)

27

7. Discussion

The aim of this study is to determine and compare common advertising effectiveness and cross-channel advertising effectiveness among and between businesses of two different product categories. The first product category was a fast moving consumer good (FMCG) product which is sold by advertising through offline channel, and the second product category was a service utilizing offline and online channels to advertise the product to its customers. Furthermore, the study intended to examine the role of brand awareness on sales through online and offline channels and in what ways brand awareness is affected. Data results from this study will help enterprise businesses to apply potential synergy and interaction effect between offline and online marketing strategies to their advertising budget accordingly to take advantages and receive product sales benefits.

When comparing the effect from offline advertising on FMCG and service products, results supported the hypothesis 1 for FMCG product sales but rejected it for service product sales, which is expected because FMCG product relies on mostly offline advertising but service product relies on offline and online channels. Another possible reason that offline advertising did not have positive impact on service sales can be because of the decrease of offline channels in recent years.

Coefficients

Variables FMCG Service

Model 1 2 3 1 2 3

Web Conversation -408.010 -159.126 -122.024 *-2.615 *-2.589 *-2.724

Offline advertising expenditures *0.470 *0.478 *0.537 .000 .000 .000

Online Advertising expenditures 49.604 14.883 61.327 -.057 -.050 .215

Brand awareness -2074.668 -2222.033 3.104 .962

Offline exp. * Web Conv. .003 **-5.449E-06

Offline exp. * Online exp. .007 1.921E-06

Web Conv. * Online exp. 4.572 .001

(28)

28 As reported by Sethuraman et al. (2011), there is an obvious decrease in offline advertising effect due to the emerging of internet and its related advertising channels. It is hypothesized in this study that the service product may have a tendency to rely more on online channels instead of offline channels in product sales. As indicated by the results, the differences between FMCG and service product advertising remain unchanged in which FMCG should continue relying on offline advertising on product sales but service type of business should use offline channels with cautious while combining it with online channels.

With regarding the web conversation effect on sales, results rejected hypothesis 3 for both FMCG and service sales indicating that there is no positive effect, which is expected for FMCG product because it uses mostly offline channels so this may be the case. Interestingly, the web conversation had significantly negative effect on service product and one of the possible explanations is that consumers only make service product a topic, such as an electricity product, when there is a problem, e.g., someone may twit about some unhappiness about the electricity business due to a shutdown. And thus this may result in negative impact from web conversation. Similar results are reported by Lewis and Reiley (2014) that increasing online advertising budget has hardly any impact on offline sales in which 93% of the sales is in offline stores while only 20% of the increase is attributed to customers who click on the advertising. This repor t confirms that what this study found has theoretical support from literature. In this case, both FMCG and service should consider reducing their advertising budgets on web conversation since this buzz did not assist both product sales but instead had negative impact on service product sales, and invest more in social media to avoid negative spillover effect.

(29)

29 Lu, 2013). Third, advertising position on conversion rate has both positive and negative effect (Rutz et al., 2012). For example, improved position is followed by improved conversion rates from keywords and click through in which about two-third of total conversion is derived from increased conversion rates, and only one-third is derived from increased click-through. Furthermore, it is found that click-through is unnecessarily the most profiting factor since higher positions of keywords result in higher conversion rates, and yielded profit can be higher even when keywords are at middle position (Ghose and Yang, 2009). Combined together, there are many possible reasons that may result in no effect from online advertising, which can be a combination of multiple factors including competition, position, advertising amount and keywords used. One support from literature is that less popular keywords are interesting to consumers to search for more information with higher tendency to generate purchase behaviors (Jerath et al., 2013).

(30)

30 time to establish brand concept, and web conversation may not be suitable for short term brand awareness either.

Finally, results rejected the hypotheses 7, 8 and 9 indicating that none of the cross effect from web conversation, online advertising and online expenditures had positive impact on sales, in which web conversation cross-effect had negative effect on service sales with partial significance. Even though the combined effect of web conversation and cross effect is negative, it is smaller than the effect from web conversation alone. Therefore, offline expenditures can be a damage control to negative web conversation. Most of the results from cross effect were unexpected because based on the literature it is expected that online and offline advertising and sales have impact cross channels effectively (Dinner, Heerde, and Neslin, 2014); and offline advertising, online advertising and direct mail all have positive impact on offline sales when multichannel impact was tested to consider their dependent variable in automobile sales (Naik and Peters, 2009), in which most customers do research online in contrast to the product categories in this study. Other supportive evidence include Wiesel et al. (2011) who found significant effect between advertising channels with very high sales from online advertising and much lower sales from flyers and faxes advertising; and Wiesel and colleagues showed that almost three quarters of Google AdWords profit is attributed to offline sales; which support the importance of connectivity and synergistic effect from cross channels between online and offline channels. It is possible that these literature reports chose different product categories, such as clothing products, that may be more sensitive or reactive to cross channel effect compared to this study. But this study did observed differences between cross channel effect on FMCG and service products although those differences had not significance statistically.

(31)

31 are many possible reasons that may result in no effect from online advertising on either product sales, and it can be a combination of multiple factors including competition, position, advertising amount and keywords used, which should be factors considering when business is making decisions on their marketing strategies. Similar to the literature review, this study indicated that short-term advertising is not enough to generate brand awareness and web conversation may not be suitable for short term brand awareness establishment. Finally, product category choice seems important to product advertising and marketing study because some product or service can be more sensitive or reactive from certain channels of advertising and cross-channel interactions.

8. Managerial Implications

Enterprise businesses need to be aware of advertising channels that can be chosen with carefulness when reaching marketing decision making steps. It is the purpose of this study to offer useful suggestions and recommendations to businesses especially those with FMCG a nd service products. Based on this study, there are four managerial implication based on the finding as the following.

First, it is clear that FMCG businesses should consider offline advertising as their major advertising target for product sales and budget distribution. On the other hand, traditional advertising channels should be considered spending fewer budgets or taken as compensational advertising channels in addition to offline channels in general. For those businesses carrying service product, even though offline advertising was found not significant it is still one of the most important advertising channels. The reason may be because service product does not rely on offline but online advertising and because offline advertising has decreasing impact on sales in recent years.

(32)

32 there are complaints and negative reactions, and thus web conversation may be avoided or invest more in social media and offline advertising expenditure to reduce the negative effect.

Third, results indicated that position and competition might play a crucial role in the success of online advertising effectiveness therefore companies should also include these components in their advertising planning. There are many factors that may result in no effect from online advertising. These multiple factors including competition, position, advertising amount and keywords used which need to be considered thoroughly when businesses are planning on their advertising channels and budgets.

Finally, it is important for enterprise businesses to know that different product categories can yield different results from the same advertising channels, and some products may be more sensitive or reactive to certain advertising channels or cross channel effect while other products may not. In addition, this study recommended that businesses consider excluding or including innovative advertising channels in their marketing planning; for example, service businesses should do better in damage control of web conversation since it is affecting them negatively to determine future advertising plan for their products. Therefore, these businesses should focus more on social media and other online advertising to avoid negative impact or profit loss and to take advantages of suitable advertising channels. Before advertising decision making, enterprise businesses may set up test in advertising to measure factors of position, competition, advertising amount and keywords with their impact on online advertising to determine suitable factors for each specific product in certain markets.

Overall, business companies should consider a better budget distribution plan by fully evaluating of single and cross advertising channels with their impact based on company product, technology and budget with awareness of factors affecting advertising impact and advertising effectiveness. Tests and trials are recommended for enterprise businesses to perform before distributing advertising budget and making strategy plan to avoid budget loss and increase chances of profiting.

9. Limitations and future research

(33)

33 from this study to other countries, especially those countries not located in Eastern Europe, such as in North America, Asian countries and other areas. Recommendations and suggestions from the study should not be generalized to any products and any markets but should be taken with sufficient consideration and caution to product and region specifications.

Second, this study selected one single product from each product category of FMCG and service which was yogurt in FMCG business and electricity in service business, and thus the information and knowledge learned from this study may not be suitable to be applied to any FMCG and service products other than yogurt and electricity. Application of the findings from this study is possible but recommended with full consideration of product and market specificities and evaluation of effectiveness and practicality before distributing major budget and making decision.

In addition, several factors were not considered in this study but may have impact on results and data analysis. For instance, it did not include competition factors for offline and online advertising which based on literature review might have significant effect on advertising efficacy and impact of product sales; the offline and online advertising examined could be broke down to more variables to obtain results with better accuracy; data obtained covered a period of two years which might need a larger timeframe to draw more accurate conclusions; and the demographics of this study did not include younger customer populations in which internet might be more significant to yield different results.

Future research should consider extended product categories and research regions to collect data from a wider range of products from areas covering several countries and regions to maximize data accuracy and applicability. For example, products under categories of different positions, competitions, advertising amount and keywords can be considered to enhance meaning and applicability of results. Furthermore, the offline and online advertising can be broken down to different variables to assist better results with higher accuracy and boarder applicability.

(34)

34

References

Abraham M. (2008). The Off-Line Impact of Online Ads. Harvard Business Review. April, 28.

Agarwal, A., Hosanagar, K., & Smith, M. D. (2011). Location, location, location: An analysis of profitability of position in online advertising markets. Journal of marketing research, 48(6), 1057-1073.

Agarwal, M. K., & Rao, V. R. (1996). An empirical comparison of consumer-based measures of brand equity. Marketing letters, 7(3), 237-247.

Allen, C., Yaeckel, B., & Kania, D. (1998). Internet world guide to one-to-one web marketing. John Wiley & Sons, Inc.

Assmus, G., Farley, J. U., & Lehmann, D. R. (1984). How advertising affects sales: Meta -analysis of econometric results. Journal of Marketing Research, 65-74.

Baron, R. M., & Kenny, D. A. (1986). The moderator–mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of personality

and social psychology, 51(6), 1173.

Barreto, A. M. (2014). The word-of-mouth phenomenon in the social media era. International

Journal of Market Research, 56(5), 631-654.

Braun, M., & Moe, W. W. (2013). Online display advertising: Modeling the effects of multiple creatives and individual impression histories. Marketing Science, 32(5), 753-767.

Chang, Y., & Thorson, E. (2004). Television and web advertising synergies. Journal of

Advertising, 33(2), 75-84.

(35)

35 Clark, C. R., Doraszelski, U., & Draganska, M. (2009). The effect of advertising on brand awareness and perceived quality: An empirical investigation using panel data. QME, 7(2), 207-236.

Danaher, P. J., & Dagger, T. S. (2013). Comparing the relative effectiveness of advertising channels: A case study of a multimedia blitz campaign. Journal of Marketing Research, 50(4), 517-534.

Draganska, M., Hartmann, W. R., & Stanglein, G. (2014). Internet versus Television Advertising: A Brand-Building Comparison. Journal of Marketing Research, 51(5), 578-590.

Dijkstra, M., Buijtels, H. E., & van Raaij, W. F. (2005). Separate and joint effects of medium type

on consumer responses: a comparison of television, print, and the Internet. Journal of Business

Research, 58(3), 377-386.

Dinner, I. M., Van Heerde, H. J., & Neslin, S. A. (2011). Driving Online and Offline Sales: The Cross-Channel Effects of Digita l versus Traditional Advertising. 2011 Marketing Science Conference, November.

Dinner, I. M., Van Heerde, H. J., & Neslin, S. A. (2014). Driving online and offline sales: The cross-channel effects of traditional, online display, and paid search advertising. Journal of

Marketing Research, 51(5), 527-545.

Ghose, A., & Yang, S. (2009). An empirical analysis of search engine advertising: Sponsored search in electronic markets. Management Science, 55(10), 1605-1622.

Glasner M. (2015). The future of multichannel marketing: marketer and consumer perspectives.

(36)

36 Hayes, A. F. (2013). Introduction to mediation, moderation, and conditional process analysis: A regression-based approach. Guilford Press.

Hausmann, A. (2012). Creating ‘buzz’: opportunities and limitations of social media for arts institutions and their viral marketing. International Journal of Nonprofit and Voluntary Sector

Marketing, 17(3), 173-182.

Huang, R., & Sarigöllü, E. (2012). How brand awareness relates to market outcome, brand equity, and the marketing mix. Journal of Business Research, 65(1), 92-99.

IAB. (2014). IAB Internet Advertising Revenue Report. IAB.

Ibeh K., Luo Y., Dinnie K. (2005). E-branding strategies of internet companies: Some preliminary insights from the UK. Brand Management, 12(5): 355–373.

Jerath, K., Ma, L., & Park, Y. H. (2014). Consumer click behavior at a search engine: The role of keyword popularity. Journal of Marketing Research, 51(4), 480-486.

Kumar, V. (2015). Evolution of Marketing as a Discipline: What has happened and what to look out for. Journal of Marketing, 79(1), 1-9.

Laroche, M., Kiani, I., Economakis, N., & Richard, M. O. (2013). Effects of multi-channel marketing on consumers' online search behavior: The power of multiple points of connection. Journal of Advertising Research, 53(4), 431-443.

Liaukonyte J., Teixeira T., Wilbur K. C. (2014). Television Advertising and Online Shopping.

Marketing Science, November.

(37)

37 Luo, X., & Zhang, J. (2013). How Do Consumer Buzz and Traffic in Social Media Marketing Predict the Value of the Firm? Journal of Management Information Systems, 30(2), 213-238.

Manchanda, P., Dubé, J. P., Goh, K. Y., & Chintagunta, P. K. (2006). The effect of banner advertising on internet purchasing. Journal of Marketing Research, 43(1), 98-108.

Nelson-Field, K., Riebe, E., & Sharp, B. (2012). What's Not to" Like?" Can a Facebook Fan Base Give a Brand The Advertising Reach It Needs? Journal of Advertising Research, 52(2), 262-269.

Preacher, K. J., & Hayes, A. F. (2008). Asymptotic and resampling strategies for assessing and comparing indirect effects in multiple mediator models. Behavior research methods, 40(3), 879-891.

Putrevu, S., & Lord, K. R. (2003). Processing Internet communications: A motivation, opportunity and ability framework. Journal of Current Issues & Research in Advertising, 25(1), 45-59.

Ratchford B.T. (2015). Some Directions for Research in Interactive Marketing. Journal of

Interactive Marketing, 29: v–vii.

Rutz, O. J., & Bucklin, R. E. (2011). From generic to branded: A model of spillover in paid search advertising. Journal of Marketing Research, 48(1), 87-102.

Rutz, O. J., Bucklin, R. E., & Sonnier, G. P. (2012). A latent instrumental variables approach to modeling keyword conversion in paid search advertising Journal of Marketing Research, 49(3), 306-319.

Sethuraman, R., Tellis, G. J., & Briesch, R. A. (2011). How well does advertising work? Generalizations from meta-analysis of brand advertising elasticities. Journal of Marketing

(38)

38 Simpson, H. B., Maher, M. J., Wang, Y., Bao, Y., Foa, E. B., & Franklin, M. (2011). Patient adherence predicts outcome from cognitive behavioral therapy in obsessive-compulsive disorder.

Journal of Consulting and Clinical Psychology, 79(2), 247.

Taylor, J., Kennedy, R., McDonald, C., Larguinat, L., El Ouarzazi, Y., & Haddad, N. (2013). Is the Multi-Platform Whole More Powerful Than Its Separate Parts? Measuring the Sales Effects of Cross-Media Advertising.Journal of Advertising Research, 53(2), 200-211.

Von Hippel, C., Issa, M., Ma, R., & Stokes, A. (2011). Stereotype threat: Antecedents and consequences for working women. European Journal of Social Psychology, 41(2), 151-161.

Wiesel, T., Pauwels, K., & Arts, J. (2011). Practice Prize Paper-Marketing's Profit Impact: Quantifying Online and Off-line Funnel Progression. Marketing Science, 30(4), 604-611.

Yang, S., Lu, S., & Lu, X. (2013). Modeling competition and its impact on paid-search advertising. Marketing Science, 33(1), 134-153.

(39)

39

APPENDIX A. Variable correlations and multiple regression analysis results

(40)

(41)

41 CORRELATION TEST FOR FMCG PRODUCT CATEGORY

CORRELATION TEST FOR FMCG SERVICE CATEGORY

(42)

(43)

(44)

44

APPENDIX B. Process-macro results

RESULTS FOR FMCG PRODUCT CATEGORY

Run MATRIX procedure:

**************** PROCESS Procedure for SPSS Release 2.13.1 ************** Written by Andrew F. Hayes, Ph.D. www.afhayes.com

Documentation available in Hayes (2013). www.guilford.com/p/hayes3 ************************************************************************** Model = 4 Y = D_sales X = D_expe M = D_brand Statistical Controls: CONTROL= D_web_cn D_onl_ad Sample size 105 ************************************************************************** Outcome: D_brand Model Summary R R-sq MSE F df1 df2 p ,4962 ,2462 15,0081 10,9961 3,0000 101,0000 ,0000 Model

coeff se t p LLCI ULCI constant 22,3276 ,6259 35,6719 ,0000 21,0860 23,5693 D_expe ,0000 ,0000 ,3594 ,7201 ,0000 ,0000 D_web_cn ,1200 ,0223 5,3838 ,0000 ,0758 ,1642 D_onl_ad -,0167 ,0078 -2,1498 ,0340 -,0322 -,0013 ************************************************************************** Outcome: D_sales Model Summary R R-sq MSE F df1 df2 p ,2791 ,0779 3990143483 2,1116 4,0000 100,0000 ,0849 Model

coeff se t p LLCI ULCI constant 342609,700 37635,5744 9,1033 ,0000 267941,646 417277,754 D_brand -2074,6684 1622,4481 -1,2787 ,2040 -5293,5656 1144,2287 D_expe ,4787 ,1857 2,5773 ,0114 ,1102 ,8472 D_web_cn -159,1259 412,1703 -,3861 ,7003 -976,8617 658,6099 D_onl_ad 14,8828 129,8064 ,1147 ,9089 -242,6499 272,4155 ******************** DIRECT AND INDIRECT EFFECTS ************************* Direct effect of X on Y

Effect SE t p LLCI ULCI ,4787 ,1857 2,5773 ,0114 ,1102 ,8472 Indirect effect of X on Y

Effect Boot SE BootLLCI BootULCI D_brand -,0085 ,0250 -,1034 ,0210 Normal theory tests for indirect effect

Effect se Z p -,0085 ,0307 -,2764 ,7822

******************** ANALYSIS NOTES AND WARNINGS ************************* Number of bootstrap samples for bias corrected bootstrap confidence

intervals: 1000

Is multichannel advertising worth it? A comparison between fast moving consumer goods and service product categories