1 University of Groningen
Faculty of Economics and Business
Master Thesis Marketing Management & Marketing Intelligence
Name Student: Shamilee Brunken Student ID number: 1811029
Student email: s1811029@student.rug.nl Telephone: +316-24366746
Date Thesis: June 23, 2014
Name Supervisor: Dr. Ir. Maarten Gijsenberg
2 Drugstores and Pharmacies: The Potential Competition due to Over-The-Counter Products
Abstract
This paper investigates whether drugstores are a potential competitive threat to pharmacies on the Over-The-Counter (OTC) medication market. Influential factors are the rising need for convenience, increasing customer informedness, and changes in the marketing mix. Based on the inferences drawn, it is suggested that pharmacies ought to reexamine their current business model for OTC product and either find other ways to maintain competitiveness in the market, or solely focus on prescription drugs.
Keywords: Pharmacy; Drugstore; Competition; Customer Informedness; Convenience; Marketing Mix
3
Table of content
Introduction ... 4
Theoretical Framework ... 5
Customer Informedness ... 5
Need for convenience ... 5
The marketing mix ... 6
Data ... 11
Methodology ... 12
Choice-based conjoint analysis ... 12
Logistic regression ... 13
Results ... 14
Conjoint analysis ... 14
Logistic Regression ... 16
Comparing the analyses ... 19
Discussion ... 20
Managerial Implications ... 21
E-pharmacy ... 22
Business Concept ... 22
Data Analysis & Discussion ... 23
Conclusion ... 25
Limitations & Recommendations ... 26
Bibliography ... 27
Appendix I: Choice-Based Conjoint Design ... 30
Appendix II: Survey ... 31
Appendix III: Factor Analysis ... 42
Price ... 42
Product ... 46
Place ... 49
Promotion ... 52
People ... 55
Customer Informedness ... 58
Appendix IV: Detailed Descriptives of E-pharmacy Poll ... 59
4
Introduction
Over the years, the amount of self-medicating products in the Netherlands has increased (Stichting Farmaceutische Kengetallen, 2006; KNMP, 2014). These self-medicating products are called Over- The-Counter (OTC) drugs, which are medications that consumers can purchase without the need of a prescription and at their own discretion (Anderson, 1971; Creyer, 2001; Ho, Mursch, Ong, & Peittula, 1998; Kim & King, 2009). OTC drugs provide temporary relief to simple discomforts and are considered safe enough to be purchased by consumers at will (Kim & King, 2009). The market for OTC products has increased by approximately 27 percent over the last 10 years (Neprofarm, 2013).
This growth has largely been stimulated by the changes in legislation regarding point of purchase of OTC product
1. Since these changes were made in 2007, the amount of products that fall within the UAD- and AV-category has also increased (CBG-MEB, 2013), and thus enlarging the supply on the market from other sources than just pharmacies. The pharmacy now has to “share” its customer base, not only with the drugstores, but also with all other retail outlets willing and able to sell AV- category OTC products. In 2013, drugstores held 75 percent of the market share in OTC products while pharmacies have over the years declined to 14 percent (Neprofarm, 2013). In addition to the legal changes, demographic and technological changes – such as the need for convenience and knowledge empowered consumers due to the internet – are inferred to further influence the function pharmacies hold within the Dutch communities (Anderson, 1971; Clemons, 2008; Kim &
King, 2009).
Due to all these changes, the question arises whether pharmacies are still a competitive force on the OTC market. Other than a faint acknowledgement from Van Mill (2005), literature regarding the dynamics in the Dutch OTC market has not been found. This study investigates whether drugstores can be considered a viable alternative to pharmacies, and aims to identify the factors that influence change in consumer behavior regarding preferred purchase location of OTC drugs. This study is a preliminary investigation that signals a larger problem for pharmacies requiring further research.
1Article 1 section s.-u. of the Dutch “Geneesmiddelenwet” (translation: Medication law) identifies three types of OTC
products: those that can only be bought at the pharmacy (UA-category), those that can only be bought at the pharmacy and drugstores (UAD-category), and those that can be bought at any retail outlet (AV-category) (Overheid.nl, 2007; CBG-MEB, 2008; KNMP, 2014). The category in which a pharmaceutical drug is placed is based on the potential risk of harm that can be done by misuse of the drug, where AV-category drugs pose minimal risk. In addition, the dosage and package size are lower in the AV-category. Although the AV-category is the main catalyst for change in the OTC market, and other retail chains besides drugstores and pharmacies can also sell these products, this study only focuses on drugstores as the main potential competitor of pharmacies. The fundamental reason for this focus is that only drugstores contain the broad product range of OTC drugs, since they are allowed to sell both the UAD- and AV-category drugs.
5
Theoretical Framework
The change in the legislation and categorization of self-medicating products is suggested to have altered the competitive position of pharmacies. However, the change in legislative rules does not fully explain a possible shift in consumer behavior. It can be argued that the degree of customer informedness and the need for convenience contribute to the preference of drugstores instead of pharmacies as a source for OTC product. These factors influence the elements of the marketing mix of both pharmacies and drugstores; however, mostly in favor of the drugstores.
Customer Informedness
Clemons (2008) argues that there was a change in consumer’s decision-making process that resulted in a change in competitive strategies of firms. Moreover, he argues that this change is mainly due to the information that is now readily available to consumers. This “informedness”, mostly caused by the reduction in search costs due to the internet and the recommendation systems between consumers though collaborative filtering, creates empowered individuals, which will play an important role in how firms determine their corporate strategies. Consumers can now easily search for additional information regarding their health and medication (Ricks & Mardanov, 2012). The need for information is thus largely reduced once the consumer reaches the pharmacy. This undermines one of the core competences of pharmacies, which is extensive knowledge. Moreover, drugstore employees are also required to have basic knowledge regarding every OTC product in the store (Overheid.nl, 2007; InfoNu, 2014), which is often sufficient for the (mostly already informed) customer.
Another argument suggesting a lower need for the superior knowledge of pharmacies is that OTC products are usually generic and well-established drugs that are used relatively frequent. Through usage, the customer gains knowledge regarding the drug, diminishes risk, and reduces the need for information at the next point of purchase (Coupey & Narayanan, 1998; Ho et al., 1998).
Therefore, the knowledge endowment of customers gained by the development of information technology and familiarity through product usage emphasizes a possible need for re-evaluation of the competitive advantages of pharmacies (Clemons, 2008; Ho, Mursch, Ong, & Peittula, 1998;
Kumar & Reinartz, 2012).
Need for convenience
Over the past decades it has been evident that time has become a scarce commodity for many households. This increase in time scarcity created a need for convenience (Kumar & Reinartz, 2012).
Where pharmacies only sell medication and occasionally a limited assortment of personal care
products, a drugstore provides a broader range of products and categories. With a single stop, a
6 consumer can purchase items from multiple categories – so called basket shopper (Cachon & Kök, 2007) or “one-stop-shopper” (Corstjens & Corstjens, 1999). This does not only save a customer another trip to the pharmacy, it also saves time standing in line; which is often a side effect of the extensive service – in the form of attention – provided at pharmacies. Moreover, often there are more drugstores nearby than pharmacies. The consumer costs in terms of time and perceived physical costs are smaller when visiting the nearest store where OTC products are sold (Corstjens &
Corstjens, 1999), therefore making a drugstore a more favorable option. Anderson (1971) calls these phenomena convenience-orientated consumption. He argues that there has been a shift in consumer attitudes from valuing products and serviced to valuing time (Anderson, 1971). Moreover, the socio- demographic trend of time efficiency (Berry, Seiders, & Grewal, 2002; Kumar & Reinartz, 2012) and the need for convenience alongside the increase in information accessibility, has led to an increase in self-medicating by means of the OTC drugs (Creyer, 2001). Thus, this change could cause a shift of the pharmacies’ OTC customer base towards drugstores.
The marketing mix
In order to infer the potential areas of change in the strategy of pharmacies regarding OTC products, the marketing mix elements will be discussed. In addition to the traditional four P’s – price, product, place, and promotion – the element “people” has been added. The main reason for this supplement is based on the notion that the current competitive advantage of pharmacies relies on the knowledge of the employees; thus making them a valuable component in the marketing mix. Each of the marketing mix elements will provide a basis for a hypothesis that will serve to answer the question whether pharmacies are losing their edge on the OTC market.
Price
A part of the marketing strategies of Dutch drugstores is price promotions. Such promotions usually alter on a weekly basis and highlight a certain product within a selected category. Price promotions, like “two for one” or “three for … Euros”, are meant to stimulate the sales of a specific product or a category as a whole (Blattberg & Neslin, 1989). OTC products are no exception to this strategy.
However, pharmacies do not participate in (frequent) sales promotions as drugstores do. And if they do, they do not communicate such promotions other than in-store – as will be argued later on.
Therefore, the presence of (widely communicated) price promotions at drugstore might draw
“bargain hunters” to drugstores instead of pharmacies.
Another factor that might influence consumer behavior is the notion of stockpiling. According to
Mela, Jedide & Bowman (1998) stockpiling refers to purchasing large quantities of a certain product
and/or shifting purchase times in order to buy before the expected time of the next purchase. It is
argued that stockpiling behavior can arise from price promotions, which might harm brand sales in
7 the end (Assunção & Meyer, 1993; Helsen & Schmittlein, 1992; Blattberg & Neslin, 1989; Mela et al., 1998). However, in this particular case, stockpiling behavior might pose as an incentive to purchase OTC products at the drugstore rather than at the pharmacy. The anticipation of a familiar and consistent cycle of sales promotions – the lying-in-wait heuristic (Mela, Jedidi, & Bowman, 1998) – might deflect the consumer from purchasing such products elsewhere. Ignoring the profitability implications on drugstores, consistent price promotions could be regarded as an argument for a change in consumer behavior in favor of drugstores.
Hypothesis 1: Price promotions negatively influences the preference towards purchasing OTC products at the pharmacy
Promotion
In the Netherlands, it is against the law to promote prescription drugs to the masses. However, it is allowed to promote OTC products, given certain conditions (IGZ, 2014). Drugstores therefore make use of in-store promotions as well as other advertising methods – e.g. flyers and TV commercials (see Figure 1). Pharmacies, in general, do not advertise promotions like drugstores do. Should they have price promotion on OTC products, they only advertise this in-store. Other media, such as TV commercials, only serve to further stimulate the need for personal health care and promote the quality of a certain pharmacy chain. Pharmacies do not view themselves as “stores” or shops, but rather as healthcare professionals (Van Mil, 2005). Therefore, their advertising focus is different from
Figure 1: Advertisement OTC Products from Dutch drugstore chain Kruitvat
OTC
Products
8 that of drugstores. Thus, only customers who have been or at the pharmacy during the period of in- store promotions will be aware of any price promotion offered by pharmacies. Therefore, customers attracted to price promotions will most likely go to drugstores instead of pharmacies for OTC products, simply because they are more aware of promotional activities of pharmacies.
Hypothesis 2: Advertising activities of drugstores negatively influences the preference towards purchasing OTC products at the pharmacy.
Product
The UAD- and AV-category drugs are fairly homogeneous across stores and pharmacies. Next to the A-brand products, pharmacies and drugstores also sell private label products. These private labels are mostly common OTC products, such as aspirin or vitamins. These products are familiar generic pharmaceutical drugs and are thus regarded as homogeneous. Along with UAD- and AV-category self-medication, both drugstores and pharmacies provide vitamin emulators, first aid products, and beauty care products. However, the product category range in drugstores extends beyond these five.
Moreover, even within some of these five categories, for instance beauty care, drugstores provide a much broader assortment. Based on the notion that consumers nowadays behave as basket shopping consumers – consumers who desire to purchase from multiple categories – it can be expected that drugstores are preferred over pharmacies (Cachon & Kök, 2007). Furthermore, findings of Baumol & Idle (1956) suggest that the probability of selecting a certain retail store is positively related to its assortment size. The greater the number of variety in items within a store, the greater, ordinarily, is the customer’s expectation of the shopping trip to be successful in finding all the items needed within one stop. Briesch et al. (2009) have also supported these findings.
Moreover, the physical cost involved in making an additional trip to another store to purchase the products from remaining categories might be significant enough to demotivate consumers from purchasing OTC products at the pharmacy (Corstjens & Corstjens, 1999; Baumol & Ide, 1956). This provides further support for the inference that the multi-category assortment range of drugstores invokes the preference among consumers over pharmacies.
Hypothesis 3: The assortment range of drugstores negatively influences the preference towards purchasing OTC products at the pharmacy.
Place
Ataman, Van Heerde & Mela (2010) found that distribution coverage is one of the most fundamental
marketing mix elements for the success of a brand. Consumers need to be able to find the product in
order to purchase it. Similarly, it can be argued that by extending the availability of OTC products to
drugstores, the distribution coverage of these products increases. Proximity, then, to either a
pharmacy or drugstore can influence the decision regarding purchase location (Baumol & Ide, 1956).
9 In addition, Reilly’s Law of Retail Gravitation
2argues that the probability of selecting a certain retail store is inversely related to the distance from a consumer’s home (Briesch et al., 2009; Brown, 1989;
Reilly, 1931). Moreover, the need for time efficiency and convenience also suggest that consumers would prefer the nearest available location since there are physical cost involved when going elsewhere (Corstjens & Corstjens, 1999; Baumol & Ide, 1956).
Hypothesis 4: The relative proximity to a drugstore (pharmacy) negatively (positively) influences the preference towards purchasing OTC products at the pharmacy.
People
As has been mentioned before, the extensive knowledge of pharmacists and other pharmacy employees is the main competitive advantage. The findings of Creyer, Hrsistodoulakis & Cole (2001) emphasize the importance of this knowledge in stating that (inter)personal experience and marketing stimuli are primary sources of information when choosing to self-medicate. However, Kim
& King (2009) suggest that interpersonal sources and mass media are the main sources of information for non-prescription – OTC – drugs rather than pharmacists or the internet, and that internet sources are considered more important for prescription drugs. Nonetheless, the findings of Rick & Mardanov (2012) emphasizes that consumers rely more on the recommendations of pharmacists when it comes to purchasing OTC products. This could be attributed to the information overload customers face during their private search for information. The expertise of the pharmacist could clear up misunderstandings and confusion that might have occurred (Ricks & Mardanov, 2012).
However, pharmacy employees are not the only in-store source of knowledge. Drugstore personnel are also forced by the Dutch law to contain a certain level of knowledge regarding any OTC drugs sold in their store (InfoNu, 2014; Overheid.nl, 2007). Thus, it can be argued that simple confusion or misunderstandings can also be clarified at the drugstore.
Despite the obvious trivia regarding the relevance of additional information, it is advocated that the extensive knowledge service provided by pharmacies can still attract consumers to pharmacies for OTC products. The main reason for this is the fact that information is the main competitive advantage of pharmacies. Whether the strength of this advantage is negligible has yet to be determined.
Hypothesis 5: Knowledge and expertise regarding OTC drugs of pharmacy personnel positively influences the preference towards purchasing OTC products at the pharmacy.
2Reilly’s Law of Retail Gravitation: “two cities attract trade from an intermediate town in the vicinity of the breaking point approximately in direct proportion to the populations of the two cities and in inverse proportion to the squares of the distances from these two cities to the intermediate town.” – W. J. Reilly (1931)
10
Table 1: Marketing Mix Summary
Pharmacy Drugstore
Price Infrequent and unpredictable
sales promotions
Frequent price promotions Product Limited additional products Wide range and variety of
additional products
Place Low distribution coverage/ Few
in large radius
Large distribution coverage/
Many in small radius
Promotion If present, only promoted in-
store
Multiple media promotion
People Extensive knowledge Basic knowledge
Customer Informedness as a moderator
Next to the hypotheses resulting from the marketing mix, which mostly speculates consumer behavior based on the characteristics of drugstores and pharmacies, additional hypotheses are needed in order to capture the influence of customer informedness. As has been mentioned, the internet allows customers to search for knowledge on their own. Experience with certain OTC products also increases knowledge and reduces risk (Ho, Mursch, Ong, & Peittula, 1998). The gain in knowledge undermines the competitive advantage of pharmacies, which is based on the customer’s lack of knowledge and experience. Therefore, customer informedness is suggested to have negative influence the preference towards purchasing OTC product s at a pharmacy rather than at a drugstore. Building on this statement, low customer informedness can increase the need for additional advice from experienced and knowledgeable personnel at the point-of-purchase. As mentioned, pharmacies excel over drugstores in this area. Thus, it can be argued that low customer informedness positively moderates the effect of knowledgeable personnel in pharmacy toward the investigated choice preference. Thus, inversely, high customer informedness negatively moderates the effect. In the same manner, high customer informedness can amplify the influence of proximity on preference towards a point of purchase of OTC drugs. When individual already have knowledge and/or experience with the products, the need for convenience is assumed to take a higher priority and stimulates the customer to purchase OTC products at the nearest available point of sale.
Therefore, high customer informedness is argued to negatively moderate the influence of the relative proximity to a drugstore on the preference towards purchasing OTC drugs at a pharmacy rather than at the drugstore. The relationship is suggested to be positive is the nearest available store is a pharmacy.
Hypothesis 6a: Customer informedness negatively influences the preference towards purchasing
OTC products at the pharmacy.
11 Hypothesis 6b: Customer informedness negatively moderates influence of the extensive knowledge of pharmacy personnel on the preference towards purchasing OTC products at the pharmacy.
Hypothesis 6c: Customer informedness negatively moderates influence of the relative proximity on the preference towards purchasing OTC products at the pharmacy.
The conceptual model, based on the hypotheses, illustrates the suggested factors that influence the preference for purchasing OTC products at a pharmacy rather than at a drugstore. The expectation is that pharmacies are less preferred as a point of purchase for OTC products. Moreover, it is expected that proximity will be the most important factor of influence in the consumers’ preference choice.
Figure 2: Conceptual Framework
Data
The data used in this research is collected through a survey. The survey has been held only among Dutch citizens over the age of 18. The 176 respondents were randomly selected from different parts of the country. The survey is comprised out of three parts: a choice-based conjoint analysis, questions on multi-item Likert- scales (7 points), and questions regarding demographics. The survey can be found in Appendix II.
The conjoint analysis will be presented at the beginning of the survey in order to reduce the formation of strong directed opinions based on the multi-item scale questions. This is an attempt to
Product
Price Promotion
People Place
Preference for Pharmacy
-/+
- - -
+
Customer Informedness
-
- -
12 ensure the choices are made via primary intuition instead of elaborate reasoning. This way, it can also be analyzed whether the intuition and reasoning of consumers are compatible. Moreover, in order to stimulate cooperation and involvement from the consumers, it is suggested that simple and non-pervasive questions are posed at the beginning of a questionnaire (Malhotra, 2010). Moreover, the questions are composed as such to suppress directional biases. Furthermore, questions are often rephrased into opposite arguments in order to validate a certain answers.
Methodology
Two methods will be used to investigate the customer preferences towards either pharmacies or drugstore, namely a logistic regression and a choice-based conjoint analysis. The logistic regression serves to analyze the effects of the previously discussed conceptual model. In addition, the conjoint analysis serves to provide additional insights regarding the importance of certain elements within a decision-making process. Moreover, the analysis complements the logistic regression method.
People often tend to answer questionnaires not only based on what they would actually do, but also on what they perceive to be the “correct” answer. The choice-based conjoint analysis triggers responses that are more intuitive. This could filter out the perception bias that could occur in the survey. Therefore, it represents market behavior more directly (Huber et al., 1992).
Choice-based conjoint analysis
There are five attributes, which are presented in Table 2. The elements regarding the knowledge of employees and the influence of advertising has not been incorporated in the design, since no suitable allocation of attributes and levels could be formulated. This might cause the weights of the attributes to be overestimated, since the “pie” is divided among a smaller amount of factors. However, the analysis still provides an indication regarding the relative importance of the remaining attributes.
Table 2: Conjoint Attributes and Levels
Store type Assortment Price Promotions Relative distance Customer Informedness
Pharmacy OTC only no sale within 500
meters
no knowledge regarding OTC use Drugstore OTC & basic
beauty care
sale between 500 -
1000 meters
basic knowledge about OTC use OTC, Beauty care,
Personal Hygiene
unknown
1further than 1000 meters
extensive
knowledge about
OTC use
13
1unknown entails not knowing beforehand whether there is a sale or not
A full factorial design would entail presenting 81 choice sets. Due to the extensive size the survey would have with 81 choice sets, a balanced and orthogonal fractional factorial design is used. This design is illustrated in Appendix I. Moreover, minimal overlap and non-dominated choice sets are ensured in order to maximize the relevance of the collected data. Furthermore, a no-choice option will be implemented into each choice set. It is important to know whether consumers are actually driven by certain factors or indifferent.
Logistic regression
As mentioned, the logistic regression an analysis is based on the multi-item scale questions from the questionnaire. Since multiple questions are asked in order to capture the same phenomenon, a factor analysis has been conducted. The details regarding the composition of the factors can be found in Appendix III. The dependent variable is also measured on a 7-point Likert-scale; however, it was re-coded into a binary variable afterwards. The cutoff points were 1-4 for a value of zero, and 5- 7 for a value of one. The middle value (4) is assigned to zero, the value of one signals a preference towards pharmacies. Since indifference does not indicate a preference, this middle value does not contribute to the meaning of the value one.
The translation of the conceptual model into a mathematical form is as follows:
where:
P1 = Price P2 = Promotion P3 = Product
P4 = Place P5 = People
CI = Customer Informedness
Demographics of the survey participants are used as control variables. These demographics concern purchase frequency, gender, age category, working hours per week, and income range.
A concern with the presented model is the possibility of multicollinearity due to the moderators. For
this reason, two models will be estimated: a model with interaction effects, and a model without
interaction effects. The performance indicators of the models will determine the adequacy of each
model.
14
Results
Conjoint analysis
Figure 3 illustrates the relative importance of the attributes in the decision-making process of consumers regarding the point of purchase of OTC products. The most notable outcome is that it appears to be relatively irrelevant whether the store is either a pharmacy or a drugstore – only 2 percent of the importance in the decision-making process is allocated to the store type. This suggests that the attributes of the store rather than the store itself mostly influences the point-of-purchase decision. The consumer preferences can be deducted when looking at the utility of the attribute levels in Table 3.
Figure 3: Relative Importance Choice-Based Conjoint Analysis
Price & Promotion
It is only natural that consumers prefer price promotions. However, the results do not clearly distinguish between the need for a product being on sale or simply the preference. More precisely, whether a customer would only purchase OTC products when these are on sale or whether they just prefer purchasing OTC products on sale. For this reason, no support can be argued for Hypothesis 1.
On the other hand, the preference towards a sale is based on the knowledge of a sale beforehand.
Knowing beforehand that there is a sale indicates the presence of promotional activities of some kind. Since drugstores promote their on-sale products, consumers are able to know beforehand whether an OTC product is on sale or not. This in contrast to pharmacies, where most often one can only discover sales items in-store. Based on this deductive reasoning, support can be found for Hypothesis 2. However, since this support is derived indirectly, no conclusions will be based on this outcome.
2%
25%
16%
24%
33%
Store Type Relative Distance Price Promotions Product Assortment Customer Informedness
15 Table 3: Choice-Based Conjoint Analysis
Log-Likelihood(null) -2927.85
Log-Likelihood -1906.85
(Pseudo) R-Square 0.1484
Hitrate 60.84%
Attributes Utility
Product
OTC only -0.4251
OTC & basic beauty care 0.0062
OTC, Beauty care, Personal Hygiene 0.419
Relative distance
within 500 meters 0.4998
between 500 - 1000 meters -0.1005
further than 1000 meters -0.3993
Price Promotions
no sale -0.2436
sale 0.3269
unknown -0.0833
Customer Informedness
no knowledge regarding OTC use -0.686
basic knowledge about OTC use 0.196
extensive knowledge about OTC use 0.49
Store type
Pharmacy -0.0382
Drugstore 0.0382
No choice
0 0.3888
1 -0.3888
Product
The presence of only OTC products decreases the utility of consumers, whereas the broadest assortment range increases it. Noticeable is that the mid-range assortment size does not provide much additional utility, meaning that consumers are somewhat indifferent about the presence of beauty care products. Thus, it can be deducted that the presence of personal hygiene products are crucial with respect to preference. Moreover, pharmacies are not known for selling personal hygiene products in their overall assortment, and if they do, the range would be very limited. Therefore, support is found for Hypothesis 3.
Place
Regardless of the store type, the store that is further away generates less utility than the one nearby.
This suggests that proximity is indeed of influential importance, and in accordance with Hypothesis 4.
16 Customer informedness
Figure 3 already illustrated that the degree of knowledge is the most important feature in the decision-making process. In addition, the results in Table 3 show that people prefer to have (extensive) knowledge beforehand. However, preference towards prior knowledge does not suggest a preference to either a drugstore or a pharmacy. Therefore, no support is found for Hypothesis 6a.
Store type
Even though the overall relevance of store type is extremely low, there is still a small preference towards drugstores. A reason for this could be that consumers are already familiar with drugstores and know that the other, and relatively more important, features are always present. However, this reason is solely based on speculation.
Logistic Regression
The results of the Logistic Regression analysis are presented in Table 4. As can be seen, the control variables are not included in de analysis. The main reason for excluding the demographic control variables from the model is that they did not improve the performance of the model. Moreover, none of these variables appeared to have any significant influence on the dependent variable. For this reason, they are excluded from the model. However, the variable regarding gender is used in order to create two segmented models. This is done for illustrative purposes regarding the possible presence and effect of multicollinearity, which will be discussed later on.
According to the results, the variables Place and People are the only two significant variables. Firstly,
Place represents the factor that captures the relative proximity to a drugstore. The results suggest
that there is a negative relationship between the relative proximity to a drugstore and the probability
of preferring to purchase OTC products at the pharmacy. The opposite holds true should a relative
proximity to a pharmacy increase – i.e. a negative relative proximity to a drugstore. Therefore, this
finding supports Hypothesis 4. Secondly, the variable People represents the factor that captures the
importance placed on the knowledge and expertise of pharmacy personnel regarding OTC drugs. The
results indicate that the more this knowledge and expertise are regarded as valuable, the higher the
probability of preferring to purchase OTC products at the pharmacy. This is in accordance with
Hypothesis 5, and is thus supported. Hypotheses 1, 2, 3, and 6a do not find any support in this
analysis.
17 Table 4: Logistic regression Analysis
Pooled Male Female Pooled Male Female
Constant -1.6791*** (0.5089) -1.0792 (0.7043) -2.3152*** (0.8145) -1.8657*** (0.6011) -2.4814** (1.2110) -2.1475 (0.8403) Price 0.2070 (0.2537) 0.2201 (0.3453) 0.4386 (4.0786) 0.1774 (0.2565) -0.0990 (0.3780) 0.4159 (0.4145) Product -0.3055 (0.2132) -0.3460 (0.2886) -0.2059 (0.3517) -0.2774 (0.2154) -0.3770 (0.3213) -0.2134 (0.3551) Place -0.5229*** (0.1970) -0.5302** (0.2536) -0.6456* (0.3465) 0.1696 (0.4427) 0.9363 (0.7363) -0.4937 (0.6884) Promotion -0.3718 (0.2534) -0.4345 (0.4129) -0.3114 (0.3466) -0.2822 (0.2584) -0.2493 (0.4448) -0.2770 (0.3578) People 0.8432*** (0.2353) 0.7400** 0.3039 0.9156** (0.3705) 1.5841** (0.6745) 3.0460** (1.2935) 0.3468 (0.9936) Customer
Informedness
0.0784 (0.1018) 0.001 (0.1400) 0.1496 (0.1635) 0.1044 (0.1190) 0.2025 (0.2123) 0.1087 (0.1772)
Customer Informedness*
Place
-0.1641* (0.0977) -0.3092** (0.1540) -0.0361 (0.1517)
Customer Informedness*
People
-0.1386 (0.1240) -0.4071* (0.2221) 0.1159 (0.1920)
Chi-Square 32.71*** 16.14** 15.35** 37.38*** 24.97*** 15.71**
(Pseudo) R-Square
0.1635 0.1626 0.1589 0.1868 0.2515 0.1630
Log-
Likelihood(null)
-100.0545 -49.6481 -48.1961 100.0545 -49.6481 -48.1961
Log-Likelihood -83.6983 -41.5759 -40.5366 -81.36 -37.1621 -40.3387
BIC 203.59 113.6488 113.168 209.2626 113.5347 121.942
Note: significance indicators: *** p< 0.01; ** p<0.05; * p< 0.1 . The coefficients are presented in odds; exp(β). The standard errors are given in parentheses. The significance of the model is calculated by the Chi-Square Statistic.
18 When looking at the constants in the models, one can see that they are always negative. This indicates that by default consumers do not prefer a pharmacy to a drugstore. Consumers will only prefer to purchase OTC products at pharmacies if the knowledge and expertise are highly valued and/or the drugstore is too far away.
There are no differences in the significance of the variables with the pooled model when segmenting for males and females. Note, however, that for females, both variables have stronger influence on the dependent variables than males do.
Multicollinearity
Before proceeding with the results of the full model (i.e. including the moderators), the concerns regarding multicollinearity mentioned before need to be addressed. Table 5 presents the Variance Inflation Factors (VIF’s) for the variables in the model. As can be seen, the variables Place, People, Customer Informedness*Place, and Customer Informedness*People display a VIF of higher than five.
This indicates that there is a possibility that multicollinearity poses a problem in the model. When looking at the results in Table 3, the significance of the variables is altered in the pooled model. Here, the variable Place loses its significance, while the interaction effect Customer Informedness*Place gains significance. An explanation could be that either Place itself does not have any direct influence, but only captured the influence that actually belongs to the interaction effect, or that multicollinearity influences these variables.
Table 5 indicates that the variable Customer Informedness on its own does not have a high VIF,
meaning that there is not much variance within this variable. Indeed, the variance appears to be
nearly constant. However, the VIFs of the variables Place and People are high, along with the
moderation of these variables with Customer Informedness. This suggests that the variance within
the interaction variables are carried over from the variables Place and People, causing them to
represent an imitation of Place and People. This imitation forms the source of the multicollinearity in
the model. However, in general, it is more likely for multicollinearity to report the interaction effect
as insignificant rather than the main effect. In addition, when segmenting for males, the variable
People and its interaction variable both appear to be significant. This again is odd if multicollinearity
is a large problem. Finally, the R-Square improves vastly for male-segmented model by adding the
moderators. For all the mentioned reasons, it could therefore be argued that the multicollinearity
problem might be a minor issue when looking at the pooled and male-segmented model. On the
other hand, it is peculiar that in the model specifically for females the previously significant variables
are no longer significant. In addition, BIC indicates that the model performance deteriorates instead
19 of improving by increasing from 113,168 to 121,942. This could be a sign that multicollinearity does indeed pose a problem in this area.
Table 5 : Variance Inflation Factor
VIF 1/VIF
Price 1.82 0.549735
Product 1.25 0.800743
Place 5.31 0.188314
Promotion 1.81 0.551963
People 5.96 0.167921
Customer Informedness 1.02 0.984321
Customer Informedness*Place 5.53 0.180820
Customer Informedness*People 6.00 0.166708
Mean VIF 3.59
Altogether, the results of the full model might very possibly suffer from multicollinearity issues, especially when segmenting for females. However, the model improves when segmenting for males and the associated significant variables does indicate that Hypotheses 6b and 6c should not simply be dismissed. Further research could provide an answer to this ambiguity.
Comparing the analyses
The two analyses provide different results. In fact, the only overlap between the two is the results regarding the importance of store proximity. Albeit the effect of location appears to be stronger in the logistic regression analysis, both confirm that proximity is of the essence in determining preference towards a pharmacy or drugstore.
The most striking differences are those regarding store type preference, product assortment size,
and customer informedness. The relative indifference in store type in the conjoint analysis is first
difference. The results of the logistic regression clearly indicate that there is a strong preference
towards drugstores as a base setting. Thus, the latter result suggests that the relative importance in
the decision-making process should be larger than a meager two percent. The second difference
refers to the importance given to the variable Product in the conjoint analysis. Whereas the logistic
regression results suggest that the variable is insignificant in determining preference towards a
pharmacy or drugstore, the conjoint analysis illustrates an increase in preference as the assortment
size increases. So far, this difference remains unexplained and forms a basis for further research. The
third difference addresses the relative importance assigned to customer informedness in the conjoint
analysis. As indicated during the elaboration regarding the methodology, not all variables of the
conceptual model have been included into the conjoint analysis as an attribute for analysis.
20 Consequently, the relative importance of some attributes might be overestimated. Therefore, it could be the case that Customer Informedness covers both the importance of itself and that of the variable People. The reason for this is that the variable People is another item that is influenced by the degree of knowledge of the consumers, especially the moderating effect between Customer Informedness and People. Additional research is needed to determine the validity of this statement.
It is difficult to compare a person’s subconscious preferences in a decision-making process with his/her conscious tendencies. When discussing the methodologies, it was mentioned that there might indeed be differences in the results due to subconscious predispositions that might not be captured in the logistic regression analysis. On the other hand, the choice-based conjoint analysis might suffer from a hypothetical bias. Further research is needed to determine the extent of these issues and the possible effects on the results.
Discussion
Out of the eight formulated hypotheses, only three are supported by the data. These three are to determine whether drugstores are to be considered a threat to pharmacies on the OTC market.
In accordance with Hypothesis 3, the extended product range of drugstores beyond the OTC category indeed seems to draw customers’ preference away from pharmacies and towards drugstores.
Therefore, the convenience associated with one-stop-shopping can be regarded as a competitive advantage in favor of drugstores. Note that this finding is not confirmed by the logistic regression analysis, and thus is only based on logical and intuitive deductions from the conjoint analysis.
When looking at the decision-making process of consumers, it appears that consumers consider the relative distance when choosing to purchase OTC products at either a drugstore or a pharmacy. The closer the store, regardless of the store type, the higher the preference of the consumer. This result is consistent with the findings in the logistic regression analysis. Hypothesis 4 is the only hypothesis confirmed by both methods. This signals the prominence of proximity regarding the preference towards either a drugstore of a pharmacy. Since the distribution coverage of drugstores is much larger than of pharmacies, the convenience created through proximity can be a vital competitive advantage for drugstores against pharmacies.
With Hypothesis 5 being the only hypothesis in favor of pharmacies, it automatically makes it the
only competitive advantage found so far. The hypothesis suggests that the knowledge of pharmacy
personnel is still the main point of attraction of pharmacies in terms of OTC products. However,
customer informedness might still pose a threat to pharmacies. Even though the hypothesis
21 regarding customer informedness is not confirmed by the analyses, both analyses demonstrate potential relevance of the variable. The results of the choice-based conjoint analysis do suggest that information is an important aspect in the decision-making process. Similarly, the logistic regression results remained rather inconclusive regarding the moderating effects involving the variable due to multicollinearity. Whether customer informedness is indeed an influential factor and whether this might undermine the competitive advantage of pharmacy – which is based on knowledge – is to be determined by further research.
Altogether, drugstores appear to have more advantages than pharmacies do. The most prominent advantage is that the level of convenience is higher for drugstore. Due to the growing need for convenience, this result could undermine the market position of pharmacies. Moreover, the regression and conjoint analysis suggests that the utility of purchasing at a drugstore is larger than that of a pharmacy. Especially when adding up all the attributes in the conjoint analysis associated with either a drugstore or a pharmacy. Even though the knowledge and expertise of pharmacy personnel contributes to higher preference, it could still be undermined by the other elements in favor of drugstores. Based on these findings, it can be said that when it comes to purchasing OTC products, drugstore are largely preferred over pharmacies. Thus, drugstores are indeed a viable, and even preferred, alternative to pharmacies and can therefore be considered as a competitive threat.
Managerial Implications
The need for convenience motivates consumers to purchase at the closest nearby location. With more drugstores available relative to pharmacies, drugstores will be the most convenient choice for consumers. Add to that the fact that drugstores provide additional products than OTC, consumers are able to raise the convenience level even further though one-stop-shopping. Thus, it can be argued that the need for convenience of consumers and the drugstores’ capabilities to service this need better than pharmacies currently do, causes drugstores to be a competitive threat for pharmacies in the market for OTC products. In order to stimulate customers to purchase OTC drugs at the pharmacy, pharmacies should increase their level of convenience. Since it will be difficult to justify expansion in the total amount of pharmacies based on solely the OTC sales, other methods need to be considered in order to improve pharmacies’ position in the market.
The current competitive advantage of pharmacy is based on the knowledge and expertise of its
personnel. However, the question remains if this is a sustainable advantage with a growing need of
customers to be informed before purchase. Leaning on a single advantage is therefore not a long-
term viable solution. By finding alternative methods to exploit their core competences and
22 capabilities, pharmacies can restructure their strategy in such a way that they might be able to compete more “aggressively” against drugstores. An example would be to engage in the rising trend of providing multichannel services to customers (this option will be further elaborated in the next section). Pharmacies should engage in competitive reasoning in order to identify competitive threats and alter their own strategies in order to adapt to changes in the market. Otherwise, competitors that are more attentive might outperform any changes made now in the future.
In conclusion, it is suggested that pharmacies restructure their strategies regarding OTC products in order to stay competitive on the market. By engaging in competitive reasoning and taking advantage of current market trends, pharmacies could maintain or even regain their market share.
E-pharmacy
Since it is inferred that drugstores are indeed a competitive threat to pharmacies, it is important that pharmacies alter their current manner of conducting business. Increasingly more brick-and-mortar retailers expending their current business model by selling products and/or providing services online.
These retailers are called click-and-mortars (Gruwal, et al., 2004). Based on the ever so increasing trend of E-tailing, would consumers value the ability of purchasing OTC products online? Which target group would value an e-pharmacy the most?
Business Concept
In order to capture the concept of online pharmacies, some aspects of the business model are discussed. Naturally, more elements could be added to make a more sustainable business model;
however, this illustration will only include the main elements that are also discussed in this paper.
Since information is a critical success factor for pharmacies, it is important that they fully utilize this capability. Information regarding the OTC products can be provided online via the same web shop. As an extra, pharmacies could even allow forums regarding products so that people could share their experiences regarding the products. This allows for a higher degree of customer informedness at the point of purchase.
By selling OTC products online, pharmacies would eliminate the issue of proximity, since the products
will be delivered at home. This in turn increases the convenience level of purchasing from a
pharmacy. Moreover, people can see online whether a product is on sale or not, which lowers the
uncertainty regarding price promotions. Furthermore, should consumers embrace the notion of
purchasing OTC products online, the reputation of pharmacies could assert trust and security for
consumers. This could ensure that consumers would purchase from an E-pharmacy rather than
23 another regular E-tailor. Especially since the principle of an e-pharmacy is not that hard to imitate by competitors.
However, setting up an E-pharmacy is not without obstacles. Even though trust might be an important factor when ordering medication online, drugstores might also be able to assert the same confidence over time. Moreover, since drugstores already have web shops and other logistic infrastructure in place, they could be able to outperform pharmacies in terms of cost and efficiency.
In addition, creating a web shop and the associated logistics would be a very costly investment.
Especially if there is uncertainty regarding the success of the endeavor.
Data Analysis & Discussion
To investigate whether people are actually interested and willing to purchase products online, the survey also included a question regarding this matter (see Appendix II). Out of the 155 people respondents to this question, roughly 41 percent answered that they would purchase OTC products online.
Two groups have been made to segment the “yea-and-nay-sayers”. An ANOVA test is performed in order to determine whether the observed differences are also significant differences. The results in Table 6 state that only Age Category, and Income Category are significantly different. Thus, only these two can be used to describe the characteristics of two groups. The details regarding the remaining variables are presented in Appendix IV.
Table 6: Analysis of Variance
F-Statistic
Frequency 3.094
Gender 2.257
Age Category 31.857***
Working Hours 2.494
Income Category 27.193***
Note: significance indicators: *** p< 0.01; ** p<0.05; * p< 0.1.
The first distinction is with respect to the Age Category. Figure 4 shows a detailed description of the
cluster and categories. Notice that the majority of the youngest age category votes “No” to the
concept of an online pharmacy. This outcome is very surprising, since the younger age groups are
considered to be more inclined to favor e-tailing activities (Hernández et al., 2011; Trocchia & Janda,
2000).
24
Figure 4: Age Category (Yes/No percentages per category group)
Another reason could be that as people age, they climb higher in their professional career and become more pressed for time. Purchasing online increases their convenience level relatively more than for younger people (Bhatnagar et al., 200). Yet the amount of no-voters remains stable at roughly 60% among all categories of working hours per week (see Figure 5)
Figure 5: Working Hours per Week (Yes/No percentages per category group)
The second significant difference is associated with Income Category (See Figure 6). The majority of the three lower income categories would respond negatively to purchasing from an E-pharmacy.
Conversely, the upper two categories present a larger group of yes-voters.
18-25 25-35 35-45 45-55 55-65 65+
Yes 38,7% 46,8% 30,8% 55,6% 40,0% 0,0%
No 61,3% 53,2% 69,2% 44,4% 60,0% 100,0%
0,0%
20,0%
40,0%
60,0%
80,0%
100,0%
120,0%
percentage
Age Category
Less than 10 hours per week
10-20 uur hours week
20-30 hours per week
30-40 hours per week
more than 40 hours per week
Yes 42,9% 40,5% 43,8% 39,6% 42,1%
No 57,1% 59,5% 56,3% 60,4% 57,9%
0,0%
10,0%
20,0%
30,0%
40,0%
50,0%
60,0%
70,0%
Percentage
Working Hours per Week
25
Figure 6: Income Category (Yes/No percentages per category group)
Hernández et al. (2011) provide an explanation for this occurrence. They suggest that the income of a consumer is only of influence during first contact with e-commerce activities. High income category consumers are better able to withstand a financial loss associated with the risk of using an unfamiliar purchase channel (Hernández et al., 2011; Lu et al., 2003). However, as experiences with the online environment increases, consumer behavior is nog longer influenced by income (Hernández et al., 2011; Al-Somali et al., 2009). Since the notion of purchasing pharmaceutical products online is not really been established yet, the negative reaction amoung low income groups could be based on risk- averse behavior. Another, and more simpler, explenation could be that there often is a correlation between age and income, where younger consumers also have a lower income level. Since , most of the respondents fall into the first two age categories, it might give a reason for the same pattern occuring in the variable Income Category. In this particular case, there is a significant correlation between age and income is 0.338, meaning that there is indeed some corelation between the two variable.
Conclusion
Even though significant differences are found between the respondents in terms of favoring an E- pharmacy or not, no strong and clear-cut target groups are established. However, this does not entail that there is no future for a click-and-mortar pharmacy. The sample size of this analysis is limited rather skewed; therefore, it is very likely that it does not represent the population accurately.
Moreover, this (preliminary) analysis is based on a market-pull principle, which does not always indicate that there is no room in the market for a certain product or service. It has been said that
“…people don't know what they want until you show it to them” – Steve Jobs (1998). Therefore, the result of the analysis remains inconclusive.
Less than
€1000,- €1000 - €2000,- €2000 - €3000,- €3000 - €4000,- Higher than
€4000,-
Yes 39,4% 42,3% 36,0% 57,1% 60,0%
No 60,6% 57,7% 64,0% 42,9% 40,0%
0,0%
10,0%
20,0%
30,0%
40,0%
50,0%
60,0%
70,0%
Percentage
