If the Car Fits, Drive It: Development of a Prototype Hypermedia Interface for Facilitating Online Vehicle Sales Through Narrowcasting

The Internet provides unlimited opportunities for “narrowcasting,” or individualized advertising, in the automobile industry. While car manufacturers attempt to drive sales with targeted advertising, efforts collapse at the dealership as customers with diverse backgrounds, lifestyles, and preferences are presented with an identical sales pitch. Likewise, existing automobile commerce sites fail to respond to a customer’s personal needs, instead relying on a filtering system for selecting models. After using schema theory to analyze car shopping and individualized advertising on the Internet, the author presents a prototype customer-driven hypermedia interface that generates unique, narrowcast sales presentations for individual users. Targeted Advertising, Narrowcasting, and Schema Theory Marketers spend billions of dollars each year trying to find the most receptive audiences for their products through targeted advertising. For example, products targeted towards men are often promoted in Sports Illustrated, GQ, and Motor Trend magazines, while advertisers in search of female customers may use Women’s Day, O, and Mademoiselle magazines. On television, niche networks such as ESPN and Speedvision, along with Lifetime and Home and Garden Television provide gender segmentation. However, targeted advertising is less effective in television and other broadcast media than in magazines. Since no methods are in place to precisely measure the composition of a television audience, advertisers inevitably end up promoting their products to segments of the population that have no interest in purchasing them. Even Nielsen Media, developers of the most popular form of audience definition the Nielsen ratings makes no exact claims about audience composition and instead relies on projected estimates from a random sample. Since 30-second television advertisements can cost over a million dollars, using a broadcast media for targeted advertising is not only uncertain, but also costly. The money spent on ineffective advertising, which fails to reach the target audience, could be used to develop new technologies and improve other aspects of the production cycle, while the advertising space could be filled with products that appeal to the audience at hand. With the advent of the Internet, advertisers now have direct access to individual consumers. Using the Internet as medium of communication, marketers can maximize the potential of “narrowcasting,” customizing an advertisement to a particular consumer’s tastes. Research has shown advertisements and the message cues contained within are interpreted differently by different people [2]. In addition, studies conducted by Joan Meyers-Levy have attributed many of these differences in perception to gender, while other studies have linked differences to other demographic contrasts such as age, occupation and personal interests [2, 4]. With narrowcasting, a marketer obtains information about a customer’s preferences and uses that information to create personalized advertisements that appeal to individual perceptions and eliminate the waste and uncertainty of broadcast media. Favorite hobbies, sports, and vacation destinations are a few examples of data that can be incorporated into narrowcast advertisements. The concept of narrowcasting stems from schema theory, developed by Roger C. Schank in the late 1970’s. Schema theory attempts to explain how situations and their identifiers can shape our perceptions and actions. Under schema theory, a script is a sequence of actions used to deal with a particular scenario [5]. Individuals have scripts for routine actions such as visiting a restaurant and buying groceries, and for lessencountered scenarios such as buying a car. Due to the existence of scripts, people at a restaurant know what to do with the menu, how to place an order and when to pay the bill. Scripts are flexible however, and different individuals use enhanced variations of scripts to process a given scenario [1]. For example, individuals have different scripts for choosing a car. A 25-year-old car enthusiast may require a 200-horsepower, V-6 engine, while the car-selection script for a 40-year-old mother of two may be more focused on safety features such as side airbags and child safety locks. Often these script variations are linked to perception and can lead to different methods of interpreting data, as in the differences between individuals in understanding the intended message of an advertisement [2, 4, 5]. Schemas are the totality of objects, events, and scenarios that signal which scripts to use in a given situation. They allow individuals to recognize and understand situations, and are obtained through personal experience, education, and through communication with others. For example, balloons, streamers, a cake with candles on top, and wrapped presents are all part of the birthday party schema. After the schema has been identified, the related scripts are employed to interact with the situation [5]. The online automobile sales industry is poised to capitalize on the use of narrowcasting. While targeted advertising has been successful in reaching prospective customers, the retail segment of the car industry has failed to take advantage of its full potential by using only one standard sales script. This article describes a prototype customer-driven hypermedia interface that generates unique, narrowcast sales presentations based on customer inputs. It operates off the user’s variations in the car selection script, highlighting personal preferences and lifestyle requirements. The interface eliminates the waste found in broadcast advertising and is more effective at recognizing individual needs than existing systems used by the automobile industry. Car Dealerships and Online Car Seller Websites: The One Script Sales Pitch Car manufacturers regularly seek a target audience for each model, and advertisements for that model are often intended to appeal specifically to the target audience. Minivans provide a classic example: Known for their “soccer mom” appeal, minivans are targeted to a “family woman” audience. Ads highlight safety and reliability, with text and graphics that promote group excursions, cargo capacity, and an overall theme of togetherness [6]. However, for most models the target audience is based on age and income ranges rather than gender. Examples include the Toyota Camry, Honda Civic, and Mercury Grand Marquis. As a result, car manufacturers often must simultaneously appeal to men and women in the same advertisement [2, 6]. Research has shown that men and women process ads differently, so the industry regularly creates separate ads for each gender [2]. Generally, ads targeted for men have a tendency to contain small amounts of copy and focus on one particular feature of the vehicle. The schema for a “female” ad includes more factual or statistical information, highlighting safety and reliability issues. In addition, advertisements designed to run in women’s magazines often contain more pictures than men’s advertisements and simultaneously promote several of the vehicle’s features [2, 3]. These targeted advertisements are designed to pick up on variations of the car-selection script held by the target audience. If “soccer moms” are known to be primarily concerned about safety when choosing a vehicle, advertisements aimed at them will focus on safety features. After spending billions of advertising dollars each year to get people to the showroom, retail car dealers drop the ball when customers arrive. While targeted advertising is often successful in creating awareness of the vehicle models among their respective target audiences, efforts to segment consumers via their car-selection scripts are wasted when all customers are given an identical point-of-purchase experience. Dealerships are often grouped together and look alike from one to the next. While a Lexus vehicle may differ greatly from a Kia automobile, retail dealerships for both companies are remarkably similar. The schema for a car dealership includes showrooms filled with glossy brochures and posters, corporate logos on the walls, and uninformed sales personnel on the sales floor. Just as the dealerships are similar, car dealers rely on one sales script to sell cars. All customers receive basically the same "presentation," or sales pitch, from the car dealer. The salesperson knows to show each customer several models, discuss the features available on each, and begin sales negotiations. Aside from sizing up a customer as they enter the lot, a salesperson only draws from casual conversation to determine a customer’s preferences. However, the salesperson often relies on stereotyping to determine each customer’s needs and which vehicles to recommend. Based on one sales script, the presentation fails to account for a customer’s lifestyle and preferences. In other words, car dealers ignore personal variations in an individual’s car-selection script by offering only one method of sales presentation. While all customers receive the same sales pitch, female customers seem to experience more frustration when buying a car. Manufacturers regularly receive complaints from female customers stating that salespeople automatically address the sales pitch to their husbands. J.D. Power & Associates also reports that in a recent study, the majority of women still find the car-buying experience intimidating [3]. Finally, 82% of women believe dealers treat them as not knowledgeable enough about cars, compared with 70% of men [7]. A hypermedia interface for online car shopping could eliminate chauvinism by acting as an unbiased salesperson whose goal is to elicit information about a customer’s personal interests and car preferences. Another problem with the existing car sales structure is that the salesperson’s interests often conflict with the consumer’s. The salesperson is paid via a commission structure through high-pressure sales. He or she will rarely suggest that a customer select a competitor’s brand, even if the competing vehicle is better suited for the consumer than their own model. Analyzing the conflict of interest one step further, the salesperson’s role is not to provide the car most suited for a particular customer, but to sell the most expensive model that a customer’s income and credit rating will allow. While the standard car-buying process has proved functional for many years, the Internet provides an opportunity for a more effective system in which buyers are responsible for declaring their own needs and preferences. Web sites for online car sellers can be designed to provide a personalized buying experience to customers based on combinations of user-defined preferences and demographic information. Whereas a sales associate works through the standard sales script with a similar delivery for each customer, online sales presentations allow for dynamic shifting of both content and overall appearance. Today, even the most "interactive’ car-buying web sites simply filter information in a process that narrows down choices until users find what they are looking for. While this selectivity process works well for customers who know exactly what vehicle they want, it fails to account for the personal preferences contained in an individual’s car-selection script, much like the dealership. Some customers do not know what category of vehicle they want, such as a coupe or a truck, and even less know the exact model they want to purchase. Unfortunately, online sites for dealers and manufacturers all begin filtering at the most specific choices make, model, and year. Another problem with automobile commerce sites is that the flow of information and the idea of need fulfillment match the dealership experience seller to buyer. The filtering process, like the salesperson, attempts to pair the customer to specific models without input about a customer’s individual preferences. Online Sales: Methods Automobiles were ideal test products for the prototype interface for several reasons. First, they represent an industry that accounts for billions of dollars in annual worldwide sales. Automobiles are one of the most expensive repeated purchases for many individuals, and the integration of vehicle sales into an online medium holds potential time and cost savings for both consumers and car sellers. Automobile branding relies on product differentiation and targeted advertising, so online car sales provide a viable platform on which to test the new narrowcasting system [3]. Finally, choosing a car is a highly individualized process, and there are numerous script variations available some customers require more cargo room for their particular lifestyle, while others specifically look for a comfortable, smooth ride. Unfortunately, car sellers have thus far failed to develop the full potential of the Internet to enhance automobile sales, as less than one percent of cars are purchased online [7]. The lack of adaptation by consumers is not irreversible and is partially due to ineffective e-commerce models for automobile sales. Many consumers spend a considerable amount of time shopping for their automobiles, but the majority of existing business-to-consumer (B2C) e-commerce models are developed for products that require less research, or “shopping time,” from the consumer. Examples include books and movies, as well as music, clothing and travel accommodations. As a result, it remains a possibility that consumers have begun to relate online shopping with small, inexpensive products. Also, current e-commerce models for automobile sales are inefficient; they fail to utilize the narrowcasting possibilities of the Internet by using one sales script for all customers. Through the filtering system online car sellers mirror the retail dealership experience of seller-to-buyer information flow and again fail to make use of the personalized variations in an individual’s car-selection script. Before designing the interface, we analyzed existing online car seller websites for content, style, level of user orientation, and overall functionality. We examined websites for retail car sellers as well as manufacturers. The sites were chosen based on information from Gomez.com, an online buying guide analogous to Consumer Reports magazine. Gomez.com provides extensive data on site functionality and quantifies its findings with scores and rankings. At each website, we observed the functionality, depth of information, and attempts at user-orientation regarding the purchasing process. By integrating the qualitative observations with Gomez.com’s numerical data, we were able to focus on the consumer experience of online car shopping without using overlapping data. After compiling all the information from the online car seller sites, we found several common problems with the websites. First, they were poorly designed with confusing layouts, difficult navigation and circular links. Second, while the sites included extensive data for particular car models, used car listings contained very little information about the particular car available for sale. Most sites provided only one photo of each used vehicle that was for sale, and some even used stock photos of the model without informing the consumer. Several sites provided links to vehicle history reports such as Carfax.com, but the vehicle descriptions contained no information on mileage and condition estimates or even links to Consumer Reports data about the particular model. Third, many of the car seller websites were owned by a consortium of car dealers, AutoNation.com for example. As a result, it appeared on several occasions that the role of the website was not to sell cars online, but to send customers to a dealer for the purchase. This scenario defeats the purpose of the online ordering option and places sales goals in question. All the online car sellers operated as if the sites were designed for consumers who knew exactly what kind of car they wanted to buy. For most sites, search capabilities were limited to the input of a make, model, and year. Since most car shoppers don’t know what make and model they want to buy before they start researching, these websites have a critical flaw in their design. For the rest of the sites, specifications were used to narrow down the choices. For example, at Autotrader.com, a user can run a search for a truck with: a manufacturer’s suggested retail price (MSRP) of less than $20,000; an extended cab; and standard power windows; and receive a list of models that fit the description. The “specification search” for features is ineffective because users are often more concerned with general feature categories, such as safety and performance. Very few users select a particular model solely because of the number of cup holders it has or because it features heated side mirrors. Likewise, for those users who don’t have a preference as to some of the specifications offered, the search can result in hundreds of models. After we examined car-related sites, the study focused on user-centric innovations used in other sectors of e-commerce. Consequently, we analyzed sector leaders such as Amazon.com (books and other media), Ebay (online auctions), and Landsend (clothing). These companies have all revolutionized some aspect of the online shopping experience, and thus were primary candidates for our observations. Most important was their focus on increased customer value through personalization, a key factor for our research. At Amazon.com, customers are presented with a customized list of recommended products based on their past product searches. Ebay uses the excitement of a live auction to bring customers back time and time again. Finally, Landsend’s website promotes a feature called “My Model,” which gives advice on certain styles and colors a customer should wear, based on the information a customer provides about their own physical characteristics. The fundamental defects with the current capabilities of online car shopping research are one of the driving forces behind our decision to create a more user-friendly hypermedia interface as used in other sectors of e-commerce. By placing the user at the forefront of the car shopping process, the Internet can be used as a viable purchase alternative to the traditional one-script dealership experience. The Interface: Results The foundation of the hypermedia interface is a database that stores all of the vehicle data, including technical specifications, photographs, and purchase information. Also included in the database is the entire catalog of options and input fields for the enduser. The interface uses six presentation modes male and female presentations for each of three age groups (under 25, 25-49, 50 and over). The modes are a form of targeted advertising, as each mode has a unique page design, prose style, and delivery approach intended to appeal to the selected group. The modes are based on a separate marketinginfluenced database that contains extensive demographic and gender purchasing data. Information from existing consumer research sources such as CACI Marketing and Claritas Express databases is synthesized to create these genderand age-differentiated modes. After entering their age and gender, users are taken to a web page that asks for their preferences concerning vehicle design and function. The interface asks customers to enter and rate the importance of their personal selection criteria, such as price, safety, and performance. Also included is a question about the user’s knowledge of vehicles and the technical terms associated with assembly and performance. In addition, the web page contains questions about the user’s lifestyle and car usage. Figure 1 is a shortened version of the "Preferences Page" that highlights several example questions. The user’s answers to these questions provide the content of the presentation. As a result, the sales pitch is modified to account for both the variations in each user’s car-selection script and the user’s personal needs. After a user completes the input process, the interface matches vehicle models to the preferences chosen based on information in the database. The information is then combined with one of the six modal frameworks, effectively generating a website presentation that is designed to appeal directly to the user. The presentation is the modified sales pitch for the models that are the most appealing to the customer, based on the customer’s selections. It emphasizes information the customer has deemed important, creating a narrowcast advertisement for the models. The presentations are combinations and permutations of the six age/gender modes and the choices selected on the “Preferences Page.” Depending on the permutation, page designs are enhanced with varying prose styles, graphics and sound use, and video clips. To clarify, a presentation for younger customers would include slang, pop references, and trendy graphics design, whereas one for older customers might rely upon historical references and brand loyalty. The difference between the male and female modes relies on the previously mentioned differences in perception between the genders. The schemas for male and female ads are employed in their respective modes. As an example, let’s say Howard Hotshot wants to buy a car. He is in his latethirties and is an avid car enthusiast. Howard isn't sure whether he wants a convertible or a sports-coupe, but he wants to spend less than $40,000 for his new vehicle. Performance and durability features are important to Howard. He prioritizes these categories and selects his appropriate level of technical knowledge from three choices (from most knowledgeable to least knowledgeable: mechanic, car enthusiast, and driver). The interface then asks Howard several questions about his lifestyle and car usage. Using the appropriate mode (male customers, 25-49 years old), the interface creates a presentation for Howard. The presentation begins with a video clip showcasing the performance of the selected models. After the video introduction, the "Presentation Page" appears, highlighting the information and features that Howard has deemed important (performance and durability). Several automobile models are presented to Howard, with an emphasis on characteristics that appeal to his individual preferences and lifestyle needs as obtained from the “Preferences Page.” All information is styled to match his level of technical knowledge, especially the text descriptions of the vehicles. As Howard browses through the different models, he finds links to outside resources such as consumer reviews, magazine articles, industry ratings, and a direct link to purchasing options. Permanent links to manufacturers’ websites are available so Howard can explore vehicles in greater detail and develop customized versions of his desired models. Consumer Reports, manufacturer specifications and press releases, government safety test data, and related sources are used to provide text information for the presentation. However, trade magazine articles are also supplied for users who label their own level of technical knowledge as enthusiast or mechanic. For all consumers, testimonials from current users are available, noting the positives as well as the negatives of owning each particular model. As another example, assume Meredith Moneybags decides to buy a car. She’s an affluent woman in her mid-forties who isn’t sure whether she wants a sports car or an SUV. Price is not a concern, but vehicle design, comfort, and safety features are important to Meredith. She prioritizes these preferences and selects her level of technical knowledge as “driver,” the option for the least knowledgeable users. Meredith then answers questions about her lifestyle and car usage, similar to those shown in Figure 1. Using the appropriate mode (female customers, 25-49 years old), the interface creates a presentation for Meredith, one that is more graphic-intensive than Howard’s and filled with descriptions and statistical information. Meredith is first introduced to basic differences between vehicle types, particularly sports cars and SUVs, highlighting differences among them in the feature categories she has selected (design, comfort, and safety). Several models are presented to Meredith, again emphasizing characteristics that might appeal to her individual preferences and lifestyle. In Meredith’s case, let’s assume she likes to take long road trips for vacations. The interface would then display the models in a vacation setting, with sections of the text focused on road-test data from Consumer Reports and Motor Trend. Since her level of technical knowledge is lower than Howard’s, the text contains more common, everyday words to describe the vehicles and relies less on the manufacturer’s specifications. As she browses through the different models, Meredith also finds links to consumer reviews, more magazine articles, and industry ratings. Finally, as in Howard’s presentation, links to purchasing options and manufacturers’ websites are prominently displayed.