Ft. Lauderdale, Florida, USA. April 5-10, 2003 Paper: Designing Applications for Handheld Devices Mobile Computing in the Retail Arena Erica Newcomb, Toni Pashley, John Stasko College of Computing/GVU Center Georgia Institute of Technol Atlanta, GA 30332 USA enewcombabellsouth.net,tpashley(@aculent.com,stasko@ccgatech.edu ABSTRACT Within the last year or two, numerous shopping aids have Although PDAs typically run applications in a"stand- emerged, from on-cart devices to applications for handheld alone mode, they are increasingly equipped with wireless devices. Easi-OrderTM is a PDa application for use in the communications, which makes them useful in new home that focuses on the creation of a personal shopping domains. This capability for more powerful information list that can be sent to the grocery store [1].This exchange with larger information systems presents a new application has been launched in Safeway stores across the situated context for PDA applications, and provides new UK. Klever-Kart was designed by Klever Marketing, Inc design and usability evaluation challenges to bring "interactive advertising to shoppers directly at the In this work we examine how grocery shopping could be point of decision"[5].Klever-Kart is an on-cart device that aided by a mobile shopping application that consumers access PDa while in a store. The interactive nutrition information, news and weather. Finally, Andersen relationship between the physical space of the store and the Consulting developed an application called Shoppers Eye this application. To better understand this interaction, we of buying opportunities[]. The user carries a handheld in studied people's grocery shopping habits, designed and a shopping mall and the system makes their shopping list evaluated prototypes, and performed usability tests within available to local stores, which in turn make bids for the the shopping environment. This paper reveals our design users busine process for this problem and a framework for designing and evaluating situated applications for mobile handhelds PDA, mobile interfaces, situated computing, wireless communication, user-centered approach, interactions INTRODUCTION personal digital assistants PDAs) have largely been used for a few narrow tasks scheduling, calendar management, and list keeping are ome of the most common ones However e pdas acquire wireless connectivity capabilities, a new set of potential uses is exposed. We became interested in exploring the use of PDAs in retail shopping situations, more specifically, in grocery stores. The biggest advantage of the pda is its portability and how well it works for Figure 1: A grocery shopper using a PDA organizing and presenting information. These characteristics are appropriate for the task of shopping How do mobile devices such as these affect the shopping experience? Is a PDa potentially beneficial for a shopping Until recently, commercial grocery shopping aids were environment? Because there are obvious limitations to hard to find. One of the first shopping applications for the PDAs such as the display size, resolution and difficulty of PDA that we encountered was simply a list management input, we speculated that most handheld devices used in tool allowing users to input and select items for a grocer mobile situations would only be appropriate for on-the-go list. The application provided little more than a physical list lookup activities and provided no assistance in the act of shopping We wondered how a shopper would shop while holding a Permission to make digital or hard copies of all or part of this work for device, as shown in Figure 1. Because shopping in any per nade or distributed fe s gfat e c wmmerc capacity almost always requires the use of hands, form bear this notice and the full citation on the first page. To copy otherwise, factor is extremely important in mobile computing. We felt or republish, to post on servers or to redistribute to lists, requires prior that focusing on the task of shopping: learning how people specific permission and/or a fee determine what they need and how they obtain it, would be CHI 2003, April 5-10, 2003, Ft Lauderdale, Florida, USA. an appropriate first step into exploring how a PDA could Copyright2003ACMl-s8113-630-7/0304..s500. interact with a larger in-store information system Volume no 5 issue no. 1
Mobile Computing in the Retail Arena Erica Newcomb, Toni Pashley, John Stasko College of Computing/GVU Center Georgia Institute of Technology Atlanta, GA 30332 USA enewcomb@bellsouth.net, tpashley@aculent.com, stasko@cc.gatech.edu ABSTRACT Although PDAs typically run applications in a “standalone” mode, they are increasingly equipped with wireless communications, which makes them useful in new domains. This capability for more powerful information exchange with larger information systems presents a new situated context for PDA applications, and provides new design and usability evaluation challenges. In this work we examine how grocery shopping could be aided by a mobile shopping application that consumers access via a PDA while in a store. The interactive relationship between the physical space of the store and the human activity of shopping are crucial when designing for this application. To better understand this interaction, we studied people's grocery shopping habits, designed and evaluated prototypes, and performed usability tests within the shopping environment. This paper reveals our design process for this problem and a framework for designing and evaluating situated applications for mobile handhelds. Keywords PDA, mobile interfaces, situated computing, wireless communication, user-centered approach, interactions INTRODUCTION Hand-held computers and personal digital assistants (PDAs) have largely been used for a few narrow tasks: scheduling, calendar management, and list keeping are some of the most common ones. However, as more PDAs acquire wireless connectivity capabilities, a new set of potential uses is exposed. We became interested in exploring the use of PDAs in retail shopping situations, more specifically, in grocery stores. The biggest advantage of the PDA is its portability and how well it works for organizing and presenting information. These characteristics are appropriate for the task of shopping. Until recently, commercial grocery shopping aids were hard to find. One of the first shopping applications for the PDA that we encountered was simply a list management tool allowing users to input and select items for a grocery list. The application provided little more than a physical list and provided no assistance in the act of shopping. Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee. CHI 2003, April 5–10, 2003, Ft. Lauderdale, Florida, USA. Copyright 2003 ACM 1-58113-630-7/03/0004…$5.00. Within the last year or two, numerous shopping aids have emerged, from on-cart devices to applications for handheld devices. Easi-Order™ is a PDA application for use in the home that focuses on the creation of a personal shopping list that can be sent to the grocery store [1]. This application has been launched in Safeway stores across the UK. Klever-Kart was designed by Klever Marketing, Inc. to bring “interactive advertising to shoppers directly at the point of decision” [5]. Klever-Kart is an on-cart device that offers users in-store information such as sale items, nutrition information, news and weather. Finally, Andersen Consulting developed an application called Shoppers Eye that “addresses the problem of shoppers’ lack of awareness of buying opportunities” [3]. The user carries a handheld in a shopping mall and the system makes their shopping list available to local stores, which in turn make bids for the user’s business. How do mobile devices such as these affect the shopping experience? Is a PDA potentially beneficial for a shopping environment? Because there are obvious limitations to PDAs such as the display size, resolution and difficulty of input, we speculated that most handheld devices used in mobile situations would only be appropriate for on-the-go lookup activities. We wondered how a shopper would shop while holding a device, as shown in Figure 1. Because shopping in any capacity almost always requires the use of hands, form factor is extremely important in mobile computing. We felt that focusing on the task of shopping; learning how people determine what they need and how they obtain it, would be an appropriate first step into exploring how a PDA could interact with a larger in-store information system. Figure 1: A grocery shopper using a PDA Ft. Lauderdale, Florida, USA • April 5-10, 2003 Paper: Designing Applications for Handheld Devices Volume No. 5, Issue No. 1 337
Paper: Designing Applications for Handheld Devices CHI 2003: NEW HORIZONS Our design focuses on a real world application of mobile checkout, 7)Within store price comparisons, 8)Map of computing carried out in a retail environment. Once in the grocery store, 9) Recipes and ingredients, 10) Health retail environment, the wireless network creates a"digita nformation, I 1)Product information space"where information is sent to the user's PDa This survey also allowed us to glean device preferer ontext for the user with task specific information, and thus among participants. We found that although 83%of population used a mobile phone (only 26% used a PDA) a computational environment to support grocery shopping they preferred a PDA as the form factor for a shopping This paper discusses our process for designing a wireless application, People felt that they could trust their personal application for a PDa with regard to the human PDA more than a device provided by the store. They nteractions and situations in the realm of grocery thought that it would give them flexibility between home shopping. First we describe our initial observations and and store usage. These responses were worth further study of numerous shoppers and how they act in the retail investigation because, in their explanations, many space. Next, we discuss how we undertake designing and participants made assumptions about the functionality and evaluating the interactions of a mobile interface. And form factor of the product even though it did not yet exist finally, we describe the initial prototype developed and the The survey, interviews, and observations provided us with results from an in-store evaluation a wealth of information from which we garnered some We utilize ideas from situated computing [4] and discuss obvious but latent facts about how people shop. Basically, evaluating and designing an interface for a mobile two kinds of shoppers emerged, those who make physical application. We assume a technological infrastructure lists and those who make mental lists. We did not find a exists to support the overall system. Tradeoffs and single person who did not at least make some type of decisions made for designing and evaluating an application mental or physical notation for items they needed before prototype for a handheld device are included in the hopes entering a grocery store hat this study may provide others with assistance for This led us to wonder what happens to the physical grocery developing these types of applications list while shopping. How is it used? Most often the list is HOW DO PEOPLE SHOP? carried on the person(in-hand or in a pocket) throughout When a consumer enters a grocery store their intention is to the store. Others leave the list in the cart as they retrieve buy products. Unlike other forms of shopping, people do items. About half of the people who carry a physical I not generally"window shop"for groceries. Our early mark off items as they shop. This is noteworthy because it investigations to uncover how people grocery shop shows that people sustain interactions with their grocery list involved three main directions in the midst of shopping 1)Observe people while they shop. How is it that people On the other hand, those people who make mental lists tend maintain their needs? Do they use a list or do they wait for to rely more on the grocery store environment to either the context of the store to drive their needs? Or does some inspire or remind them of items they might need combination hold? To answer these questions, we Therefore, browsing was the main catalyst for obtaining anonymously observed shoppers in 3 different grocery their grocery needs In our observations we discovered one of the most 2)Shop with people and ask questions about their shopping important factors in the mechanics of shopping: shoppers as it unfolds. This ethnographic approach offered very must use their hands. People require their hands and eyes to valuable insight. We followed 5 participants throughout the look and feel for fresh and ripe produce. They must pick up task of shopping. Shopping is a repetitive task and it was packages and read labels. Grocery shopping is an especially easy to see people's shopping habits emerge, especially in a hands-busy, eyes-busy task. As Underhill states, "it's hard grocery store with which they were familiar to overemphasize the importance of the hand issue to the 3)Ask people how they shop. We designed a survey with a world of shopping. if the shopper can't pick [items] up set of wide-ranging questions on topics including shopping It's all for naught[81 frequency, list making, habits, etc The retail Environment 46 participants responded to the survey. We asked The motivation of a retail store opposes the users ons a bout shopping, technology usage, and potenti intention;a retail store wants shoppers to spend as much features for a grocery application. Participants ranged in time as possible in the store because this can directly age from early 20s to mid-60s with a mix of single translate to more sales. The longer a shopper is in a store married, working, retired, and with or without children the more likely they are to buy something [ 8]. According to market research, almost 60 of all products bought at the When presented with features of a possible grocery grocery store are not premeditated purchases [2 shopping application, participants ranked them in order Underhill's fieldwork techniques present the environmental from highest to lowest priority: I)Create a grocery list, 2) psychology behind public spaces and reveal how important Arrange a grocery list, 3)Provide product location by aisle, the store and the aisle are to the total shopping experience 4)Itemized pricing, 5)Specials and coupons, 6) Electronic Volume no. 5. issue no. 1
Our design focuses on a real world application of mobile computing carried out in a retail environment. Once in the retail environment, the wireless network creates a “digital space” where information is sent to the user’s PDA contingent on their shopping list. The digital space provides context for the user with task specific information, and thus a computational environment to support grocery shopping. This paper discusses our process for designing a wireless application for a PDA with regard to the human interactions and situations in the realm of grocery shopping. First we describe our initial observations and study of numerous shoppers and how they act in the retail space. Next, we discuss how we undertake designing and evaluating the interactions of a mobile interface. And finally, we describe the initial prototype developed and the results from an in-store evaluation. We utilize ideas from situated computing [4] and discuss evaluating and designing an interface for a mobile application. We assume a technological infrastructure exists to support the overall system. Tradeoffs and decisions made for designing and evaluating an application prototype for a handheld device are included in the hopes that this study may provide others with assistance for developing these types of applications. HOW DO PEOPLE SHOP? When a consumer enters a grocery store their intention is to buy products. Unlike other forms of shopping, people do not generally “window shop” for groceries. Our early investigations to uncover how people grocery shop involved three main directions: 1) Observe people while they shop. How is it that people maintain their needs? Do they use a list or do they wait for the context of the store to drive their needs? Or does some combination hold? To answer these questions, we anonymously observed shoppers in 3 different grocery stores. 2) Shop with people and ask questions about their shopping as it unfolds. This ethnographic approach offered very valuable insight. We followed 5 participants throughout the task of shopping. Shopping is a repetitive task and it was easy to see people’s shopping habits emerge, especially in a grocery store with which they were familiar. 3) Ask people how they shop. We designed a survey with a set of wide-ranging questions on topics including shopping frequency, list making, habits, etc. 46 participants responded to the survey. We asked questions about shopping, technology usage, and potential features for a grocery application. Participants ranged in age from early 20s to mid-60s with a mix of single, married, working, retired, and with or without children. When presented with features of a possible grocery shopping application, participants ranked them in order from highest to lowest priority: 1) Create a grocery list, 2) Arrange a grocery list, 3) Provide product location by aisle, 4) Itemized pricing, 5) Specials and coupons, 6) Electronic checkout, 7) Within store price comparisons, 8) Map of grocery store, 9) Recipes and ingredients, 10) Health information, 11) Product information. This survey also allowed us to glean device preferences among participants. We found that although 83% of our population used a mobile phone (only 26% used a PDA), they preferred a PDA as the form factor for a shopping application. People felt that they could trust their personal PDA more than a device provided by the store. They thought that it would give them flexibility between home and store usage. These responses were worth further investigation because, in their explanations, many participants made assumptions about the functionality and form factor of the product even though it did not yet exist. The survey, interviews, and observations provided us with a wealth of information from which we garnered some obvious but latent facts about how people shop. Basically, two kinds of shoppers emerged, those who make physical lists and those who make mental lists. We did not find a single person who did not at least make some type of mental or physical notation for items they needed before entering a grocery store. This led us to wonder what happens to the physical grocery list while shopping. How is it used? Most often the list is carried on the person (in-hand or in a pocket) throughout the store. Others leave the list in the cart as they retrieve items. About half of the people who carry a physical list mark off items as they shop. This is noteworthy because it shows that people sustain interactions with their grocery list in the midst of shopping. On the other hand, those people who make mental lists tend to rely more on the grocery store environment to either inspire or remind them of items they might need. Therefore, browsing was the main catalyst for obtaining their grocery needs. In our observations we discovered one of the most important factors in the mechanics of shopping: shoppers must use their hands. People require their hands and eyes to look and feel for fresh and ripe produce. They must pick up packages and read labels. Grocery shopping is an especially hands-busy, eyes-busy task. As Underhill states, “it’s hard to overemphasize the importance of the hand issue to the world of shopping…if the shopper can’t pick [items] up, it’s all for naught” [8]. The Retail Environment The motivation of a retail store opposes the user’s intention; a retail store wants shoppers to spend as much time as possible in the store because this can directly translate to more sales. The longer a shopper is in a store, the more likely they are to buy something [8]. According to market research, almost 60 % of all products bought at the grocery store are not premeditated purchases [2]. Underhill’s fieldwork techniques present the environmental psychology behind public spaces and reveal how important the store and the aisle are to the total shopping experience. Paper: Designing Applications for Handheld Devices CHI 2003: NEW HORIZONS 338 Volume No. 5, Issue No. 1
Ft. Lauderdale, Florida, USA. April 5-10, 2003 Paper: Designing Applications for Handheld Devices At an early stage in our analysis we too realized the cannot be whittled down into one succinct description importance of this physical space and began to deconstruct Therefore, we developed a high-level profile of a shoppe who might benefit most from a mobile shopping aid. This In our initial fieldwork we observed three major grocery profile involves a person who would frequent the same stores, analyzing the layout and product location within the grocery store and use a"loyal shopper"card store. We found that most of the time the general location Grocery stores attempt to lure shoppers through various of products were the same across stores with only the end sal Includ ng"loyal shopper aps displaying varying contents. A grocery store is an cards. Specifically, these cards entitle the bearer to overw whelming repository of information in the form of numerous items at a lower price than those who do not visual, tactile, aural and olfactory stimuli. There are signs, have a card. While shoppers believe that the cards are a labels and prices to read, produce to squeeze and items togimmick'", many still use the cards when they grocery grab from shelves. In addition there are noise factors, shop. The use of these cards enables the grocery store to scanners beeping, cash registers buzzing and carts rolling. track what and when an individual buys Music is playing, and announcements are being made while IOVING FROM TASK TO DESIGN conversations are taking place. All these factors serve, even What tasks in the shopping process could be aided by a implicitly, to keep the shopper in the store and influence mobile context? Answering this question required studying their spending the physical task of shopping as well as people's internal As Nardi points out, it is not the store or the individual the individual objectives. Common among shopping are three separate shopper that require analysis but, instead, the relationship high-level tasks; pre-shopping, in which a person plans or between the two [7]. In the store, the shopper is presented decides to go shopping, shopping, where the person is with an enormous retail environment that they must"edit" actually in the store shopping, and checkout, where the according to their task at hand. According to Lave, a person makes payment for the items and leaves the store shopper in a grocery store experiences the store as a The time spent in preparation outside of the store is personally ordered, edited version"of the larger store [6]. contextual and results in a list being created whether it is a shopper may shop" only for certain items in certain physical or mental. It can be a collaborative effort or a last aisles, depending on needs and habits. Thus the store has minute solo act prior to visiting the store.Regardless,the been"edited to"match personal preferences"[7]. We result is the creation of the shopping list. fer to this as the edited space While shopping, we found that shoppers want more Our research supports the notion of an edited space,even knowledge. control. and convenience. In contrast the retail more so when the shopper is engaged in interacting with a industry constructs the store environment to move a mobile interface. As mentioned earlier, grocery shopping is shopper in a contrived route, having constructed zones to a "hands-busy, eyes-busy"task. Shoppers are already so influence the intentions of the shopper. These competing involved in finding what they want that adding PDa goals and the complexities of the act of shopping offered interactions may create further distractions and further more possibilities for designing the relational interactions editing of the store Any design in this area must carefully between the store and the shopper. Therefore, we decided consider this fact to focus our design on the actual act of shopping The Shopper's Profile Our observations within the store revealed that most people We studied many varied shoppers, single and married, male nd female, mixed in age. We found innumerable have a good understanding of the store environment and characteristics of shoppers across a range of demographics how to act in it. "The integration of human-computer Lifestyle emerged as a relevant characteristic because it interaction and the user's situation in a particular working xt In determines how people shop. If a shopper is purchasing interaction"[4]. Situated interactions, where the user roceries for a household, the list may be a collaborative utilizes the interface while shopping, became the greatest effort where communication is important. If a shopper is on challenge for designing the mobile interface. We were a budget then coupons and specials become relevant to their buying decisions challenged with identifying appropriate times of interaction where a situated application could offer assistance to the In general people shop in grocery stores with which they user while shopping are familiar In our survey 72% visit the same grocery store During the task analysis, we uncovered breaks in the shopping task where a situated application could special case if a person shops at a grocery store they do not complement the task at hand and assist a person with just- frequently visit. Similarly, shoppers tend to navigate the in-time information. For example, one break could happen while searching for a product in the store, or trying to Since grocery shopping appears to be based on lifestyle and remember an item to purchase. The application at this point personal pre pecific pattern of shopping would provide a clue or reminder specific to the situation emerges. It seems that the profile for a grocery shopper Volume no 5 issue no. 1
At an early stage in our analysis we too realized the importance of this physical space and began to deconstruct the arranged space within the grocery store. In our initial fieldwork we observed three major grocery stores, analyzing the layout and product location within the store. We found that most of the time the general location of products were the same across stores with only the end caps displaying varying contents. A grocery store is an overwhelming repository of information in the form of visual, tactile, aural and olfactory stimuli. There are signs, labels and prices to read, produce to squeeze and items to grab from shelves. In addition there are noise factors, scanners beeping, cash registers buzzing and carts rolling. Music is playing, and announcements are being made while conversations are taking place. All these factors serve, even implicitly, to keep the shopper in the store and influence their spending. As Nardi points out, it is not the store or the individual shopper that require analysis but, instead, the relationship between the two [7]. In the store, the shopper is presented with an enormous retail environment that they must “edit” according to their task at hand. According to Lave, a shopper in a grocery store experiences the store as a “personally ordered, edited version” of the larger store [6]. A shopper may shop “only for certain items in certain aisles, depending on needs and habits.” Thus the store has been “edited” to “match personal preferences” [7]. We refer to this as the edited space. Our research supports the notion of an edited space, even more so when the shopper is engaged in interacting with a mobile interface. As mentioned earlier, grocery shopping is a “hands-busy, eyes-busy” task. Shoppers are already so involved in finding what they want that adding PDA interactions may create further distractions and further editing of the store. Any design in this area must carefully consider this fact. The Shopper’s Profile We studied many varied shoppers, single and married, male and female, mixed in age. We found innumerable characteristics of shoppers across a range of demographics. Lifestyle emerged as a relevant characteristic because it determines how people shop. If a shopper is purchasing groceries for a household, the list may be a collaborative effort where communication is important. If a shopper is on a budget then coupons and specials become relevant to their buying decisions. In general people shop in grocery stores with which they are familiar. In our survey 72% visit the same grocery store due to convenience and familiarity of store layout. It is a special case if a person shops at a grocery store they do not frequently visit. Similarly, shoppers tend to navigate the store using the same route. Since grocery shopping appears to be based on lifestyle and personal preferences, no one specific pattern of shopping emerges. It seems that the profile for a grocery shopper cannot be whittled down into one succinct description. Therefore, we developed a high-level profile of a shopper who might benefit most from a mobile shopping aid. This profile involves a person who would frequent the same grocery store and use a “loyal shopper” card. Grocery stores attempt to lure shoppers through various sales, specials and offerings including “loyal shopper” cards. Specifically, these cards entitle the bearer to numerous items at a lower price than those who do not have a card. While shoppers believe that the cards are a “gimmick”, many still use the cards when they grocery shop. The use of these cards enables the grocery store to track what and when an individual buys. MOVING FROM TASK TO DESIGN What tasks in the shopping process could be aided by a mobile context? Answering this question required studying the physical task of shopping as well as people’s internal objectives. Common among shopping are three separate high-level tasks; pre-shopping, in which a person plans or decides to go shopping; shopping, where the person is actually in the store shopping; and checkout, where the person makes payment for the items and leaves the store. The time spent in preparation outside of the store is contextual and results in a list being created whether it is physical or mental. It can be a collaborative effort or a last minute solo act prior to visiting the store. Regardless, the result is the creation of the shopping list. While shopping, we found that shoppers want more knowledge, control, and convenience. In contrast the retail industry constructs the store environment to move a shopper in a contrived route, having constructed zones to influence the intentions of the shopper. These competing goals and the complexities of the act of shopping offered more possibilities for designing the relational interactions between the store and the shopper. Therefore, we decided to focus our design on the actual act of shopping. Our observations within the store revealed that most people have a good understanding of the store environment and how to act in it. “The integration of human-computer interaction and the user’s situation in a particular working context in a mobile environment is identified as the situated interaction” [4]. Situated interactions, where the user utilizes the interface while shopping, became the greatest challenge for designing the mobile interface. We were challenged with identifying appropriate times of interaction where a situated application could offer assistance to the user while shopping. During the task analysis, we uncovered breaks in the shopping task where a situated application could complement the task at hand and assist a person with justin-time information. For example, one break could happen while searching for a product in the store, or trying to remember an item to purchase. The application at this point would provide a clue or reminder specific to the situation. Ft. Lauderdale, Florida, USA • April 5-10, 2003 Paper: Designing Applications for Handheld Devices Volume No. 5, Issue No. 1 339
Paper: Designing Applications for Handheld Devices CHI 2003: NEW HORIZONS From the task analysis of shopping, and the user-preferred list. Shoppers can shop for the items on the list, check off Teatures, our design criteria emerged Top on our list was to items as they acquire them, review and add specials to the develop interactions in the mobile interface, which list, view and save recipes and watch for in-store coupons complement the situations that users encounter while and specials, Preferences would allow the user to modify shopping. The grocery list became the cohesive force these features achieving this criterion. The grocery list is the key artifact Upon checkout the shopper scans the grocery items and the used within the system of grocery shopping because it is a system compares the original list with the items that have representation of the premeditated intentions of the user been scanned. An updated"scanned"list is beamed to the while in the store. By creating a shopping list, a person has PDA allowing the shopper to verify the total of the grocery specified particular situations for the context of shopping. bill. When satisfied with the purchases, the shopper can The presentation of a list on the PDa merges the interface into the context of the store beam the absolved list along with payment information back to the checkout. A receipt is beamed back to the PDa Early Ideas and Concepts for the shopper to later reconcile with a checking account In order to ensure that we thoroughly canvassed possible One of the resounding complaints we heard from design ideas, an early design activity focused on creating cenarios in which a PDA shopping application would be participants was that checkout lines were too long and that useful. The scenarios were wide-ranging and were aiting in line to checkout was the worst part of grocery shopping. Although the scenario addresses this issue motivated by real shoppers who described how it is that alleviating the checkout line would be another research they shop, what frustrates them about grocery shopping and how their shopping could be improved. Eventually, we project in and of itself. Design Evolution and Rationale implemented today and offers many of the features that The scenario in combination with our task analysis prepared us for the next stage, in which our concepts were In the scenario, the local grocery store contains an"always interpreted into a visual language, the interface. W on"information system. The shopper, who is a member of attempted to merge the features from the design scenario with the appropriate mappings and interactions from the the local grocery store's frequent shoppers club, would be immediately recognized upon entering the store situated space into a mobile interface There are four types of lists the shopper can use We presented three interface designs to a focus group session for feedback. While the first design reflected a 1)A grocery list from the store based on their previous basic Palm OS application, our second and third design purchases. This list is generated by the store and is created attempts were more appropriately focused on the spatial based on the frequency of the shopper's past purchases. A and contextual aspects of shopping shopper who frequents the store on a regular basis enables The spatial design, shown in Figure 2, incorporated a store the store's information system to predict their grocery map that displayed icons to show where items on the grocery list were located. This idea was well received 2)A personal grocery list created in the home however, the store map commanded most of the screen real 3)A grocery list that family members have phoned in or tate, resulting in problems with occlusion when other interactively paged to the system. When a shopper joins the features of the interface, such as coupon alerts or search club, a profile can be created for every person the shopper functions, appeared in pop-up windows cares to include. Therefore, those items that family The contextual design, shown in Figure 3, served two members need or items that, for instance, are remembered purposes. In a person's home, the interface served to aid the in the middle of a meeting, can be captured as they are shopper in creating a shopping list. Therefore, the interface was divided in to four portions: a refrigerator, a kitchen 4)A combined list of all three. Those items that may be pantry, a utility closet and an"additional items"area.Since duplicated across the lists are consolidated and one master we discovered that lists were created contextually, these list is created four segments provided context for the shopper when dentify ing items needed at the store. Once in the store, the The availability of these different lists attempts to reach all interface retained four segmented areas, but they were now lists can still do so using the application. Those who do not and the grocery list created at home. This interface hay types of list-makers. Those who currently make grocery divided into a store map, a promotional area, a grocery total make lists are also catered to through the store-generated merits but sacrificed some of the spatial designs features st. And for those who want to ensure they have gotten such as the size of the store map and the location of the verything they need in one trip, the master grocery list items within it. serves them best The focus group helped us to advance through the first Once the shopper enters the store, the list is reordered to iteration of an interface. Continuing to balance the data provide the most efficient route to obtain every item on the gathered from our survey and observations with the Volume no. 5. issue no. 1
From the task analysis of shopping, and the user-preferred features, our design criteria emerged. Top on our list was to develop interactions in the mobile interface, which complement the situations that users encounter while shopping. The grocery list became the cohesive force for achieving this criterion. The grocery list is the key artifact used within the system of grocery shopping because it is a representation of the premeditated intentions of the user while in the store. By creating a shopping list, a person has specified particular situations for the context of shopping. The presentation of a list on the PDA merges the interface into the context of the store. Early Ideas and Concepts In order to ensure that we thoroughly canvassed possible design ideas, an early design activity focused on creating scenarios in which a PDA shopping application would be useful. The scenarios were wide-ranging and were motivated by real shoppers who described how it is that they shop, what frustrates them about grocery shopping and how their shopping could be improved. Eventually, we sculpted a final scenario that could possibly be implemented today and offers many of the features that shoppers want most. In the scenario, the local grocery store contains an “alwayson” information system. The shopper, who is a member of the local grocery store’s frequent shoppers club, would be immediately recognized upon entering the store. There are four types of lists the shopper can use: 1) A grocery list from the store based on their previous purchases. This list is generated by the store and is created based on the frequency of the shopper’s past purchases. A shopper who frequents the store on a regular basis enables the store’s information system to predict their grocery needs. 2) A personal grocery list created in the home. 3) A grocery list that family members have phoned in or interactively paged to the system. When a shopper joins the club, a profile can be created for every person the shopper cares to include. Therefore, those items that family members need or items that, for instance, are remembered in the middle of a meeting, can be captured as they are thought of. 4) A combined list of all three. Those items that may be duplicated across the lists are consolidated and one master list is created. The availability of these different lists attempts to reach all types of list-makers. Those who currently make grocery lists can still do so using the application. Those who do not make lists are also catered to through the store-generated list. And for those who want to ensure they have gotten everything they need in one trip, the master grocery list serves them best. Once the shopper enters the store, the list is reordered to provide the most efficient route to obtain every item on the list. Shoppers can shop for the items on the list, check off items as they acquire them, review and add specials to the list, view and save recipes and watch for in-store coupons and specials. Preferences would allow the user to modify these features. Upon checkout the shopper scans the grocery items and the system compares the original list with the items that have been scanned. An updated “scanned” list is beamed to the PDA allowing the shopper to verify the total of the grocery bill. When satisfied with the purchases, the shopper can beam the absolved list along with payment information back to the checkout. A receipt is beamed back to the PDA for the shopper to later reconcile with a checking account. One of the resounding complaints we heard from participants was that checkout lines were too long and that waiting in line to checkout was the worst part of grocery shopping. Although the scenario addresses this issue, alleviating the checkout line would be another research project in and of itself. Design Evolution and Rationale The scenario in combination with our task analysis prepared us for the next stage, in which our concepts were interpreted into a visual language, the interface. We attempted to merge the features from the design scenario with the appropriate mappings and interactions from the situated space into a mobile interface. We presented three interface designs to a focus group session for feedback. While the first design reflected a basic Palm OS application, our second and third design attempts were more appropriately focused on the spatial and contextual aspects of shopping. The spatial design, shown in Figure 2, incorporated a store map that displayed icons to show where items on the grocery list were located. This idea was well received; however, the store map commanded most of the screen real estate, resulting in problems with occlusion when other features of the interface, such as coupon alerts or search functions, appeared in pop-up windows. The contextual design, shown in Figure 3, served two purposes. In a person’s home, the interface served to aid the shopper in creating a shopping list. Therefore, the interface was divided in to four portions: a refrigerator, a kitchen pantry, a utility closet and an “additional items” area. Since we discovered that lists were created contextually, these four segments provided context for the shopper when identifying items needed at the store. Once in the store, the interface retained four segmented areas, but they were now divided into a store map, a promotional area, a grocery total and the grocery list created at home. This interface had its merits but sacrificed some of the spatial design’s features such as the size of the store map and the location of the items within it. The focus group helped us to advance through the first iteration of an interface. Continuing to balance the data gathered from our survey and observations with the Paper: Designing Applications for Handheld Devices CHI 2003: NEW HORIZONS 340 Volume No. 5, Issue No. 1
Ft. Lauderdale, Florida, USA. April 5-10, 2003 Paper: Designing Applications for Handheld Devices feedback from the focus group, we redesigned the interface. and visual encoding. Of the 8 participants, 7 were not PDA It is at this point that crucial design tradeoffs occurred. Our users and 5 did not routinely make redesign is shown in Figure 4 For the evaluation a moderator walked the participants step by step through a shopping scenario. The overall system was described and each participant was shown two or three screens at a time( depending on the actions taking place on the interface). They were asked what action they would take on the interface to complete the intended task. Along the way the participants thoughts and suggestions were encouraged 容 o En Fit Figure 2: The spatial design 88 s VALUE Figure 4: The formative evaluation redesign The evaluation uncovered several interesting items Perhaps the most important was the fact that participants who were unfamiliar with how to use a pda offered mixed responses. On one hand they could not"traverse"the interface in a manner required when using a PDA. As a ted their actions to what they knew they could do with a mouse on a computer. On the other hand because they were not limited by knowledge of how to interact with PDAs, they offered insight through their Figure 3: The contextual design interpretations and comments. Additional findings were Locations of interface functions (buttons or icons) needed Clearly, PDA displays have extremely limited space. The closer proximity to the objects or areas on which they take merging of the desired features from both the spatial and action. For example, the Search button in the upper middle contextual interfaces presented design obstacles. We would right opened a search field at the lower center of the have to use pop-up windows that would, at different time interface. Participants did not immediately see where their cclude the main interface. Designers for handheld devices action had effect. We speculated that this problem would be refer to this as"deck stacking". It is a metaphor commonly heightened when in a store environment using a PDA used on small screen real estate The iconography we used did not convey the inter In the task of shopping, people's priorities are the list of meanings to participants. We decided that the icons she items they need and knowledge of where those items are be replaced with text for clarity. We discovered that located. Therefore, we left the grocery list at the top of the designing icons for a handheld display is even more screen and devoted the center of the screen to a store complex than a desktop application because of the low layout. This redesign required a heavier use of icons due to resolution. One icon, an exclamation point, was placed on limited display area. The revolving promotional area was aisles where items from the list were located. At one point moved to the bottom of the screen so that the user could during the evaluation, several exclamation points were on continue to be informed and take advantage of store the screen and a participant said this made him feel coupons and speci anxious and overwhelmed -almost stressed out from We mocked-up this design into nine screens for a paper shopping. This was not our intention for the icon an prototype, and evaluated it for the information architecture Volume no 5 issue no. 1
feedback from the focus group, we redesigned the interface. It is at this point that crucial design tradeoffs occurred. Our redesign is shown in Figure 4. Clearly, PDA displays have extremely limited space. The merging of the desired features from both the spatial and contextual interfaces presented design obstacles. We would have to use pop-up windows that would, at different times, occlude the main interface. Designers for handheld devices refer to this as “deck stacking”. It is a metaphor commonly used on small screen real estate. In the task of shopping, people’s priorities are the list of items they need and knowledge of where those items are located. Therefore, we left the grocery list at the top of the screen and devoted the center of the screen to a store layout. This redesign required a heavier use of icons due to limited display area. The revolving promotional area was moved to the bottom of the screen so that the user could continue to be informed and take advantage of store coupons and specials. We mocked-up this design into nine screens for a paper prototype, and evaluated it for the information architecture and visual encoding. Of the 8 participants, 7 were not PDA users and 5 did not routinely make physical grocery lists. For the evaluation a moderator walked the participants step by step through a shopping scenario. The overall system was described and each participant was shown two or three screens at a time (depending on the actions taking place on the interface). They were asked what action they would take on the interface to complete the intended task. Along the way the participant’s thoughts and suggestions were encouraged. The evaluation uncovered several interesting items. Perhaps the most important was the fact that participants who were unfamiliar with how to use a PDA offered mixed responses. On one hand they could not “traverse” the interface in a manner required when using a PDA. As a result, they limited their actions to what they knew they could do with a mouse on a computer. On the other hand, because they were not limited by knowledge of how to interact with PDAs, they offered insight through their interpretations and comments. Additional findings were: - Locations of interface functions (buttons or icons) needed closer proximity to the objects or areas on which they take action. For example, the Search button in the upper middle right opened a search field at the lower center of the interface. Participants did not immediately see where their action had effect. We speculated that this problem would be heightened when in a store environment using a PDA. - The iconography we used did not convey the intended meanings to participants. We decided that the icons should be replaced with text for clarity. We discovered that designing icons for a handheld display is even more complex than a desktop application because of the low resolution. One icon, an exclamation point, was placed on aisles where items from the list were located. At one point during the evaluation, several exclamation points were on the screen and a participant said this made him feel anxious and overwhelmed –almost stressed out from shopping. This was not our intention for the icon and we Figure 2: The spatial design Figure 3: The contextual design Figure 4: The formative evaluation redesign Ft. Lauderdale, Florida, USA • April 5-10, 2003 Paper: Designing Applications for Handheld Devices Volume No. 5, Issue No. 1 341