Project proposal
- 1. User Centred Design Project 04.12.2015 ─ Diego Garaialde, Peter Clarke, Satish Narodey, Cathal O’Leary, & Lynn Earls
- 2. 1 Introduction Grocery shoppingisanarea of immense interestand growth at the moment, which makes it an ideal market for revolution in technology and design (Galante, Lopez & Munby, 2013; Warschun, Delfmann, & MuBig, 2012). We were assigned with the task of venturing creatively into this field to create an innovative shopping experience for a fictional supermarket “PriceValue”. From our beginnings, we first gathered and introduced ourselves in a small conference room. We triedtocreate an atmosphere of thinkingoutsidethe reigns of current online shopping trends. Our initial couple of sessionsweresomewhat restrained as perhaps contrary to our initial plan. Unease has never found itself a feature within a creative stronghold. Despite our early tardiness, we eventually loosenedupandourideas began to flow and unravel. In order to be innovative in our thinking, fi rst of all we decided that we needed to find out what was currently available or ‘out there’ in the online grocery shopping world. To do this we each took upon the task of researching around the area individually and reporting back to the group with our different findings. Research and Requirements Gathering The goal of gatheringliterature wastocollectivelygraspthe fieldof online shopping,itscurrent trendsand mostimportantlytofindoutwhypeople seemedtoprefertoshopinan ‘oldfashioned’way despite the remarkablepotential technologycouldholdforit.Currentliterature onthe subjectmay informusof howmoderninnovatorsare viewingonline shoppingandperhapsmaysparkan ideaaswell as informuson the topic.Undoubtedlythere isgreat potential tomake moneyfromonline shoppingin thisday and age,yetno organisationhasproperlybrokenthroughwithanoutstandingdevice oridea that people genuinelywanttouse.We alsowanteddiscoverthe general opinionspeopleholdtowards online shopping,barriersthatpreventtheiruse,andthe motivationsbehindtheircurrentusage patterns. Of the informationgatheredduringthe initial research,one particulararticle thatstoodoutwas entitled‘The Future of Online GroceryinEurope’.Here Galante andcolleagues(2013) suggestthat there isa lotof potential formarketgrowthinthe fieldof onlineshopping,andthatpeople wanttoshop online butjustdon’tfeel thatthe effortinvolvedisworthit.Theygo onto say that a clickand collect system(selectingandpurchasingonline butcollectingfromthe store) maybe of greatpopularityinthe
- 3. 2 future,cuttingoutthe unpopulardeliverycharge.Essentially,thismethodof drive throughshopping allowsthe customertoavoidbothqueuingand carryingtheirbags to theircar. The authorscontinue to suggestthatthe reasononline shoppinghasn’ttakenoff isithasadditional costsinthe line of labour deliveries,vehiclesandtheirfuel andmaintenance costs,withcustomersreluctanttofork outthe extra expensesforthe additional convenience.Theycontinue tosuggestthatthe progressof online grocery shoppingdependsona single retailertomake the firstboldandassertive move towardsinnovationfor the othersto quicklyfollow suit.Theymake averygoodpointaboutthe requirementof aneasyto use and effective interface,asonline shopping experiencesdiffergreatlyacrossgrocers,yetare still severelylackingconsideringhowmanyitemsthatare purchasedona weeklybasis.If the interface is difficulttouse anditemscan’tbe located,itcan resultinthe user gettingfrustratedandabandoning theirtransaction.The article continuestooutlinepotential risksinvolvedinonlinegroceryshopping, such as systemcrashesthatcan resultinhavingto start all overagain,a late deliverythatleadsto ruineddinnerplansandunclearlabellingresultinginthe wrongitemsbeingpurchased. Anothernoteworthyarticle thatwasencounteredisentitled‘A FreshLookat Online Grocery Shopping’.Here Warschunandcolleagues(2012) declare thatthe major hurdle tobuyinggroceries online isthatpeople are generallyalreadyhappywiththe waytheyobtaintheirgroceries.Thiswould suggestthatin orderfor online shoppingtotake off,customersmustprefertodotheirshoppingonline above doingitinperson.Anotherimportantpointmade inthisarticle worthmentioningisthatthere is verylittle roomforerror inselectinganddeliveringgoodsonline,tosuchan extentthata usermay neverattempttouse the service againif a product isincorrector if a deliveryisn’tquickandefficient enoughfortheirstandards. From our researchanddiscussionswe foundthatacommonkeybarrierthat put people off online groceryshoppingwashavingtodeal withdifficultinterfaces.Anotherirritableflaw thatseemed to regularlyappearwasthe cost of deliveryandthe durationthatittook forthe groceriesto arrive, particularlywithinpopulartime slots.People generallywanttheirfoodthe same daythattheyorderit, and thislimitationof currentmethodsseemedtobe a keydeterrentinonlineshoppingoverall. Additionallyif the costof deliveryisquite highitcanbe a matterof principle tonotwant to have to pay a great deal fora generallypoorservice.Anotherservice thatwaswidelyavailable instoresbutnot online,wasthe abilitytoavail of popularoffersandreduceditems.Peopleseemtolove offersand catchinga deal evenif theydidn’tsetouttobuythe itemorperhapseven wantit,so itwouldbe
- 4. 3 importantto make the customerfeel asthoughtheyweren’tmissingoutonanythingbyavoidingthe supermarket.We foundthatdifferentpeople have differentstandardsof selectinggoodswhile shoppingsoitwouldbe importanttomaintaina veryhighstandardto satisfyeverybody’sneeds, providingthe freshestproductsandensuringthatthe packagingwasn’tdamagedorout of orderin any way. We foundthat eventhe ideaof online shoppingpeoplefoundtobe quite tediousand uninterestingsobearingthisinmindwe wantedtocreate a funand enjoyableexperience outof the online shoppingtask. Afterthe groupfeltthat theyhadcollectedsufficientinformationtounderstandthe underlying potential andlimitationsof the online shoppingexperience,theystartedtobrainstormideasforthe project.Itwas agreedthat a focuson creatinga betteronline shoppinginterface wasnotenough,and insteadthe groupdecidedtothinkof interestingandnovel experiences centredaroundfood. The aim was to create a productthat people wantedtouse initsownright,because of the functionalitythe productitself offered,whichcouldthencouldbe connectedtothe worldof online groceryshopping. Thisallowedustobroadenour scope for ideaswhichfocusedonuserexperience attheircore. From a broad varietyof potential ideasvaryingfromvirtual realityandgamificationconceptsof shoppingtoa smart trolleythatwouldscanand weighitemsastheywere placedintothe basket,we eventuallydecidedonaSmart Fridge anda CookbookAppas our final twochoices.Bothseemedto have the most potential incatchingonwiththe publicandalthoughthe previousideaswere goodthey perhapslackedthe depth,dimensionanddurabilitythatthe final twopossessed.The smartfridge idea rose out of current interestinthe “internetof things”,amarketof massive potential andinnovation (Kortuem,Kawsar,Fitton,&Sundramoorthy,2010).The pitchwasthat smart algorithmscouldbe combinedwithavarietyof sensorstoallow forthe analysisof fridge contents,andthisinformation couldbe effectivelyandappropriatelyparsedforuseful informationtobe impartedtothe enduser.The Cookbookapptookinspirationfromthe currentinterestinrecipe andmeal planningapplications (Breton,Fuemmeler,&Abroms,2011) buttook advantage of the potential integrationwithonline shoppingforan easierandmore streamlinedmeal preparationexperience.Itwasthoughtthatif people enjoyedusingthe designedproduct,andthe integrationwithPriceValueallowedforaneasierwayto shopfor groceries,itwouldincrease the amountof people shoppingonlineaswell asincreasing PriceValue’srevenue.Throughtheirincreasedrevenue,theywouldbe able topayfor developmentand improvementsof suchproducts,whichwouldinturnleadtomore revenue.The popularonlineretailer
- 5. 4 Amazonhas gaineditsmarketdominance throughasimilarmodel,where thereisahighinitial investmenttoincrease brandawarenessandshoppinginteractions(Rayport,1999).Customerswould thenbe enticedintoanecosystemthatwouldpromote highlevelsof transactions,whichwould eventuallyleadtoanexponential increase inprofitsandinvestmentreturn. A targetdemographicwaschosen foreach projecttoallow forgreaterfocus onspecificuser requirements,whichwouldthenleadtouseful,applicableandtailoredsolutionsthatwouldserve to entice the userto increase theirshoppinghabits.The maindemographicdecidedforthe Cookbookapp was students,astheyare oftenthrustintoforcedindependence,whichisperfectforthe realmof followingrecipesandplanningmeals(Holdsworth,2009).Parentswere decidedtobe the best demographicforthe smart fridge,astheyare typicallyhomeownerswhowouldhave aneedto purchase theirownfridge.The familyunitwouldalsoleadtolargerlevelsof interactionwiththe appliance,whencomparedtostudentsandbusyprofessionalswhichare more inclinedtowardstake out,or the elderlywhichare tendtorejectnew technology(Rouillard,2012). From here itseemedappropriatetoconductan online questionnaire tofindoutabouthowboth studentsandfamiliesfelt aboutthe appandfridge.Anonline surveyseemedtobe the mostappropriate and efficientmethodtogatherinformationfromthe publicasitallowedustoaccessa large pool of informationquickly,andwasquite easytoadminister.Additionaltothis,studiesconductedbyLuce and colleagues(2007),Smyth,Dillman,ChristianandStern(2006) and Funke,ReipsandThomas(2011), revealedthatparticipantsare more revealingwithinthismediumwhilealsoelaboratingfurtherontheir responses.We formulatedaconsent formto beginwithinordertooutline whatinformationwe were investigatingandwhywe were lookingforit.We thengatheredsome general demographicinformation to give some contexttosubsequentresponses. Aftergatheringthe resultsfromthe SmartFridge questionnaire,we foundthatpeople preferred to shopin a store as it waseasiertocompare itemsandsource the bestdealsandoffers. Trustin the supermarketstaff wasanotherimportantfactor,ascustomerspreferredtoselectthe freshestmeats and vegetablesthemselves,asgenerallythe date of the goodsselectedbythe supermarketstendedto suitthe store rather than the customer.People were alsoveryunhappywiththe shopssubstitution system.Theydidn’twantthe hassle of receivingthe wrongproductasit wasenoughto irritate the shopperandput themoff shoppingonlineagain.Technologyanddifficultieswithaninterface was anothercommonresponse andwasenoughforpeople toavoidonline shoppingcompletely.With
- 6. 5 regardto the CookbookApp,similarconcernsoverthe interfacesusabilitywasof greatimportance and occurredfrequently.Anotherimportantnote wasthatcurrentdietaryappsfor those withrestricted dietsneverconsideredthe user’sbudget,renderingthe suggestedmealsquiteexpensive.Bothof the surveyssuggestedthatboththe appand the fridge were innovativeandwouldbe usedbythe vast majorityof the participants. Creative Design Withthe resultsfromthe userquestionnaire, we were able to organise a database of common issuesusersencountered.The database wasfilteredforthe mostsalient and important concerns raised by participants, which could then be formulated into a set of user requirements. Producing a set of stable requirements from users means that you create a sound foundation in which to build design solutions, and allows for an end product that is focused and targeted at improving user experience (Chavan et al., 2009; Rogers et al., 2015). Requirements were separated into three categories: Fridge Specific, App Specific, and Mutual Requirements. Due to the underlying theme of online grocery shopping, some requirements were consistent across both product ideas, and this allowed for the formulation of solutions that were subsequently applicable to both products as well. One issue that many participants expressed was that they did not trust the supermarket to deliverthe bestqualityperishablestothem.Complaintswere raised about bruised fruits or vegetables that did not look particularly fresh, as well as meat and dairy products that were already close to their expiration dates. Basically, customers had certain standards for different foods, and did not trust the employeesthatwere selectingtheiritemstodosoto theirunique specifications.A proposed solutionto this issue was named “Shopper Trust”, and involved a rating system that would allow users to leave public ratings and feedback on their specific shopper. In principle, this solution aimed to allow for customers to vocalise their preferences and received suited service accordingly. Deliverywasarecurringtopicinthe userresponsesasa large deterrentagainstonline shopping. Waiting for deliveries within certain time slots, no same day delivery and high delivery price were all commonlystatedconcerns.Customerswould feel like they weren’t able to commit themselves to any
- 7. 6 task within the time slot, as they could be interrupted at any time to accept the delivery. A proposed solution to this issue was a real-time delivery tracker, which would give an estimate to arrival time, allowingforgreatertransparencyanduser satisfaction. This solution was inspired by applications such as Uber, as well asreal-time bustimetablesandDomino’sPizzadelivery indicator. Same day delivery is beingcurrentlyexplored by both Amazon and Google in the United States to great customer feedback (Zych,2015), thusinspiringthe team to include this functionality to potentially solve this issue. Lastly, deliveryprice wasindicatedtobe toohigh,especially when the number of items ordered is low. Again takingfromthe Amazonbusinessmodel (Rayport,1999), we decidedtoinclude free delivery whenever possible. It would seem reasonable to include free delivery when the customer spends over a certain amount, earns enough loyalty points to balance the cost, or as a service that comes packaged with purchasing the fridge, and so these principles were included in our design. When purchasing items online, customers frequently stated that there wasn’t enough information onthe product,andthisleadto erroneouspurchasing of wrong items or items that are too large. As such, we proposed that PriceValue would invest in creating a large, extensive and detailed database of all the products they currently sell, so that information could be presented clearly to the customer, but also incorporated into the smart algorithms behind the two products. Nutritional value was alsoimportanttocustomers,andso includingthisinformation in the database would allow for the automatic creation of meals and recipes that met certain criteria, such as dietary needs or allergies. Particularly, the fridge was seen to benefit immensely from this database, as it could communicate effectivelywiththe appliance tocalculate quantitiesandother relevantdetailsbasedonthe contents of the fridge.Thiscommunicationbetweensmartalgorithmswouldallowforsuggestions,notificationsand extradetailstobe presentedtothe customer when needed, such as notifying them when certain food items were running low, or that certain food was about to expire. The database could also connect to the customer’sorderinghistory,toallow forthe emergence of shopping trends that could be exploited for furtherautomation.Thisdatabase isalsoimportantfor the organisation of meals and recipes in the Cookbook app, as questionnaire participants were very interested in getting suggestions based on budget and dietary requirements. With our solutions formulated, we could move on to create scenarios that would show the systeminaction,allowingforthe creationof contextaroundthe userexperience,aswell ashighlightthe tasksthe userwill perform(Carroll,2000).To allow fordiversityindesignideaswheneventually moving
- 8. 7 to a lo-fi prototype, we created essential use cases (Constantine & Lockwood, 1999), user groups (Rogers et al., 2015) and storyboarding scenarios (Jacobson, Christerson, Jonsson, & Overgaard, 1992; See Appendixfolderforexamples),whichallowedforadiverse spreadof useractionsand reactions. For example, a storyboard was sketched of the different situations in which notifications could be used within both the designed products to impart information which was thought to be relevant to the current situation. Detailed personas were also created to induce a more empathetic perspective towards users, as the team could make decisions with specific fictional characters in mind (Dix, 2004; Rogerset al.,2015). One of the designedpersonaswasastudentnamedMaud,whohad specificdietary requirements as she was allergic to nuts and a vegan (See Appendix folder for further examples). Once enoughappropriate andrelevantdocumentationwascreated,we were prepared to move onto the next step of design, the low-fidelity prototype. The conceptualisation of our ideas into a prototype wouldthenallow us to evaluate our solutions with the end-users, to gauge their efficacy as solutionstotheirraisedconcerns,aswell asgainingadiversityinperspective regarding how they could be improved (Schon, 1983). Low Fidelity Prototype and Evaluation Low-fidelity prototypes are a cheap, efficient and easy way to transform abstract ideas into tangible andtestable artefacts(Rogersetal.,2015). Prototypescanrange fromwoodenblocks,sketches on paper to complex pieces of software (Schon, 1983), but all are built with the purpose of testing a specificconcept or feature (Baskinger, 2008). Our team had two purposes in mind when designing our low fidelity prototypes: to show users a more concrete representation of the potential experience available fromthe product,andto gauge engagementandexcitementtowardsthe products as a whole. To create our prototypes, we focused on the method of sketching outlined by Greenberg, Carpendale, Marquardt, and Buxton (2012) as it is the simplest form of artistic expression and communicative tool. Artistic talent is not a requirement of sketching as the focus lies on the design instead of the aesthetics. Lo-fi prototypes of different interfaces, such as the ones described by Baskinger’s (2008) index cards, get designers into the mind-set of graphical user interfaces, as well as navigation,buttonplacementsandinterface metaphors(Rogersetal.,2015). Lo-fi prototypes carry a lot
- 9. 8 of advantages (Dix, 2004) which were behind our reasoning for dedicating a large portion of our resources to their extensive use. These include cross-team communication, clarification of requirements,formation of design alternatives, extremely low cost and its low requirements to entry (Dhillonetal.,2011). However,there were certain limitations attached to this undertaking. The length of the development cycle was greatly increased, especially as all team members were involved in designing their own subsections of our products. Additionally, this compartmentalisation of design meant there was little consistency and integration between interfaces. When presenting these prototypestothe user,thislack of underlyingsequence couldhave negativelyimpacted the evaluation. The team devised a structure of components required for each product prototype, with some overlapwhenfeatures connect to more general online shopping were involved. In total, 19 interfaces items were planned for the fridge and 18 interface items were planned for the cookbook app (See Appendix folder for further details). These included sketches of icon placements and navigational structure for itemssuch as the delivery tracker, the shopper trust system and the substitution system. For the purposesof evaluation,the imageswere amalgamated into a PowerPoint presentation that was shown to participants as an illustrative tool while we conducted a structured interview with them. The team member ran through a prepared script while the participant sat in front of the demonstrationof the application,where theycouldrelaytheiropinionsaboutthe productfreelyasthey saw it.Thiscognitive walkthrough technique allowed for the greater insight into the exact thoughts of the participants and their feelings toward the product (Dix, 2004). Overall,the responsestothe lo-fi were overwhelmingly positive, with participants exclaiming theyreallylikedbothproductideasandwere excitedto see a finished product. One participant stated, “The idea that the fridge can tell me if the items I am purchasing will fit is amazing”, with another interested in the glass door interface, “The glass door is a great idea. Why don’t any fridges have this already? Being able to overlay information on the contents would be fantastic – it’s kinda Star Trek”. Participants also highlighted that they enjoyed the personal touch that the Shopper Trust solution offered, and that they thought it could become “a hit in real stores”. It is important to state that the script used to evaluate the lo-fi prototype did carry a very positive tone, highlighting the potential advantages and benefits of the system which could have biased the participants towards positivity in theiropinions.Thisconfoundwillneedtobe addressedinfurtherevaluationstoprevent contamination of opinion between designers and users.
- 10. 9 The information collected was analysed to see what advantages and disadvantages each product offered so we could better decide which product to move forward onto the hi-fi prototyping stage (See Appendix folder for further examples). The fridge received positive reactions towards its smart notifications and suggestions based on the contents, and customers seemed to react positively towardsthe experience.However,withthe parentdemographic,careful care would be needed to keep usability as clear and easy to navigate as possible, as they may not be used to normal technological conventions like students would be (Rouillard, 2012). Positive responses were also seen for the cookbook’sabilitytoorganise andsuggestmealsbased on price and dietary preferences, as usual apps of the nature would regularly suggest healthy but expensive meals. Additionally, the student demographicismore eagertoadopt new technologies,andwould be more receptive to this product in general. Cookbook applications are very common in the current market, so the PriceValue offering wouldneedtobe extensivelyfeaturedandverywelldesignedtostandabove the restof the market and capture an audience. Takingall considerationsmentionedabove,the teamdecided to move on with the smart fridge productas itwas seenasthe most innovative,creative and interesting solution to the online shopping problem.People seemedgenuinelyinterested in using the smart fridge and stated that they thought it would make their shopping experience easier, and thus increase shopping overall. Additionally, the featuresandideasconnectedtothe cookbookappcouldbe integrated into the smart fridge design at a later date, to create a more complete and feature rich product. High Fidelity Prototype The hi-fi prototype,while atime-consumingandresource-intensive commitment,isknowntobe an effective way to test functionality, define a navigational scheme, portray the look and feel of a product, and serve as a living manifestation of a product (Rogers et al., 2015). According to Lim and colleagues (2008), prototypes can act as a filter to emphasise particular key aspects of the product, which can then be presented to the end user for feedback. Taking the primary qualities that were perceived most positively and innovative by the previous evaluations, we were able to create a fundamental principlefordesign.The designthatemergedwouldallow foranefficient measure of new possibilitiesandunderlyinglimitationsof the product. Finally, the hi-fi prototype allows for previously abstract conceptual ideas to be concretised and externalised (Lim et al., 2008). While certain studies
- 11. 10 foundcomparable feedbackfromusersfrombotha lo-fi andhi-fi prototype, regardless of the different in interactivity and cost (Sauer & Sonderegger, 2009; Dhillon et al., 2011), the advantages of the hi-fi prototype appeared to be worth the commitment necessary to take the product design to the next iteration. However,asisinherentin the nature of prototypes, some compromises had to be taken. Time limitswere aneverpresentfactor,andsowe had to focuson horizontal prototypingstyle where arange of functionswere shownwithminimal detail andinteractivity. As the intention was to quickly test user opinion to the design, overall experience and functionality of the product, this approach was seen as appropriate.Afterdeepimmersioninthe requirementsoutlined by the users and their opinions on our proposedsolutions,we feltwe could create a conceptual model of user experience we desired for our users (Rogers et al., 2015). We decided to use a rapid prototyping application known as Invision (http://www.invisionapp.com/),which allows for easy and quick formation of high fidelity prototypes. These prototypescanbe easilyanimatedandcarriedacrossplatforms,which made the final evaluation easier to accomplish. The framework allowed us to create static images that could be animated into a clickable andinteractiveexperience.The interfacewasdesignedusing Photoshop,andthenuploadedto the Invision Platform and connected with other static images to create a coherent interface. Due to issuesof continuityof designseenwiththe lo-fi prototype, we chose to allow one designer full control of the interface,whichallowedforaconsistentvisiontobe carriedacrossthe whole interface.However, this also meant that it was more vulnerable to oversights and carried biases in perspective (Dix et al., 2004). The final interactive version shown to users can be found at https://invis.io/UX54TY1QT. The design included 13 screens that were the minimum necessary to portray the range of features the smart fridge would provide (See Appendix Folder for all images used). Interactivity remained low and was restrained to basic navigation between screens. This allowed for a focus on overall experience and functionality, instead of being bogged down on the more intricate details of extensive in-depth design. As prototyping is an iterative process (Beyer & Holtzblatt, 1997), the subsequentversionsof the productcoulddelve deeper into the specific details after general feedback was received from the overarching system. The first image was of an overlay screen that appeared over the glass door of the fridge. The overlaywould give information on items within the fridge, such as expiry dates, remaining content as
- 12. 11 well as suggestions or notifications. To prevent an overwhelming display that was too information dense,iconswere usedwhereappropriatewithminimal text.The iconsrelied on the design metaphors principlesoutlinedbyErickson(1990),which allow for the user to understand a product by transferring knowledge fromfamiliarconcepts.Tosuggestremainingcontentwithin packaging, a percentage metre was usedsimilartothose seeninbatteryindicators.Toindicate expiration dates, a clock metaphor was used. This clock is surrounded by an arrow to suggest a consumption or reuse, similar to the arrow foundinrecyclingicons.Anexclamationmarkwaschosenfor notifications, as they are a regular design conventionforitemsof importance.Lastly, a gear icon was chosen for the menu, as it’s conventionally knownto meansettings, or more features. The notification itself required actual text, to allow for the full explanation of the details being provided. The menuscreenwas dividedintoagrid-basedformat,as this is regularly seen with computers and phones, and is an easy to navigate and a familiar system. Icons were picked that were thought to best describe the main functions of the fridge, with titles underneath for ease of use. Overall, the system aimed to be aesthetically pleasing, with a visually focused interface with large, high quality images.The goal wasalso to dothe mostprocessingaspossible in the background, to allow the user to be presentedwiththe mostrelevantinformationatthe mostrelevanttime.Whetherthiswasclearfrom the presentation will be highlighted in the later evaluation. In the My Fridge section, items were highlightedthatwere abouttoexpire andwere runninglow,allowinguserstoordermore automatically or add it to their shopping list. The online shopping section focused on regularly shopped items and itemsthatwere runningout.Thiswas chosenbecause usersdescribedthe dull andrepetitiousnature of shopping online for items that they’ve bought over and over again. Usually there’s a set of items that shopperswouldsticktoo,toppingupwhenevertheyranlow on one item or another. The system could potentiallyguessthatitemswere runninglow,basedonthe analysisof frequencyatwhichcertainitems were bought, even if they’re not placed inside the fridge itself. The shoppertrust initiative received positive feedback from lo-fi evaluations, and so received prominence inthe hi-fi design.Twoscreenswere createdtobothexplainthe innovative features of the system, and to highlight the possible benefits it carries. Additionally, the substitution function and deliverytrackerwere populardesignsolutions,receivingprominence in the hi-fi prototype as well. The inspiration for the real time tracker came from services such as Uber, which give an estimation of the time until arrival of the vehicle.
- 13. 12 Withthe hi-fi prototype,detailshadtobe ironedout,suchas procedure of navigation,interface interaction and material, as well as nature of exposed content. This helped to further concretise the conceptual ideas and better highlight the potential of the product to users for evaluation. Final Evaluation Evaluatinginterfacesshouldbe seenasaprocess rather than one specific task (Dix, 2004). With thisattitude takentoheart,the group begana secondevaluation stage, with the goal of evaluating the hi-fi prototype of the smart fridge. Takingthe busy parentpersona,the hi-fi prototype was evaluated using a group of parents in a controlledenvironment. The evaluation took place with the user sitting in front of the demonstration application and the interviewer progressing through a pre-prepared script which highlighted features and sought specific responses from the participant. A cognitive walkthrough (Dix, 2004) was used for features that were not yet implemented in the prototype. By observing users’ reactions to the interactive prototypethe developersof the appliance couldgainagreaterunderstandingof the usability issues. This usability testing, the team felt, would encourage active feedback and highlight errors or reactions to interactions that could not have been imagined prior to the evaluation. The initial userrequirementswere quitebroadbutitwas hopedthatthrough refinement of the lo-fi prototype, the application would be a step closer to meeting these requirements. A number of elements of user experience were targeted during the evaluation; namely the shopping and shopper trust experience,the deliveryexperienceandthe userinterface.Greater care was taken at this junction to create a more neutral tone in the evaluation script, to avoid contamination of biases between designers and users, a potential issue with the lo-fi prototypes. Galante and colleagues (2013) have suggested that there is significant potential for market growthin the fieldof onlineshoppingandthatpeople want to shop online but just do not feel that the effortanddifficultyinvolved is worth it. With this in mind, the hi-fi prototype attempts to simplify the shoppingexperience byremovingthe more mundane aspectsof shopping with, for example, the smart fridge notifying users of the items they are running low on and suggesting additional items they may wish to purchase.
- 14. 13 In the requirements gathering stage it was identified that a key barrier to broader adoption of online groceryshoppingwasusershavingtodeal withdifficultinterfaces.Withthisinmindthe interface experience had to be handled carefully. The prototype attempted to simplify the process by reducing tasks to minimal touch interactions within the fridge’s application. Of the usersevaluated,the majorityhadissueswiththe glass door upon which the information overlayswouldbe presented. What had been assumed to be a practical and interesting design feature provedtobe a turn-off forusers.User1 and User 5 were each reluctanttohave theirfridge contents on general display.User3,perhapsjokingly,suggested this presentation of fridge items would impact her future purchases as she may be shamed into purchasing healthier groceries. The menu interface was streamlined to the most basic tasks identified and was popular across all users interviewed. It was considered to be intuitive and functional. During the evaluation, it was found that users perceived the notifications of low stock to be useful butwere vehemently against Price Value suggesting items they may wish to buy. User 1, User 3 and User 5 each consideredthistobe tantamounttoforcedadvertisingandfounditdistasteful.Perhaps the experience would be improved if it was made clearer that these suggestions were based upon shopper behaviour and analytics of previous purchases. In the article entitled ‘A Fresh Look at Online Grocery Shopping’, Warschun, and colleagues (2012) state that a significantbarriertopurchasinggroceriesonline isthat people are generally already quite contentwiththe waythat theyobtaintheirgroceries.The evaluations found that there remains a reticence by shoppers to relinquish the eye-ball test of fresh products. With this considered, the ShopperTrustfeature attempted to instil trust in the shopper that the quality checks they themselves would perform would be taken care of by a trusted shopper. This feature was popular amongst all the usersinthe evaluation.Theyeachfoundthe ShopperTrustandYour Shopperscreenstobe intuitive and to the point. Priorto the evaluation,itwashighlightedinthe requirementsgatheringphase thatpeople were also very dissatisfied with the shop substitution systems. In the fridge interface, users were notified whenitemswere unavailable and suggestions for replacement were offered. User 1, User 3 and User 5 each foundthisaspectof the interface tobe misleadingandall userswere uneasyabout suggestions by the store. It was common amongst each of those who identified this as a weakness that they would prefer to manually select substitutions or receive no item at all in its place.
- 15. 14 Althoughthe userinterface was generally perceived to be intuitive and navigable, there were some design flaws identified by the users during the evaluation phase that could easily be improved upon. For example, the Best Before feature was unpopular with User 1 and User 2 and a number of icons were determined to be misleading or vague. User 3 identified the lack of pricing information on the shoppingandproductdescriptionpagestobe a fundamental flaw.Therecanbe no denyingthatthis should be rectified in the next redesign. User 5 considered there to be insufficient information contained in the Shopper Trust screen, although they found it easy to guess the details. Galante and colleagues(2013) alsosuggestthatthe reason online shopping has not taken off is it hasadditional costsassociatedwithlabour,deliveries,vehicles and their fuel and maintenance costs, with customers reluctant to incur extra expense for the additional convenience. This opinion was validatedbythe evaluation.The DeliveryandDeliveryTrackinginterface featureswere consideredto be adequate butthe notionof payingadditional chargesfordelivery,havingpurchasedorsubscribedtouse the smart fridge was vexing to users. User 2 stated they considered free delivery should be one of the contractual terms of the smart fridge and service. The requirementsgatheringphase identified another bugbear that seemed to regularly appear inuser responses.Thiswasthe costof deliveryandthe durationof time that it took for the groceries to arrive,particularlywithinpopulartime slots.Peoplegenerallywanttheirfoodto arrive on the same day that theyorderit.The priority delivery feature of the interface was considered by User 1 and User 3 to be unclearand unpopular.Also,the feesexhibitedwithinthe prototype,while purelyfordemonstration purposes,were exorbitantandmayhave impactednegativelyon the user evaluation. Perhaps the next iteration, should priority delivery remain a feature, could look at including realistic charges for such a service. The inability to avail of the same offers online as in-store was another feature that was identifiedasimportanttousersduringthe requirementsgatheringphase.Customerswishedtofeel like theywere notmissingoutondealsand offers by avoiding the supermarket. Thus, the Rewards feature and associated screens were created. User 1, User 3 and User 5 each suggested that rewards were a positive move, common in all grocery retailers but felt the implementation of this in the Price Value fridge application was handled incorrectly. The points system, again for demonstration purposes, was considered to be wildly high. Those same users each remarked that they would prefer money-back
- 16. 15 discounts rather than store-suggested loyalty items. Rewards and offers are highly regarded by consumers so this should be prioritised in the next iteration of the interface. Rogers and colleagues (2015) identify anumber of factors for interpreting evaluation data. The reliability of the evaluation seems to have been quite high, in that many of the results were echoed across eachof those usersthat evaluatedthe hi-fi prototype.Despitethis,itisasmall sample size and it wouldbe pertinenttogrowthissample size should the product be developed further. The validity can also be considered to be high, as the participants evaluated each of the features of the application as well asthe interface.Thiscouldof course be increasedwithahigherdefinitionprototype, i.e. a working smart fridge,foruserstoevaluate.Sucha prototype wouldseemlikeapositive nextstep,allowingusers to evaluate the prototype in a in a natural setting or “in the wild” (Rogers et al., 2015). There were a number of unintended negative biases introduced, namely around the use of arbitrary figures for prioritydeliveryandsuggestedrewards.Inthe nextiterationof the productthese will be addressed.The scope of the findings is quite narrow and can be used to implement positive changes to the product quickly. Conclusion The team were able to evolve an idea from an initial abstract concept to a concrete and extensive realisation of both design and functionality. The final hi-fi, while basic in its interactivity allowed for a material representation of the concept that was presented to real potential users and evaluated.There wasadistinctshiftinperceptionandemotionalvalence towards the product from the lo-fi prototype to the hi-fi, which could suggest that the lo-fi evaluation methods had too positive a slant. While the glass door was immensely popular initially, it appears that after seeing it in action, participants were made acutely aware of the aesthetic needs of an organised fridge, which they no longer felt was appealing. Participants even suggested that the glass door be removed altogether because of thisor made frostedinstead.Due tothe overlaytechnologyof the design,thiscouldbe easily integrated into the design, were a slider of sorts could change the opacity of the screen to the user’s liking. Highlighting this feature in future iterations could aptly cover these concerns. The participantswere generallynegative aboutcertainfeaturesof the fridge, particularly about placeholder item suggestions and pricing. This could be easily fixed in a future iteration where prices
- 17. 16 wouldbe includedasaneutral sign,whichwouldallow customerstonote thata price wasavailable,but not invoke the emotional valence attached to judging their perceived value. Users also felt that suggestionswere aformof intrusive advertising,perhapsbecause the itemspickedwere not items they wouldusually shop for themselves. A small user survey of regularly purchased items could be used to tailorthe evaluation experience to the customer and highlight the unique features of the product in a mannermore similar to what they would experience if they owned the product themselves. Finally, a greater effort should be made to highlight why certain suggestions were chosen, as these can act to increase usertrustinthe algorithm,asopposedtoleadingtosuspicionof ulteriormotives.Lastly,due to the small sample size,the reactioncouldhave been just a coincidental distribution of a segment of the population adverse to the idea, and a larger sample can be used in the future to combat this. Overall, the smart refrigerator has great potential as a product and would fit well into the current trend towards home automation and the internet of things (Kortuem et al., 2010). With modifications and further innovative thinking, the product could serve to be the boost needed to increase growth in the online shopping market and increased adoption. Even without considering the productitself,keyideas highlighted in the design could be incorporated into existing online shopping experiences to increase their sales and revenue (e.g. delivery tracker and shopper trust). This project allowed the team to be immersed in the design experience and discover the limitationsof the process,butalsothe potential forinnovation.Itisclear from our experience however that the perspectives of the users and the designers do not always align, and it is necessary to include frequent and extensive user input to assure a quality experience centred around the user, who will ultimately be the one benefitting from the experience.
- 18. 17 References BASKINGER, M.(2008). PENCILS BEFORE PIXELS:A PRIMER IN HAND-GENERATED SKETCHING.INTERACTIONS,15(2), 28-36. BEYER, H., &HOLTZBLATT,K.(1997). CONTEXTUAL DESIGN:DEFINING CUSTOMER-CENTEREDSYSTEMS. ELSEVIER. BRETON,E.R., FUEMMELER,B.F., &ABROMS,L. C.(2011). WEIGHT LOSS—THERE IS AN APPFOR THAT!BUT DOES IT ADHERE TO EVIDENCE-INFORMEDPRACTICES?TRANSLATIONAL BEHAVIORAL MEDICINE, 1(4), 523-529. CARROLL,J.M.(2004). BEYOND FUN. INTERACTIONS, 11(5), 38-40. CHAVAN,A. L., GORNEY,D., PRABHU,B., &ARORA,S.(2009). THE WASHINGMACHINE THATATE MY SARI - MISTAKES IN CROSS-CULTURAL DESIGN,INTERACTIONS,16(1), 26-31. CONSTANTINE,L.L., &LOCKWOOD,L. A.(1999). SOFTWARE FOR USE:A PRACTICAL GUIDE TO THE MODELS AND METHODS OF USAGE-CENTEREDDESIGN.PEARSON EDUCATION. DHILLON,B., BANACH, P., KOCIELNIK, R., EMPARANZA,J.P.,POLITIS,I., PĄCZEWSKA,A.,&MARKOPOULOS, P. (2011). VISUAL FIDELITY OF VIDEO PROTOTYPESAND USERFEEDBACK:A CASE STUDY.IN PROCEEDINGS OF THE 25TH BCSCONFERENCE ON HUMAN-COMPUTER INTERACTION (PP.139-144). BRITISH COMPUTER SOCIETY. DIX,A. (2004) HUMAN-COMPUTER INTERACTION.UNITEDKINGDOM:PEARSONEDUCATION. ERICKSON,T.D.(2000). WORKING WITH INTERFACE METAPHORS.IN B.LAUREL,AND MOUNTFORD,J. (1990) THE ART OF HUMAN-COMPUTER INTERFACE DESIGN.ADDISON-WESLEY, BOSTON. FUNKE,F., REIPS, U.D., &THOMAS,R. K.(2011). SLIDERS FOR THE SMART:TYPE OF RATING SCALE ON THE WEB INTERACTS WITHEDUCATIONAL LEVEL.SOCIAL SCIENCE COMPUTER REVIEW,29(2), 221-231. GALANTE,N., LÓPEZ,E.& MUNBY,S.(2013). THE FUTURE OF ONLINE GROCERY IN EUROPE. RETRIEVED FROM HTTP://WWW.MCKINSEY.COM/INSIGHTS/CONSUMER_AND_RETAIL/THE_FUTURE_OF_ONLINE_GROCER Y_IN_EUROPE GREENBERG,S., CARPENDALE,S., MARQUARDT,N.,&BUXTON,B. (2012). SKETCHING USER EXPERIENCES. MORGAN KAUFMAN.
- 19. 18 HOLDSWORTH,C.(2009). 'GOING AWAY TO UNI':MOBILITY,MODERNITY,AND INDEPENDENCE OF ENGLISH HIGHER EDUCATION STUDENTS.ENVIRONMENT ANDPLANNING.A, 41(8), 1849. JACOBSON,I., CHRISTERSON,M., JONSSON,P., AND OVERGAARD,G.(1992) OBJECT-ORIENTEDSOFTWARE ENGINEERING –A USE CASE DRIVEN APPROACH. ADDISON-WESLEY,HARLOW, ESSEX. KORTUEM,G., KAWSAR,F., FITTON,D., &SUNDRAMOORTHY,V. (2010). SMART OBJECTS AS BUILDINGBLOCKS FOR THE INTERNET OF THINGS.INTERNET COMPUTING,14(1), 44-51. LUCE, K. H., WINZELBERG,A.J., DAS, S., OSBORNE,M.I.,BRYSON,S.W., &TAYLOR, C.B.(2007). RELIABILITY OF SELF-REPORT:PAPER VERSUS ONLINE ADMINISTRATION.COMPUTERSIN HUMAN BEHAVIOR,23(3), 1384-1389. RAYPORT,J. F.(1999). THE TRUTH ABOUT INTERNET BUSINESSMODELS.STRATEGY AND BUSINESS, 5-7. ROGERS,Y., SHARP,H., PREECE, J.(2015). INTERACTION DESIGN:BEYONDHUMAN COMPUTER INTERACTION. UNITED KINGDOM:JOHN WILEY &SONS. ROUILLARD, J.(2012). THE PERVASIVE FRIDGE.A SMART COMPUTER SYSTEM AGAINST UNEATEN FOODLOSS. IN SEVENTH INTERNATIONAL CONFERENCE ON SYSTEMS (PP-135). SAUER,J., &SONDEREGGER,A.(2009). THE INFLUENCE OF PROTOTYPE FIDELITY AND AESTHETICS OF DESIGN IN USABILITYTESTS:EFFECTSON USER BEHAVIOUR,SUBJECTIVE EVALUATION AND EMOTION.APPLIED ERGONOMICS,40(4), 670-677. SCHÖN,D.A.(1983). THE REFLECTIVE PRACTITIONER:HOW PROFESSIONALSTHINK IN ACTION.BASICBOOKS, NEW YORK. SMYTH,J.D., DILLMAN,D.A., CHRISTIAN,L. M., &STERN, M.J. (2006). COMPARING CHECK-ALL AND FORCED- CHOICE QUESTION FORMATS INWEB SURVEYS.PUBLICOPINION QUARTERLY,70(1), 66-77. WARSCHUN,M., ALBERS,S., DELFMANN,W., &MUBIG,R.(2012). A FRESHLOOK AT ONLINE GROCERY. RETRIEVED FROM HTTPS://WWW.ATKEARNEY.DE/DOCUMENTS/856314/1214618/BIP_A_FRESH_LOOK_AT_ONLINE_ GROCERY.PDF/2F4E07B2-3859-4D79-9049-33A2F25A8B0D XIAO,B., &BENBASAT,I.(2007). E-COMMERCE PRODUCT RECOMMENDATION AGENTS:USE,CHARACTERISTICS, AND IMPACT.MIS QUARTERLY,31(1), 137-209.
- 20. 19 ZYCH,L. (2015) GOOGLE VS.AMAZON:PUTTING SAME-DAY DELIVERY TO THE TEST.RETRIEVED FROM HTTPS://SPEEDYBOXCOURIERS.COM/GOOGLE-VS-AMAZON-PUTTING-SAME-DAY-DELIVERY-TO-THE-TEST/