Configuring Configurators : The creation of UX guidelines for parametric product customization

(1)

Linköping University | Department of Management and Engineering Master’s Thesis, 30 Credits | Design Engineering Spring 2021 | ISRN: LIU-IEI-TEK-A–21/04064—SE

Configuring Configurators

- The creation of UX guidelines for parametric

product customization

Saga Gotthold, Anna Karlsson

Supervisor: Maria Gustin Bergström Examiner: Renee Wever

Linköping University SE-581 83 Linköping, Sweden +46 013 28 10 00, www.liu.se

(2)

Copyright

The publishers will keep this document online on the Internet – or its possible replacement – for a period of 25 years starting from the date of publication barring exceptional circumstances. The online availability of the document implies permanent permission for anyone to read, to download, or to print out single copies for his/hers own use and to use it unchanged for non-commercial research and educational purpose. Subsequent transfers of copyright cannot revoke this permission. All other uses of the document are conditional upon the consent of the copyright owner. The publisher has taken technical and administrative measures to assure authenticity, security and accessibility. According to intellectual property law the author has the right to be mentioned when his/her work is accessed as described above and to be protected against infringement. For additional information about the Linköping University Electronic Press and its procedures for publication and for assurance of document integrity, please refer to its www home page: https://ep.liu.se/.

(3)

Abstract

The aim of the thesis is to identify which functions may be included in product configurators in order to create a good experience for a user. The findings are compiled into a set of guidelines outlining how the user experience is affected by functions, which functions are either necessary- or supporting functions, and how these functions may be realized to create a good user experience. The thesis is conducted by firstly studying literature about both UX-theory and product configurators. Subsequently, a state of the art study is performed by investigating how ten different product configurators on the market fulfill specific functions. A benchmarking study is performed with users. This study is conducted in two different instances, one with an existing product configurator, and one with a hi-fi prototype of the same configurator updated according to the developed guidelines. In between these two instances an iterative user study is carried out to co-design configurators with users.

The established guidelines are the following: The configurator should be accommodated to different levels of creative confidence for customizing a product, The configurator should let the user express themselves when customizing a product, The configurator should allow the user to fulfil their goal of customizing a product, The configurator should show the user how to fulfil their goal of customizing a product.

(4)

Acknowledgements

Thank you to all who helped us in this project. We want to start by thanking our supervisor Maria Gustin Bergström for listening and providing good advise throughout the entire thesis process, and to our examiner Renee Wever for insightful feedback.

Thanks to Kristoffer Skyttner at SkyMaker for initiating this interesting research topic, and to Alexander Nilsson for creating our hi-fi prototype and providing inspiration to our project. We want to thank our opponents Isabella Marklund and Simon Karlsson for a great collaboration, insightful feedback and an interesting report to read in return. Thanks to Louise Josefsson for creating a great report template which has been very helpful to us. Lastly, we want to send thanks to our test participants for devoting their time to participate in our user studies: Adam, Daniel, Daniel, Emma, Jawed, Louise, Lovisa, Lovisa, Moa, Nathalie, Niklas, Nina, Per.

(5)

List of Tables

3.1 How well each product configurator fulfills the set criteria. Dark means it is fulfilled, white is not fulfilled, and half coloured means it is somewhat fulfilled. 29 3.2 Data about the participants in the benchmarking study. Tests with ID type

B1 are performed in the initial benchmarking study, while test with ID type B2 are performed in the follow-up benchmarking study. . . 43 3.3 Data about the participants in the iterative user study. . . 47 4.1 A summary of how the investigated product configurators realize features

re-lated to enabling product customization. . . 54 4.2 A summary of how the investigated product configurators realize features

re-lated to purchasing the product. . . 55 4.3 A summary of how the investigated product configurators realize features

re-lated to the product visualization. . . 56 4.4 A summary of how the investigated product configurators realize features

re-lated to how they guide the user. . . 59 4.5 A summary of how the investigated product configurators allow the users to

perform trial and error in the configurators. . . 61 4.6 A summary of each individual test in the iterative study. The choices made by

(10)

List of Figures

1.1 An example of a product configurator created by SkyMaker AB (2021), created in DynaMaker. Screenshot by author. . . 1 1.2 An example of a product configurator for a robot gripper. The configurator is

created in DynaMaker (SkyMaker AB, 2021). Adapted screenshot by author. 2 1.3 Illustration of how the different research questions help create the guidelines. 5 1.4 Illustration of the work process of this thesis project. . . 7 2.1 Hassenzahl’s model of the user experience. The faded boxes cannot be altered

in their respective perspectives. Adapted from Hassenzahl (2003). . . 10 2.2 The Components of the User Experience (CUE) model created by Thüring and

Mahlke (2007). Adapted from Minge et al. (2017). . . 13 2.3 An example of a check button. Screenshot by author from Microsoft (2019) . 14 2.4 Example of a list box. Screenshot by author from Microsoft (2019). . . 14 2.5 An example of radio boxes. Screenshot by author from Microsoft (2019). . . . 14 2.6 Example of check boxes. Screenshot by author from Microsoft (2019). . . 14 2.7 An example of a text box. Screenshot by author from Microsoft (2019). . . . 14 2.8 An example of a slider. Screenshot by author from Microsoft (2019). . . 14 2.9 An example of a dialogue box. Adapted screenshot by author from Microsoft

(2019). . . 15 2.10 An example of a balloon help. Adapted screenshot by author from Microsoft

(2019). . . 15 2.11 An example of a tool-tip. Screenshot by author from Microsoft (2019). . . 15 2.12 A general example of how the framework for a customer journey may look. . . 16 2.13 A general example of an established HTA. . . 17 2.14 An example of how a perfume configurator explains the function of a perfume

(i.e. the smell) by using images. Screenshot by author from the website Fra-grance by Me (2021). . . 19 2.15 An example of how a couch configurator provides modules. Providing modules

like this gives the user a semi finished part consisting of seat, legs, backrest, and arm rests. Hence, they do not have to design these independently. Screenshot by author from Ikea (2021a) . . . 20

(11)

2.16 The modules of the meCUE questionnaire, adapted from Minge and Thüring (2018b) . . . 24 3.1 A collage with overviews of the investigated product configurators. 1) Shelf

Help, (Shelf Help, 2020) (2) DynaMaker (DynaMaker, 2021), (3) Anylamp (Anylamp, 2017), (4) Pickawood (Pickawood, 2018), (5) Mellby Home (Mellby Home, 2020), (6) Elfa (Elfa, 2017), (7) Ikea Pax (Ikea, 2021b), (8) Modulor (Modulor, 2021), and (9) Boston Tec (Boston Tec, 2019). Screenshots by authors. 30 3.2 The function tree established for product configurators. . . 32 3.3 The (Shelf Help, 2020) configurator. Screenshot by author. . . 33 3.4 The HTA of using Shelf Help (2020), focusing on the 1.1 Familiarize/1. Modify

sub-goal of the analysis. . . 34 3.5 The HTA of using Shelf Help (2020), focusing on the 1.2 Specify/1. Modify

sub-goal of the analysis. . . 35 3.6 The HTA of using Shelf Help (2020), focusing on the 2. Confirm sub-goal of

the analysis. . . 35 3.7 The HTA of using Shelf Help (2020), focusing on the 3. Go to Cart sub-goal

of the analysis. . . 36 3.8 The start page of the hi-fi prototype of the Shelf Help configurator. . . 38 3.9 The HTA of using the hi-fi prototype of Shelf Help, focusing on the

Familiar-ize/Modify sub-goal of the analysis. . . 39 3.10 The HTA of using the hi-fi prototype of Shelf Help, focusing on the

Spec-ify/Modify sub-goal of the analysis. . . 40 3.11 The HTA of using the hi-fi prototype of Shelf Help, focusing on the Confirm

sub-goal of the analysis. . . 41 3.12 The HTA of using the hi-fi prototype of Shelf Help, focusing on the Go to cart

sub-goal of the analysis. 3.2 is grayed out since it cannot be fully integrated into the prototype, but the feature should be included when integrated in a web shop. . . 42 3.13 The template for the customer journey maps used for analysing the

bench-marking study, based on the work process identified in Shelf Help (2020). The graph at the top is an example of the participant’s emotional journey. Adapted from a template available in Miro (2021). . . 46 3.14 An overview of the testing page which the participants interact with in its initial

state, before any changes are made between the individual tests and iterations. The three templates are shown to the left, and the UI-representations of the functions are shown to the right. More legible figures are shown below. . . 49

(12)

3.15 Template 1 from the iterative study in its initial state, before any changes are made between individual tests and iterations. . . 50 3.16 Template 2 from the iterative study in its initial state, before any changes are

made between individual tests and iterations. . . 50 3.17 Template 3 from the iterative study in its initial state, before any changes are

made between individual tests and iterations. . . 51 3.18 The UI-representations of the functions the participants can choose from in

the iterative study in their initial state, before any changes are made between individual tests and iterations. . . 52 3.19 An overview of the process for one iteration in the iterative user study. . . 53 4.1 How the 3D-view looks in the Mellby Home (2020) configurator. Screenshot

by author. . . 56 4.2 How the first person view looks in the Mellby Home (2020) configurator.

Screenshot by author. . . 56 4.3 How the measurements are shown in a 3D-view along the model in Shelf Help

(2020). Screenshot by author. . . 57 4.4 How the measurements are shown in a 3D-view along the model in Ikea (2021b).

Screenshot by author. . . 57 4.5 How the room environment looks in Pickawood (2018). Screenshot by author. 58 4.6 How the room environment looks in Schmidt (2020). Screenshot by author. . 58 4.7 How the product visualization is obstructed by tools in Shelf Help (2020).

Screenshot by author. . . 58 4.8 How the product visualization is obstructed by tools in Elfa (2017). Screenshot

by author. . . 58 4.9 One available environment in Mellby Home (2020). This environment resembles

an archipelago. Screenshot by author. . . 60 4.10 One available environment in Mellby Home (2020). This environment resembles

a suburban hill. Screenshot by author. . . 60 4.11 How the decorative elements in Ikea (2021b) look. Screenshot by author. . . . 61 4.12 How the decorative elements in Schmidt (2020) look. Screenshot by author. . 61 4.13 A compilation of the participants’ emotional journeys during the initial

bench-marking test. For the individual customer journeys, see Appendix D. . . 63 4.14 The category View possibilities contains Templates, Add Modules, Reviews, and

(13)

4.15 The category Help contains Video Tutorial, Error Balloon, Tool-Tip, Error Balloon Help, Balloon Help, and Interactive Tutorial at the start of the iterative

user study. . . 66

4.16 The category Rotation contains Preset Views, Box Rotation, Sphere Rotation, and Floor Rotation at the start of the iterative user study. . . 67

4.17 The category Explanation of Product contains Environment Catalogue, Re-views, Descriptive Text, and Module Overview at the start of the iterative user study. . . 67

4.18 The category Interaction contains Yellow Area, Arrows, Blue Area, and +/- at the start of the iterative user study. . . 68

4.19 The iterative prototype adapted by A7. . . 70

4.25 The final version of Template 1 in the iterative prototype. . . 79

4.28 The final version of the alternatives in the iterative prototype. . . 81

4.29 A compilation of the participants’ emotional journeys during the follow-up benchmarking test. For the individual customer journeys, see Appendix E . . 83

4.30 The meCUE results for the first module, showing the mean score for each of benchmarking studies. . . 86

4.31 The meCUE results for the second module, showing the mean score for each of benchmarking studies. . . 86

4.32 The meCUE results for the third module, showing the mean score for each of benchmarking studies. . . 87

4.33 The meCUE results for the fourth module, showing the mean score for each of benchmarking studies. . . 87

4.34 The meCUE results for the fifth module. The lighter version of the bar is the result after the initial benchmarking study, and the darker bar is the result after the follow-up benchmarking study. . . 87

(14)

5.1 The function tree established for product configurators. Revised to show which sub-sub-functions are necessary- and supporting functions, based on the state of the art study. The faded boxes are functions which are not investigated. . . 93 6.1 An example of how to provide templates. The available templates vary in their

purpose and look, according to the functionality of the product configurator. . 117 6.2 An example of how to allow comparing designs. Two similar designs are

com-pared based on their image, price, weight, and plywood length. . . 117 6.3 An example of how to provide a sharing function. In this example the user can

both see other users’ designs and share their own design on the web-shop, and also share it to their personal social media accounts. . . 119 6.4 Example of how to let the user place their product in an environment. The

user can edit the size of the room and the colour of the walls and floor through a RGBA colour picker. . . 119 6.5 An example of how to show measurements in the 3D-model. . . 120 6.6 Example of how to show data about product. In this case it is a product

summary showing price, weight, and plywood length of a product. . . 120 6.7 An example of clear interaction points. The blue area is the active interaction

point, which appears when the user hovers over the shelf. . . 122 6.8 Example of a process structure. The five tabs provide a clear structure the user

(15)

Nomenclature

Attribute: Benchmarking study: Characteristic: GUI: Hi-fi prototype: HTA: Iterative: Mass Customization: meCUE: Parametric: Product configurator: TAP:

From Hassenzahl’s model (Hassenzahl, 2003). Attributes are either hedonic or pragmatic.

A study performed to evaluate one product compared to an-other product.

From Hassenzahl’s model (Hassenzahl, 2003). The character-istics in the model are: Stimulation, identification, evocation, utility and usability.

Graphical User Interface. It allows the user to interact with the interface of a product by including graphical elements. It includes windows, cursors, and buttons and other graphical objects.

High fidelity prototype. A prototype which is as close to the end product as possible in terms of GUI and functionality. Hierarchical Task Analysis. A tool used by user experience researchers to understand the users’ goals, and the tasks re-quired to reach the goals.

Repeating the same process several times.

A manufacturing technique which offers customized products at a mass produced scale.

Modular Evaluation of key Components of User Experience. A questionnaire created by Minge et al. (2017) to evaluate the user experience of a product.

A model which is constrained by set parameters. Other ge-ometries in the model are dependent on the parameters, and thus automatically changed when a parameter is altered. A tool provided by manufacturers to be used by consumers. In a configurator consumers can design their own version of a product offered by the manufacturer.

Thinking-Aloud Protocol. A method used in user studies. Users are encouraged to verbalize their thoughts when par-ticipating in a test.

(16)

UI:

UX:

User Interface. It allows the user to interact with the inter-face of a product.

User Experience. The theory about all experiences the user of a product experiences in relation to that product. In this thesis report, the term "UX" refers to the theoretical topic, while the term "User Experience" refers to the actual expe-rience of users.

(17)

1 Introduction

Product configurators allow consumers to become co-creators of their own products. Con-sumers want to be more creative and play a larger role in creating their products, instead of just consuming them as many currently do (Sanders, 2005). Product configurators let customers design their own products independently in a web shop, which in turn allows busi-nesses to have mass customization of their products. Many different types of product could be implemented in a product configurator. Anything from shoes, houses, and trucks. As product configurators become more advanced they may also become harder for users to understand and use. For this reason it is important to look at how the user experience can be improved in such software. In this introduction the project which this report concerns is defined and motivated. Firstly, a background of the project, product configurators, and UX is given. Then the research aim, goals and deliverables, and delimitations are explained.

1.1 Thesis Background

This thesis is conducted in collaboration with the company SkyMaker AB. SkyMaker is a company based in Linköping which has developed the CAD-configurator platform DynaMaker (SkyMaker AB, 2021). A configurator for a greenhouse which is created in DynaMaker can be seen in Figure 1.1. The company is interested in investigating best practices for creating a user experience where the customer can understand that they are using a product configurator even if they do not have any previous experience in CAD or similar product design programs.

Figure 1.1: An example of a product configurator created by SkyMaker AB (2021), created in DynaMaker. Screenshot by author.

Product configurators are tools used by businesses to offer customized products to their cus-tomers. They are often integrated on websites where the customers themselves can interact with the product and modify measurements, colours, materials or other parameters. One example is seen in Figure 1.2, which shows a product configurator for a robot gripper. The

(18)

customers can see how their customized product will look. This is key in achieving mass customization, which can be described as a combination of mass production and customiza-tion (Salvador et al., 2009). Several customers can create customized products and when the configurations are finished the software sends an estimated price offer to the customer and/or sends the order for production to the manufacturer.

Figure 1.2: An example of a product configurator for a robot gripper. The configurator is created in DynaMaker (SkyMaker AB, 2021). Adapted screenshot by author.

There is a risk of users becoming surprised by more advanced product configurators, which might lead to them abstaining from using them (Salvador et al., 2009). According to the owner of this project, SkyMaker AB, the customers might misinterpret the functionality of the product configurator. To counteract this, the UX (User Experience) in the software can be adapted to make it easy to use and clarify the functionality of the product configurator. UX does not only include clarifications of the functionality of a product configurator. The subject concerns everything a user experiences in relation to a product or service (Christensen et al., 2020). This can entail both physical and emotional experiences, and according to Hassenzahl (2003) it includes aspects a designer can control, and aspects they cannot control. Due to the holistic nature of UX, there are many different ways of describing it, and working with it (Pucillo and Cascini, 2014).

1.1.1 User Experience Background

To develop best practices for creating a good user experience, it is important to understand what UX is comprised of. The term “User Experience” first appeared during the 1990s in the article What you see, some of what’s in the future, and how we go about doing it: HI at Apple Computer by Norman et al. (1995). However, a singular definition of UX and its components does not currently exist despite scholars having attempted to define the term ever since its

(19)

creation (Christensen et al., 2020). Hence, there exist numerous models trying explain and encompass UX (Pucillo and Cascini, 2014). However, as Christensen et al. (2020) gathered, despite the unclarity, several academics seem to agree on the holistic nature of UX. It is not simply a singular event of an interaction, but everything in relation to the product; from the product’s services to the company’s brand which forms the user’s perception and thus, the user experience. Pucillo and Cascini (2014) also points out that the user experience is heavily dependent on the context or situation of the user, and hence, it is difficult to design an experience. For example, a designer cannot affect what happens to the user earlier during their day, and therefore the designer cannot know which mood the user will be in when interacting with the product.

Due to the holistic nature of UX, its components are not set in stone. By conducting an empirical study on UX, Hassenzahl and Tractinsky (2006) define different research perspec-tives within UX. However, they conclude that none of the perspecperspec-tives are capable of truly encapsulating all that is UX, but do contribute to aspects of it. In the end, Hassenzahl and Tractinsky (2006) argue that the core of UX is to create products which actively provide pleasure, joy, and not just simply being the "absence of pain" (s. 95).

As evident, UX is an elusive and holistic term. Therefore, it is important to determine which features of UX are under consideration when applying it to specific products, such as product configurators.

1.1.2 Product Configurator Background

In order to explore the area of interest for this thesis an understanding of product configurators needs to be developed. A brief background to answer what types of configurators exist, what product configurators are used for, and why they are used, is introduced in this section. Hermans (2012) divides configurators into four different types: Veneers where only the surface of a product is adapted, modular in which the user chooses between pre-determined parts, parametricwhere certain parameters are changed by the customer, and generative customiza-tion which generates a design from simple sketches. All four types of product configurators may be used in order for companies to let customers customize their products.

Product configurators became an important tool for businesses in the 1990’s (Stevens and Jouny-Rivier, 2020). They allow users to co-design their products since a configurator can be placed on a website and thus be accessed at any time by any customer, and it provides constraints and aid for configuring the product (Franke and Piller, 2002; Von Hippel, 2001). The reason being that a product configurator is a means for the user to identify and express their own needs, which provides more accurate data than interviews and more inexpensive than ethnographic studies (Von Hippel, 2001). A product configurator is especially suitable when a company needs to find the needs of many different users or the changing needs of a customer (instead of, for example, performing new user studies each time) (Von Hippel, 2001). Implementing product configurators is one way of achieving mass customization for businesses.

(20)

Product configurators are one of the main enablers of mass customization (Blecker et al., 2004), and mass customization has been on the rise since the 1980’s (Pine, 1993). Pine (1993) predicted that it would be necessary for companies to survive in an increasingly turbulent market due to globalization, quicker changes in technology, and ever evolving customer de-mands.

In conclusion, there are four main types of configurators and the main difference between them is how the product is altered, through veneers, through modules, parametrically, or genera-tively. Product configurators are utilized by businesses to achieve mass customization of their products, and they help businesses better understand their customers through performing co-design.

1.2 Research Aim and Questions

The aim of the thesis is to identify which functions could be implemented to create a good user experience in product configurators with regards to inexperienced users. The functions identified in the project concern elements the private individual user interacts with in the configurator. Hence, no back-end functions or functions utilized by a manufacturer are in-vestigated. Furthermore, functions are not considered to be purely aesthetic GUI-elements (Graphical User Interface), and also not raw written code. For example, a function considered in this thesis could be a slider, but it is not the code behind the slider nor the appearance of it. The project is expected to culminate in a set of guidelines for product configurator developers which will help them design user friendly products. The aim is fulfilled by answering the following research questions:

• RQ0: Which generally applicable guidelines can be developed from the answers to the following sub-research questions?

– RQ1: How do different functions affect the user experience in a product configu-rator for private individuals?

– RQ2: Which functions are either necessary- or supporting functions for the user to reach their goals in a product configurator for private individuals?

– RQ3: How might the necessary- and supporting functions be realized to create a good user experience in a product configurator for private individuals?

The necessary functions mentioned in RQ2 and RQ3 refers to functions required for a software to be a configurator. Supporting functions, on the other hand, are those which are not required but appreciated by users. The sub-research questions each answers a different aspects of the guidelines (RQ0). RQ1 gives a main, overarching guideline while RQ2 and RQ3 answers how the this guideline can be realized through different functions. These relationships are illustrated in Figure 1.3.

(21)

Figure 1.3: Illustration of how the different research questions help create the guidelines.

1.3 Goals and Deliverables

The goal of the thesis is to deliver a set of guidelines regarding what functions are needed in a configurator in order for it to provide a good user experience to the user. A good user experience in this case refers to the users ability to reach their goal in a positive manner. The guidelines are intended to be used by product configurator developers while designing configurators.

1.4 Delimitations

The delimitations of this project regard the investigated product configurators, configurator users, guideline users, and technical delimitations. These delimitations are explained further in the following sections.

1.4.1 Product Configurator Delimitations

In reality, product configurators may have different requirements for creating a good user experience. Such differences may depend on the brand image, manufacturing capabilities, and the product type. However, in this project it is assumed that the developed guidelines are applicable to all product configurators which fulfill the selected criteria. In order for the guidelines to be considered applicable to a specific configurator, the configurator should fulfill the following four criterion. The criteria are as follows:

(22)

individuals are the target group of the thesis, only product configurators geared towards this group are investigated. Another reason for this distinction is that professional individuals might have access to courses and training in using the product configurators which is not available to the authors.

• The product configurators allow the user to configure a piece of furniture. The reason is to keep the functionalities of the product configurators as similar as possible.

• The product configurators must contain a 3D-model which is configured by the user. This is important since the potential problems in 3D and 2D are different (H. Zhao et al., 2019).

• The product configurators deal with parametric design of the product. Veneers, modu-lar, and generative configurators have different functionalities compared to parametric configurators (Hermans, 2012) and are avoided to make relevant comparisons.

The project focuses on the user experience from the moment the user first sees the product configurators to the moment the user adds the product to their online cart. Hence, the external design of the websites is not taken into consideration. The user experience is further limited to activities only related to the product configurators and no external services the web shop may offer, such as delivering the product to the home of the user.

1.4.2 Product Configurator User Delimitations

The intended users for the researched product configurators are adults with average skills in using computers. The intended users are without professional experience from CAD-software or product configurators. They are assumed to have the goal of owning a customized product, and this should be the reason as to why they are utilizing a product configurator. Hence, the product configurator should not be intended to be used as a game, and its purpose not only to provide entertainment for the user. Instead, the product configurators should be meant to enable ownership of customized products. The participants partaking in different instances of the thesis project match the requirements of the intended users of the product configurators.

1.4.3 Guideline User Delimitations

The product configurator developers are expected to have knowledge in programming of prod-uct configurators. Therefore, the guidelines do not showcase a step-by-step process of imple-mentation but rather some general guidelines for best practice.

1.4.4 Technical Delimitations

To further limit the scope of the research only interactions through computers are investigated. This includes interaction with a keyboard, computer mouse or touch pad, and a desktop or laptop monitor.

(23)

The project utilizes an existing product configurator created by SkyMaker, Shelf Help (2020). This is the only configurator used for validating the guidelines. Furthermore, the creation of a functioning prototype is outsourced to the project owner company. This is due to the limited programming experience of the authors. For this reason the code and creation of the hi-fi prototype is not detailed in the report.

1.5 Thesis Process

As showcased in Figure 1.4, the process starts with an initial planning of the project followed by collecting literature for the 2 Theoretical Frame of Reference. Then a State of the art study is carried out, and an Initial benchmarking study of an existing product configurator is performed. This is followed by an Iterative user study in a lo-fi prototype. Based on the information collected in the previous phases a first draft of the Guidelines are compiled. These are implemented in a hi-fi prototype, this prototype is then used in a Follow-up Benchmarking study.

Figure 1.4: Illustration of the work process of this thesis project.

The 2 Theoretical Frame of Reference is constructed by gathering literature regarding UX, product configurators, and user studies. The sections about UX and Product configurators are used as a basis for the later discussion regarding the results of the benchmarking study and the iterative study. Furthermore, the section 2.1 User Experience focuses on describing different aspects of the user experience while the section 2.2 Product Configurators focuses on theoretically established guidelines for which functions should be considered in a configurator. Hence, the UX and the Product configurators sections covers RQ1: How do different functions affect the user experience in a product configurator for private individuals? Moreover, the two sections also serve as the framework for the guidelines. Lastly, the user studies section lays down the framework for how the benchmarking study and the iterative user study should be carried out.

The state of the art analysis is conducted to gain an understanding of launched product configurators. Before the state of the art study is conducted, a function tree is created which includes the necessary sub- and support functions for product configurators. By noting which functions are included in the studied product configurators, the function tree is revised n terms of which functions are considered sub- or support functions based on the state of the art study i. Hence, the state of the art study covers RQ2: Which functions are either necessary- or supporting functions for the user to reach their goals in a product configurator for private

(24)

individuals? as well as RQ3: How might the necessary- and supporting functions be realized to create a good user experience in a product configurator for private individuals?.

The benchmarking study is modelled after the findings in the 2.3 User Studies and is carried out on two different occasions, as shown in the third and seventh steps in Figure 1.4. The initial benchmarking study is conducted with an existing product configurators, after the state of the art study is completed, while the follow-up benchmarking study occurs after a hi-fi (High Fidelity) prototype is created with the guidelines implemented. The two test rounds of the study are conducted to be able to validate the created guidelines. Importantly to allow a fair comparison, the two tests are planned, carried out, and analysed in the same manner. The first test serves the additional purpose of identifying current problems and benefits in the chosen configurator and acts as a starting point for the creation of the prototype used in the subsequent iterative study.

The iterative study is conducted after the initial benchmarking study is completed, as shown in Figure 1.4. The iterative study consists of six tests conducted in sequence, with each test building upon the previous one. The test includes an iterative prototype which is based on the 2.1 User Experience and 2.2 Product Configurators of the section 2 Theoretical Frame of Reference, as well as the results of the initial benchmarking study. The aim of the study is to quickly gain users’ opinions and reactions to design suggestions for potential representations of the identified necessary functions in the state of the art study. Hence, the iterative study is used as a basis for answering RQ3: How might the necessary- and supporting functions be realized to create a good user experience in a product configurator for private individuals? The results of the iterative study is thus the participants’ opinions of which functions should be included in a product configurator, all of which are used as the basis for developing the guidelines.

After the iterative study is finished and analysed the guidelines are created. The guidelines are based around the framework identified in the frame of reference. The necessary functions identified in the state of the art study. As well as the participants opinions and reactions of representations of the functions during the initial benchmarking study and the iterative study. The guidelines and the results from the meCUE questionnaire filled out by the participants in the subsequent testing serve as the answer to RQ0: Which generally applicable guidelines can be developed from the answer to the following research questions? The follow-up benchmarking study is performed on a hi-fi prototype which is a modified version of the product configurator used in the initial benchmarking study. The modifications done in the configurator are based on the first draft of the guidelines, and the aim of the study is to validate the implementation of the guidelines.

(25)

2 Theoretical Frame of Reference

In this chapter relevant literature is presented. The main topics are; user experience, product configurators, and user studies. All these topics are essential to understand and apply in order to compile guidelines for designing a good user experience in product configurators. To start, the user experience deals with the general concept of how users of a product are affected by it, and this is presented below.

2.1 User Experience

When it comes to user experience in product configurators several interesting topics to in-vestigate exist. In the following section the components of the user experience, the graphical user interface, and ways of understanding the user experience through creating empathy are described.

2.1.1 The Components of the User Experience

The user experience is affected by many different factors, as such several different perspec-tives exists on the subject (Hassenzahl and Tractinsky, 2006). Hassenzahl (2003) presents a model of how the user experience is created by both the designer and the user. The model proposed by Hassenzahl (2003) has been used by Hassenzahl in later works to describe other relationships within UX, such as the interplay of beauty and goodness (Hassenzahl, 2004). The elaborated version by Hassenzahl (2004) has been further researched and confirmed by several scholars (Karapanos, 2010). The original model consists of the relations between product features, product character and consequences (Hassenzahl, 2003), and is illustrated in Figure 2.1. The original model by Hassenzahl (2003) is split between the aspects of the user experience which the designer and the user respectively can control respectively. The components are further described below.

(26)

Figure 2.1: Hassenzahl’s model of the user experience. The faded boxes cannot be altered in their respective perspectives. Adapted from Hassenzahl (2003).

A product feature is an aspect of the product which the designer has control over (Hassen-zahl, 2003). Hassenzahl (2003) describes it as being the product’s content, presentation, functionality, and way of interaction. These are designed to have a specific product character (Hassenzahl, 2003).

A product character is how the product is interpreted and its assigned attributes (Hassenzahl, 2003). Hassenzahl (2003) uses "interesting", "useful" and "novel" as examples of potential attributes. Notably, Hassenzahl (2003) splits the product character into the intended and the apparent, which can be seen by the different perspectives in Figure 2.1. The split is due to the designer not being able to control how the user will interpret their intended product character (Hassenzahl, 2003). As such, while the user cannot affect the product features, the apparentproduct character is solely determined by the user’s own interpretation of the product features (Hassenzahl, 2003). Furthermore, Hassenzahl (2003) splits the product character into two attributes: pragmatic and hedonic. Diefenbach et al. (2014) note that the importance of each attribute varies depending on the product.

The pragmatic attribute is described by Hassenzahl (2003) as the Manipulation of the en-vironment, as seen in Figure 2.1. Hassenzahl (2003) describes the focus of the pragmatic attribute as to allow the user to fulfill an intended goal. He adds that in order to fulfill the goal, the user needs to be supplied with the necessary tools, as such the product needs to

(27)

consider Utility. Naturally, the user also needs to know how to access and use the tools, as such the product needs to consider Usability. Hassenzahl and Monk (2010) describe the pragmatic attributes as being the “what” (utility) and “how” (usability) of an interaction. The hedonic attribute differs from the pragmatic attribute (Hassenzahl, 2003). Instead of focusing on fulfilling the user’s intended goal, it considers how the user is feeling while they fulfill it (Hassenzahl, 2003). Hassenzahl and Monk (2010) describe the hedonic attribute as being the “why” to an interaction. The hedonic attribute of UX-design has been used by several scholars (Diefenbach et al., 2014). In a study on the use of the hedonic attribute in literature, Diefenbach et al. (2014) discovered that 65% of the studied literature referenced Hassenzahl’s model as the origin. To examplify the difference between the hedonic and the pragmatic, Hassenzahl (2003) uses the functionality of a hammer. Buying the hammer because it fulfils the need to drive in a nail would be basing it on its pragmatic attribute (Hassenzahl, 2003). Then if you choose a specific brand since you believe it reflects your values the choice of hammers is based in the hedonic attribute (Hassenzahl, 2003). Hassenzahl (2003) further breaks down the hedonic attribute into three components: Stimulation, Identification and Evocation. These three components are further described in the list below:

• Stimulation refers to the user’s need for personal improvement. Hassenzahl (2003) explains that even though certain functionalities are not currently utilized by the user, their existence makes the user feel like they will gain the competence to use them in the future. Moreover, a product’s novelty is also part of stimulation (Hassenzahl, 2003). • Identification refers to the user being able to express themselves through the product

(Hassenzahl, 2003). Hassenzahl (2003) describes this characteristic as being social, since it enables the user to show other people their personality .

• Evocation is the product’s ability to evoke memories within the user (Hassenzahl, 2003). Hassenzahl (2003) exemplifies evocation by old video games being enjoyable partly due to their nostalgia. Since evocation is created by context rather than the product itself, it is difficult to design intentionally (Roto and Rautava, 2008). Notably, Evocation is no longer considered to be included in Hassenzahl’s model (Bevan, 2008). Diefenbach and Hassenzahl (2011) argue that when users are given a choice between a prag-matic attribute or hedonic counterpart they will most likely choose the pragprag-matic attribute. The reason being that it is harder to justify choosing the hedonic attribute over the pragmatic attribute (Diefenbach and Hassenzahl, 2011). However, as Diefenbach and Hassenzahl (2011) note, this is not due to the hedonic attribute being less important to the user. The pragmatic attribute could be described as the bare minimum, being sorely missed when not present but otherwise going unnoticed (Diefenbach and Hassenzahl, 2011). Hedonic attributes on the other hand will enrich the experience itself with its presence. Diefenbach and Hassenzahl (2011) describe other studies which observe that when the hedonic attribute is given a suit-able justification it tends to be chosen more than its pragmatic counterpart. For instance, a consumer may be more likely to buy a non-necessity such as ice cream if it is discounted, compared to buying essentials such as toilet paper when it is discounted. Moreover, when presented on its own, without comparing it to a pragmatic counterpart, studies have shown

(28)

the hedonic attribute to be more favorable than the pragmatic attribute (Diefenbach and Hassenzahl, 2011). As Diefenbach and Hassenzahl (2011) explain, this means forcing a choice between the hedonic and pragmatic attributes could lead to people choosing the option they find more justifiable but not the one they actually want. Furthermore, favoring the creation of mostly pragmatic products could fail in generating an emotional attachment (Diefenbach and Hassenzahl, 2011). Without emotional attachments products can easily be replaced (Diefen-bach and Hassenzahl, 2011). However, no matter whether the product includes hedonic or pragmatic attributes, they both affect the product’s consequences (Hassenzahl, 2003). The consequences are determined not only by the apparent product character but also on the user’s own situation (Hassenzahl, 2003). This is shown in the illustration of Hassenzahl’s model, in the user’s perspective in Figure 2.1. This entails that certain product characters might be more or less suitable for different situations (Hassenzahl, 2003). Hassenzahl (2003) exemplifies this by using an ATM with the intended attribute "understandable" which it achieves by taking the user through numerous steps. Hassenzahl (2003) notes that if you are a new user this would indeed be very helpful, but if you are used to the machine or are in a hurry it could be more annoying than satisfying. As such, Hassenzahl (2003) concludes by saying that the attribute of "understandable" might not have changed, but the situation clearly changed the consequences of the product.

Since the situation is such a crucial aspect of creating the final appeal of the product, Has-senzahl (2018) recommends designers to have the users’ overall motivations in mind through recognizing their usage modes. A usage mode is how the user acts in order to fulfill their cur-rent goal (Hassenzahl, 2018). Hassenzahl (2018) describes two diffecur-rent usage modes: action modeand goal mode. Action mode is when the goal is secondary to the actions the user takes (Hassenzahl, 2018). Hence, the users are spontaneous when performing the actions and not limited to one set goal, instead new goals are created continuously throughout the process (Hassenzahl, 2018). While in action mode, as Hassenzahl (2018) describes, users needs stimu-lation to avoid boredom. In contrast, too much stimustimu-lation during goal mode can instead be a cause of anxiety or frustration (Hassenzahl, 2018). This is due to the users focusing more on fulfilling the goal rather than the actions they take during goal mode. Consequently, efficiency and effectiveness are qualities considered to be important to a user in goal mode (Hassenzahl, 2018). Hassenzahl (2018) explains that while the usage mode affects the overall appraisal, it does not change the perceived product characters, rather the characters are judged based on their suitability to the current usage mode.

Another model which describes the components of the user experience is the CUE-model (an abbreviation of Components of User Experience model) by Thüring and Mahlke (2007), it can be seen in Figure 2.2. It uses the differentiation between the Instrumental and Non-Instrumental qualities (Thüring and Mahlke, 2007). These qualities derive from Hassenzahl’s distinction between the pragmatic and the hedonic attributes (Minge and Thüring, 2018a). With the instrumental being the pragmatic while the non-instrumental is the hedonic at-tributes of a product.

As seen in Figure 2.2, the CUE-model divides the user experience into three components which affects the overall appraisal of the product (Thüring and Mahlke, 2007). These components are the perception of the instrumental and non-instrumental qualities as well as the user’s

(29)

emotional reaction to the product (Thüring and Mahlke, 2007). The first iteration of the CUE-model only considers the user’s emotional reaction to be affected by the perception of the qualities (Thüring and Mahlke, 2007) and not that the user’s emotions are also affected by the perception. However, as Minge and Thüring (2018a) describe, the updated version of the meCUE model takes into account that the emotions can affect the perceptions as well.

Figure 2.2: The Components of the User Experience (CUE) model created by Thüring and Mahlke (2007). Adapted from Minge et al. (2017).

Importantly, the CUE-model states that the user experience components are established by the functionalities of the product (system), the user’s characteristics and skills, as well as the context of the task (Thüring and Mahlke, 2007). This can be seen in the "Interaction characteristic" box in Figure 2.2. Hence, according to the CUE-model, the user experience is considered to be an accumulation of the product features, the user, and the context. In turn these create the perception of the qualities as well as the user’s emotional reaction, which eventually become the overall appraisal of the product (Thüring and Mahlke, 2007). In the model this is shown as a consequence of the perception of instrumental and non-instrumental qualities, along with emotions, which is visible in Figure 2.2.

As described both by Hassenzahl (2003) and Thüring and Mahlke (2007) among others, the user experience is affected by many different aspects of a product as well as the user of a product. One of the many features which could have an effect is the GUI of a product.

2.1.2 The Graphical User Interface

There are several guidelines and principles proposed by scholars to heighten usability within GUIs (Yee et al., 2012). In this section, common GUI elements, theory regarding vision and using icons and labels in GUI will be presented.

(30)

Graphical User Interfaces from Variability Models. A check button (Figure 2.3) can either be true or false. A list box (Figure 2.4) can include numerous items or allow choosing between true and false. A radio box (Figure 2.5) entails only choosing one option out of a set of items. Check boxes (Figure 2.6) can be used to select numerous items in a set. A text box (Figure 2.7) can be integers or real numbers. A slider (Figure 2.8) can also be either an integer or real.

Figure 2.3: An example of a check button.

Screenshot by author from Microsoft (2019) Figure 2.4: Example of a list box. Screenshot by_{author from Microsoft (2019).}

Figure 2.5: An example of radio boxes.

Screen-shot by author from Microsoft (2019). Figure 2.6: Example of check boxes. Screenshot by author from Microsoft (2019).

Figure 2.7: An example of a text box. Screenshot

by author from Microsoft (2019). Figure 2.8: An example of a slider. Screenshot by author from Microsoft (2019).

A dialogue box is an interactive box which informs the user about an issue, and forces the user to click on buttons in the dialogue box in order to make a choice (Tulaskar, 2018). This is shown in Figure 2.9 where the user is asked to say "OK" or "Cancel" regarding a choice offered to them. Balloon help can appear spontaneously, without any input from the user, and they include text or images that informs the user about anything (Tulaskar, 2018). See 2.10 for an example of how balloon help may be represented. Tool-tips appear when a user hovers over a specific element in the GUI, and they may include text or images to inform the user about a topic relevant to the hovered element (Tulaskar, 2018). Figure 2.11 showcases how a tool-tip may look.

(31)

Figure 2.9: An example of a di-alogue box. Adapted screenshot by author from Microsoft (2019).

Figure 2.10: An example of a bal-loon help. Adapted screenshot by author from Microsoft (2019).

Figure 2.11: An example of a tool-tip. Screenshot by author from Microsoft (2019).

When creating GUI-elements the user’s visual focal point is important to take into consid-eration to ensure important information is not missed (Johnson, 2021). As Johnson (2021) explains objects in the peripheral vision are easily missed or misinterpreted, especially if the object is still. As such, a common issue is the user missing error messages because it appeared in the peripheral and failed at drawing attention to itself (Johnson, 2021). GUI can incorpo-rate icons or text, which each have different advantages and disadvantages when it comes to learnability to present information to a user (Wiedenbeck, 1999).

The choice of using exclusively labels, exclusively icons or a combination of both have an im-pact on the actual usability and the perceptional usability of a software (Wiedenbeck, 1999). For inexperienced users, Wiedenbeck (1999) argues that labels or labels in combination with icons are preferable since they aid in the initial learning process of using a new software. In contrast, for more experienced users, including icons is favorable as the labels become redun-dant and exclusively using labels in general decreases the perceived usability (Wiedenbeck, 1999). As such, the usefulness of labels decreases the more the user familiarizes with the software and its functionalities (Wiedenbeck, 1999). Moreover, as Johnson (2021) describes, perception is also affected by what the user is searching for or expecting. Users scan a website to quickly find the information they need and filter out the rest (Johnson, 2021).

These small GUI elements are important parts of an interactive website. They can be designed and combined in numerous ways, and the manner in which they are portrayed plays a large part in creating the user experience. In order to choose which GUI elements to utilize, the designer needs to understand the user experience.

2.1.3 Understanding the User Experience

When developing the user experience it is important to gain empathy for the user and under-stand their current experiences as well as their potential reactions to implemented concepts (Gibbons, 2019). This section describes the tools Customer Journey Map and Hierarchical Task Analysis both of which are used to gain an understanding of the user experience. A tool to gain empathy is a customer journey map (Wikberg Nilsson et al., 2015). It showcases the different aspects of the user’s journey by defining several features of the interaction, as well as the user’s emotions and opinions. They can be based on a persona, real research and observations of a user, or the designer putting themselves in the user’s shoes. The customer journey map can be used for several purposes (Stickdorn et al., 2018 ; Wikberg Nilsson et al.,

(32)

2015). For instance, it can be used to evaluate and compare concepts, gain empathy for the user (Wikberg Nilsson et al., 2015), as well as showcasing potential improvement areas (Stickdorn et al., 2018).

A customer journey map consists of two axes (Wikberg Nilsson et al., 2015). A general example of how a customer journey is established is shown in Figure 2.12. The horizontal axis contains the time from when the user starts their journey to when it ends. The vertical axis contains different categories, which are to be filled in with information about the user at different points throughout their journey (Wikberg Nilsson et al., 2015). The categories can differ between projects (Wikberg Nilsson et al., 2015) but the categories suggested by Stickdorn et al. (2018) include: specific phases of the journey, the user’s actions, and an emotional journey which is a graph of the user’s positive and negative emotions throughout the journey. Wikberg Nilsson et al. (2015) exemplify a customer journey map which include some of the categories covered by Stickdorn et al. (2018), as well as categories including what the user says and their experience. Furthermore, Wikberg Nilsson et al. (2015) include categories for evaluating the system, such as a category for pros and cons, and potential improvements. Customer journeys can explain what the user actually does and feels during an interaction, but there are other methods which can be used to explain what the user should do.

Figure 2.12: A general example of how the framework for a customer journey may look.

To showcase and map out the activities required during human machine interactions, several different task analysis methods have been established (Stanton et al., 2005). One such method is the Hierarchical Task Analysis (HTA), which is widely used due to its adaptability (Stanton et al., 2005). A general example of how the framework of an HTA may look is shown in Figure

(33)

2.13. In an HTA the main task is divided into sub-tasks which are then further broken down into operations. As Stanton et al. (2005) describe, to fulfill the main task, the sub-tasks have to be fulfilled, and in order to fulfill the sub-tasks the user has to perform the operations. Hence, the operations are at the lowest level in the HTA and cannot be further divided. In contrast, the sub-task can, if needed, be broken down into more sub-tasks, creating more levels to the HTA until proper operations can be identified (Stanton et al., 2005). The data used for creating the HTA, can be collected in various ways, such as walkthroughs or observations of usage (French et al., 2019; Stanton et al., 2005). French et al. (2019) suggest visualising an HTA in a hierarchical format since it easily shows the interconnected tasks and order of completion as exemplified in Figure 2.13.

Figure 2.13: A general example of an established HTA.

The user experience may be difficult to truly define, and many different methods for trying to understand it exist. This theory must then be applied to the product or service of interest, which in this case is product configurators. While this section has explored the different aspects of the user experience, the next section looks further into best practices regarding functionalities within product configurators.

2.2 Product Configurators

In this section product configurators are explained in terms of their important functions, such as allowing trial and error, adapting to the competence and personality of the user, and helping the user though the process.

The features needed to create a successful product configurator can be identified according to Fettermann et al. (2017), but in contrast Streichsbier et al. (2014) say the user’s purchasing decision is not majorly influenced by the availability and positioning of certain features in a configurator. Mostly, the realism of the visualization of the configured product and the feeling of being entertained throughout the process matters to the user (Streichsbier et al., 2014). Furthermore, users say they also care about the trustworthiness of the interface and how easy it is to use (Streichsbier et al., 2014). The user must feel like the options within the configurator match both their own needs and the business’ brand image (Randall et al., 2005). Kang (2017) investigated six different elements of a user interface which could

(34)

potentially increase loyalty among customers. The results were that three of the elements proved to have a positive impact: character, choice and customization. Character refers to aesthetics of the website and the brands characterization. Choice refers to the amount of available products to the consumer, while customization refers to the degree of freedom the consumer has while customizing. Four areas are reoccurring themes within the literature about the user experience in product configurators. They are the following three areas: Trial and error, user competence, user personality, and helping the user.

Trial and error. The user should be aided in understanding their needs from trial and error within the configurator (Von Hippel, 2001), this may be described as understanding the product subjectively. H. Zhao et al. (2019) say that the user’s trial and error learning in product configurators is supported when the process is adapted to the product type, the personality of the user, and the type of configuration that is applied. Users should be able to test their design within the toolkit to see the implications of the design and thus learn by doing (Von Hippel, 2001). This could also be facilitated by letting users archive their designs and come back to a previous configuration in future iterations, and by showing how a user can reach a certain functionality by changing the design (Randall et al., 2005; H. Zhao et al., 2019). The configurator should help the user compare different variants of their design by displaying information about the variants side by side (Randall et al., 2005; H. Zhao et al., 2019). The configurator should also allow the user to undo decisions and navigate backwards in the configuration process as a way of encouraging trial and error within the configurator (Abbasi et al., 2013).

User competence. The information about the configured product and the available options must be clear, sufficient, and easy to understand for a novice user in order for them to continue to use the configurator (Stevens and Jouny-Rivier, 2020). This is typically achieved by using terms the user is familiar with, such as the functionality of a component (Von Hippel, 2001). This could also entail using photos, videos, and scales to help the user understand parameters that are otherwise lost in a digital interface, such as size, weight, or smell (Randall et al., 2005). One way of showing the functionality of a product is using images and describing text as in Figure 2.14, where the product is described in terms of similar smells instead of the combination of chemicals that is ordered. This would aid the user with understanding the product subjectively. This is also supported by Randall et al. (2005), as untrained users need aid in understanding the implications of the options while advanced users can use parameter values. The way the options are presented should follow a standard that differentiates the type of decision that has to be made (Abbasi et al., 2013). Such as if it is mandatory or optional, bounds of intervals, what the constraints are, search mechanisms, selection and deselection, spell check, price indicator, and default values (Abbasi et al., 2013). Due to the different abilities of users Randall et al. (2005) suggest that the configurator should be customized to this. It should let a user decide which type of interface they want to face and at what point in the configuration process they want to start. Moreover, H. Zhao et al. (2019) suggest having multiple viable paths to achieve the end goal. Another way of achieving a more customized interface is suggested by Abbasi et al. (2013), which is to let the user alter whichever variables they wish to, while the unedited values are automatically assigned default options.

(35)

Figure 2.14: An example of how a perfume configurator explains the function of a perfume (i.e. the smell) by using images. Screenshot by author from the website Fragrance by Me (2021).

User personality. The personality of the user also affects their configuration experience (Ling et al., 2020). Users with high self-esteem enjoy having more creative freedom in the configurator, i.e. having a larger solution space, while less confident users benefit from in-spiration from other designers (Ling et al., 2020). Ling et al. (2020) suggests that a way of satisfying both ends of the spectrum is to first guide the user through a configuration process, and then let them edit further if they feel dissatisfied. A configurator should not immediately guide the user to the next step of the process (Abbasi et al., 2013). The reason is that even though all constraints are fulfilled it does not mean the user is happy with the chosen options, and the user should therefore have the freedom to continue modifying their design (Abbasi et al., 2013). The user should be aided with standardized modules to use for their design, this lets them focus on other, more unique design elements (Von Hippel, 2001; H. Zhao et al., 2019). In Figure 2.15 an example of how a couch configurator provides modules is shown, image is taken from Ikea (2021a). Providing modules like this gives the user a semi finished part consisting of seat, legs, backrest, and arm rests. Hence, they do not have to design these independently. Similarly, Randall et al. (2005) advocates for limiting the design process by giving the users a pre-determined, recommended product, which they can improve upon with the configurator. For example, the configurator could gather data about the user and then suggest a good start-configuration for them (H. Zhao et al., 2019). The inclusion of a sharing function where users can show the designs they feel proud of can benefit the user experience, especially for users with a high level of creative confidence (Stevens and Jouny-Rivier, 2020). This should both heighten the enjoyment of finishing a product configuration, and aid other users with their creativity in the design process (Stevens and Jouny-Rivier, 2020). The latter perk is mostly enjoyed by users with lower creative confidence (Stevens and Jouny-Rivier, 2020). Sharing could occur via chat rooms, galleries, comments, ratings or a historical record of configured designs (H. Zhao et al., 2019).

(36)

Figure 2.15: An example of how a couch configurator provides modules. Providing modules like this gives the user a semi finished part consisting of seat, legs, backrest, and arm rests. Hence, they do not have to design these independently. Screenshot by author from Ikea (2021a)

.

Helping the user. The configurator should provide explanations of relevant constraints when the user wants or needs to know more about them (Abbasi et al., 2013). Furthermore, the configurator should automatically fix as many errors as possible without involving the user (Abbasi et al., 2013). The user should be provided with such real-time help when working in the configurator, in addition to tutorial videos and written solutions to common questions and problems (H. Zhao et al., 2019). This type of guidance makes the user feel safer while working in the configurator (H. Zhao et al., 2019). Abbasi et al. (2013) writes that the user should be informed if a constraint causes options to be enabled or disabled. By indicating where conflicting decisions have been made and showing ways to solve the conflicts the configurator can aid the users in solving their own problems (Abbasi et al., 2013). H. Zhao et al. (2019) claim that the user needs to be able to manipulate the 3D-model directly and receive instant feedback on their actions in order to feel like they are in control of the configurator. The users should be taken through stages of increasing difficulty, and always have information about how much of the configuration has been completed (H. Zhao et al., 2019). This can be achieved by numbering steps in the process, letting them go back and forth between steps, as well as seeing what decisions have been made and which ones still remain (Abbasi et al., 2013).

In conclusion of this section, there are many opinions about what makes a good or bad product configurator, but there also seems to be some consensus among these opinions. One way of learning how various features may be implemented to create a good user experience is to perform user studies.

2.3 User Studies

The designer cannot always be trusted to create the most usable product, they are affected by their own biases and knowledge, and may not be part of the target group (Quesenbery, 2004). Therefore user tests are often needed to perfect a design (Quesenbery, 2004). When performing user studies there are three aspects of the user behaviour which can be investigated;

Configuring Configurators : The creation of UX guidelines for parametric product customization