Does Enhanced Usability Increase Trust? : A case study on a Swedish snus e-commerce application

(1)

Spring Semester 2020 | LIU-IDA/LITH-EX-G--20/004--SE

Does Enhanced Usability Increase

Trust?

A case study on a Swedish snus e-commerce

application

Filip Brunander

Albin Folkesson

Henrik Hörnlund

Simon Larsson

Johan Mäkikaltio

Matthew Soulaka

Amanda Teir

Carl Ågren

Handledare/Mentor, Diba Rezaie Examinator/Examiner, Aseel Berglund

(2)

Abstract

The purpose of this study was to examine the importance of usability to enhance perceived consumer trust in a web application selling snus. This was done by answering the research question “Does usability affect Swedish snus consumers trust in snus e-commerce applications?”. The concept of trust is especially important in an e-commerce environment where consumers have to rely heavily on the seller to follow through on the transactions, and the web application is the main way of interacting with the seller. In order to do so, a web application, “Dosan”, was developed and tested to collect data about usability and the users perceived trust. The development was split into three iterations. Each iteration was composed of a development phase and a testing phase where five users tested and evaluated the web application. To approximate the level of usability and trust, the users’ task efficiency and effectiveness was recorded. After completing the test the participant was asked to fill out a questionnaire containing Likert scale questions measuring satisfaction and trust. In addition, participant comments were recorded and used as a basis for further development. Each iteration showed an increase in trust and usability, which led to the conclusion that there is a positive correlation between usability and trust, and indicates a causal relationship between the two.

Sammanfattning

Syftet med denna studie var att undersöka vikten av användbarhet för att förbättra up pfattat konsumentförtroende för en webbapplikation som säljer snus. Detta gjordes genom att besvara frågeställningen “Påverkar användbarhet svenska konsumenters förtroende för en e-butik för snus?”. Konceptet förtroende är i synnerhet viktigt i en e-butiksmiljö där konsumenter måste förlita sig kraftigt på att företagen fullföljer transaktionerna, och webbapplikationen är det huvudsakliga sättet att interagera med säljaren. För att genomföra detta utvecklades och testades en webbapplikation, “Dosan”, för att samla in data om användbarhet och användarnas uppfattade förtroende. Utvecklingen var uppdelad i tre iterationer. Varje iteration bestod av en utvecklingsfas och en testfas där fem användare testade och utvärderade applikationen. För att uppskatta nivån av användbarhet och förtroende registrerades användarnas prestationsförmåga och verkställighetsförmåga under testuppgifterna. Efter testet ombads deltagaren att fylla i ett formulär innehållande frågor med Likert-skalor för att mäta belåtenhet och förtroende. Tilläggningsvis spelades deltagarnas kommentarer in och användes som bas för vidare utveckling. Varje iteration visade förhöjd användbarhet samt förtroende, vilket ledde till slutsatsen att det finns en positiv korrelation mellan användbarhet och förtroende och indikerar ett kausalt förhållande mellan dessa två.

(3)

1 Introduction

The use of e-commerce has exploded since it was introduced, internationally as well as in Sweden. A report made by Postnord [1] states that 66% of the Swedish population purchased goods online at least once a month in 2018, and that the total revenue of business to consumer e-commerce amounted to a total of 92.8 billion SEK for the year. The report claims that consumer reasons for choosing e-commerce solutions compared to traditional in-store purchases are many, such as convenience, easy price comparison, greater choice of payment methods and not being restricted by geographical distance or store opening hours. In the article Building trust online [2] it is stated that a lack of consumer trust towards marketers is the greatest impairment for digital transactions. Consequently, a lack of trust has a negative effect on revenue, as it hampers intent to purchase.

According to the authors of the article The Doctors’ Choice Is America’s Choice [3] claim that trust has been a key factor for tobacco sales for a long time. The authors further elaborate with the example of the well-known ad campaign “Doctors smoke camels” by Camel in the 1940’s. In this campaign, doctors were used as a sales pitch since they convey a message of trustworthiness and dedication [3]. The authors of Consumers’ evaluation toward tobacco companies: implications for social marketing [4] claim that the tobacco industry has also contributed to a variety of CSR initiatives, such as charita ble donations and assisting tobacco growers, in order to improve company perception. According to the authors, the benefits of CSR initiatives include but are not limited to increased trust and insulation from bad publicity. Since snus retailers are a part of the sales chain they are affected by, and able to affect, consumer trust in tobacco wares.

According to the article An integrative model of organizational trust [5], trust is the willingness to take a risk without being able to control the actions of the second party. As stated in the article The impact of interface usability on trust in Web retailers [6], this is especially important in an e-commerce context where the consumer is entrusting the business with not only their personal details [7] but also a payment without having the product in hand. Furthermore, Roy et al argues that usability of a web application has a significant effect on the perceived ability of the marketer, which in turn increases consumer trust [6]. Usability is defined by ISO [8] as “the extent to which a system, product or service can be used by specified users to achieve specified goals with effectiveness, efficiency and satisfaction in a specified context of use”. Roy et al. [6] specifies the interface areas of navigation, consistency, learnability, user guidance and perception as the critical factors when discussing usability in relation to trust.

1.1 Aim

The purpose of this study was to examine whether increased usability would increase the consumers trust in a web application selling snus to Swedish consumers. In order to test this, the web application ‘Dosan’ was constructed and tested during development to see how usability and trust changed throughout the development phase.

1.2 Research Question

Does usability affect Swedish snus consumers trust in snus e-commerce applications?

According to a report from Postnord [9] e-commerce is on the rise in Sweden and have been so for several years. The report states that Swedish consumers continue to increase their online spending. This behavioural shift might be due to factors including but not limited to convenience, price and selection.

(7)

2

For a consumer to engage in an agreement with a retailer the consumer must trust in the company’s ability and intention to actually deliver on what’s promised [5]. For retailers this means that the ability of gaining the trust of consumers is an essential element of being a successful business.

Something that has been proven to increase consumers trust in a company is face-to-face interactions with its staff [10]. Online retailers generally do not offer such interactions to the same extent as brick-and-mortar businesses which is an almost defining characteristic of online retailers, that can cut costs and increase scalability of the business. This does however raise the question of how to achieve consumer trust under such conditions. The authors of Trust and satisfaction, two stepping stones for successful e-commerce relationships: A longitudinal exploration [11] established that consumer trust is an interesting topic for online retailers. This is a relevant question for online retailers of the Swedish snus market, as consumers still tend to make most of their snus purchases in physical stores, as found in appendix 9.4 (Market survey). Higher levels of trust in online snus retailers could increase consumers’ proclivity to convert their physical purchases to online purchases.

To a large extent a consumer’s interaction with an online retailer takes place on the company’s web application [7]. Under such circumstances it is interesting to examine how the web application itself can affect consumers trust in the business.

Usability is an influential dimension with regards to user’s interaction with a web application. ISO’s definition of usability can broadly be summed up by the user’s ability to perform intended tasks successfully, the time it takes for the user to perform these tasks and the level of overall satisfaction the user experiences from these interactions.

This paper aims to shed some light on the question whether the usability of a web application can affect consumers trust in a retailer on the Swedish snus market.

1.3 Delimitations

In order to narrow down the definition of usability, the ISO 9241-11:2018 [8] standard is used. This is a widely used and accepted standard which enables comparison with several similar studies.

The test groups consisted of participants between 18 and 29 years old. This is not representative of the Swedish population, and any conclusions may therefore not be fully representative in a national

perspective. Testing was limited to usability and the user’s experienced trust. Furthermore, the testing will be done on computers to minimize the need of Responsive Web Design. The conclusions drawn about usability design therefore only applies to computer browsers.

Moreover, due to the restricted time allowed for the study, the learnability aspect of usability was discarded.

(8)

3

2 Background

Snus is a tobacco product consumed orally, and comes either as a loose, moist powder, or in portion bags. The predecessor to snus, snuff, a dry powder consumed through the nose, reached Sweden in the late 16th century and became a status symbol for the wealthy nobility [12]. It was only with the introduction of moist snus in the early 19th_{century that the product became widely consumed by common people, however mainly} among the male part of the population [13]. When Sweden in 1991 applied for membership in the European Union, directives from Brussel indicated that all forms of oral tobacco was to be forbidden, but due to extensive protests from the Swedish population, Sweden was exempt from the ban [14].

In Sweden snus is a product regulated by law and is only to be sold to customers who are 18 years or older [15]. According to a study conducted by SCB [16], 13.6% of the Swedish population in 2018 said they consume snus daily, which is an increase by 2.1 units since 2008 . The study stated that the number of female consumers has increased with almost 50% in the same time period, likely because of product diversification such as the development of flavoured, white snus with a lower amount of nicotine.

(9)

4

3 Theoretical Framework

This section will lay the theoretical framework needed to answer the research question. This is done by introducing the necessary theory that is relevant for answering the research question. First, the theory behind the connection between usability and trust is presented. Then the concept of usability is covered by presenting the definition of usability and its fundamental factors. Consequently, trust is described in a similar way. To complete the chain from usability to trust to snus e-commerce applications, the relation between trust and selling tobacco products online is described. Furthermore, the underlying theory for the methods used to answer the research question is reviewed, this includes measurement methods of usability and trust. The theoretical framework connects usability to trust, and trust to e-commerce applications. Throughout the theoretical framework two concepts about parties involved in a trusting relationship will be used to better describe these parties. First, the trustor which is the party that trusts the other party. Secondly, the trustee which is the party that is trusted. These two parties and a model connected to them will be presented in section 3.3.1.

3.1 The connection between usability and trust

In the article The role played by perceived usability, satisfaction and consumer trust on website loyalty [17] the authors came to the conclusion that when a user perceive a website as usable, the trust is increased. The authors confirm that usability influences the perceived competence significantly, which is a part of the trustor’s ability and therefore plays an important role in the perceived characteristics of the trustor. Carlos et al. [17] argues that greater usability might be perceived as the characteristic benevolence of the trustee. It is stated that the consumer’s (trustor) needs are an easy to use web application, and by creating such a web application the perceived benevolence by the consumer could be increased.

According to Roy et al. [6] trust is a dynamic concept that begins developing when the customer experience their first impression, but develops as the customer obtain new information or experiences. The authors explain that the first impression is the most critical in an online transaction since it is during a very short period of time compared to B2B-cases.

Roy et al. [6] conducted a survey where the Mayer trust model [5], described in section 3.3.1, is connected to different factors. The result from this were that ease of navigation, ease of learning, perception and support influence the perceived trust of an online retailer. The authors explain that ease of navigation is based on the ease of finding what one is looking for in a website and being able to tell where you are on the website. Ease of learning is achieved by the use of simple and clear language, clear display and logical grouping. Perception applies to the interface design which should consider the human perceptual organization limitations. The authors further explain that the mental workload decreases while the learnability increases as the support to the user increases. Consistency is not proven to affect trust in the article, but it is possible that it affects usability without affecting trust. Moreover, the article states that consistency increases the more a website follows the internal and external standards and conventions. The factors that were tested in this survey seemed to mostly affect perceived ability.

(10)

5

Figure 1: The connection between usability factors and perceived trustworthiness [6].

Once the user perceives that the required level of honesty, benevolence and competence is reached, the user will feel satisfied [17]. Furthermore, the authors claim that trust was found to be direct and positive dependent on the degree of consumer website satisfaction.

3.2 Usability

This section covers the concept of usability by presenting the definition of usability and its fundamental factors.

3.2.1 Definition of usability

ISO’s [8] definition of usability is “the extent to which a system, product or service can be used by specified users to achieve specified goals with effectiveness, efficiency and satisfaction in a specified context of use”. ISO continues to define effectiveness as “accuracy and completeness with which users achieve specified goals”, efficiency as “resources used in relation to the results achieved” and satisfaction as “extent to which the user’s physical, cognitive and emotional responses that result from the use of a system, product or service meet the user’s needs and expectations” [8].

3.2.2 Designing for usability

In the article Usability evaluation of users’ experience on some existing e-commerce platforms [18], an extensive survey was performed where questionnaires were sent out to 222 users of five different e-shops, containing questions with an intent to get information about in what context users are utilizing the internet and their opions about the websites. Based on this, a clear connection was identified from how the respondents have answered the questions in appendix 9.1 (User experience questionnaire) and how they then rated the general services offered.

The survey results were then broken down by Babatunde et al. [18] into a set of recommendations of things to consider when building an e-commerce platform:

• Design with a focus to be easily navigable

• Have a lot of product details to make the products “come to life” • Efficient site and payment security

• Prompt delivery time

(11)

6 • Easy way of exchanging information

• User reviews

Furthermore, in Enhancing the Explanatory Power of Usability Heuristics [19], Nielsen lists the top nine usability heuristics to take into consideration for user interface design:

Visibility of system status

The system should provide the user with sufficient feedback on actions performed on the interface.

Match between system and the real world

Discourse in feedback should be adapted to the user. The language should be recognizable to the user and come in a logical order.

User control and freedom

Recovering from an error should be easy, the user should be able to return to the previous state without extensive dialogue.

Consistency and standards

Follow platform conventions in order for the user to feel in control. ”Express the same things the same way”.

Error prevention

Remove elements prone to errors, and if the action is irreversible, ask the user for confirmation before execution.

Recognition rather than recall

Make the important elements visible in each view, the user should not have to recall important information.

Flexibility and efficiency of use

By making shortcuts to important functions available, the system can cater to experienced and inexperienced users simultaneously.

Aesthetic and minimalist design

Only display relevant information to avoid clutter.

Helping users recognize, diagnose and recover from errors

As opposed to error prevention, this heuristic adheres to the use rs understanding of what went wrong, how to fix it and also how to avoid it in the future. This can be provided with dialogue and error codes.

In the article Assessing Web Site Usability in Retail Electronic Commerce [20] the authors found several constructs and their respective connections to usability:

Identity

The purpose of the web application should be clearly stated to increase web site traffic. Furthermore, the website should be unique with characteristics that differ from similar web applications to increase its usability.

Responsiveness

Customer service plays a major part in the customers perceived service quality. Feedback options and FAQ are example of important functions that increases the usability of a web application.

(12)

7

Additionally, in the article E-Commerce Website Quality Assessment based on Usability [21] it is stated that there are several more factors that affect usability, some of which are built up by sub-parameters:

User Satisfaction

• Customer support services

Essential in creating a good image of the company and web application. • Successful Transactions

Important with trustworthy and successful transactions, connects usability and trust. An easy checkout helps both the company and the customer.

Simplicity

Important in order to minimize the risk of losing the interest of customers. • Ease of Use

Important for web usability since more customers will look for the web application if it is easy to use.

• Frequently Asked Questions (FAQ)

Plays a key part since a well-developed FAQ page shows that the developer understands the needs of the customers.

• Error Notification

Errors that occur, either because of the developer or because of mistakes by the customer, should have great error notifications to help remove the error. The information should not be redundant but at the same time enough to identify the error.

Attractiveness

Designed to attract customers. • User Friendly

The web application should be friendly enough to enable a user to u se the developed application quickly and easily. Menus and links should be placed in appropriate places and each page should include an easy way to return to the home page. Services like the shopping cart should also include other functions that the user might find useful, for instance shipping options, change the quantities of products or to continue shopping. • Appearance

The design of the web application should be aesthetically pleasing.

Efficiency

Important factor since many of other factors may depend on it. • Product Categorization

Keep things simple and organize the products in a hierarchy.

Searching

It should be possible to search for products and to do so with efficiency and ease. It should also be easily accessible. Furthermore, filtering could be used to make it easier for the user to find a specific product or type of product.

Product Information

(13)

8

Furthermore, in the article Key factors for developing a successful e-commerce website [22], several of these factors above are highlighted as important aspects to develop a successful website. Additionally understandability, interactivity and response time are mentioned as particularly important.

3.3 Trust

Mayer et al. [5] defines trust as: “the willingness of a party to be vulnerable to the actions of another party based on the expectation that the other will perform a particular action important to the trustor, irrespective of the ability to monitor or control that other party.” In the article the authors differentiate trust from taking a risk, describing trust as the willingness to take risk. The authors explain that trust is often mixed up with cooperation, confidence, and predictability, therefore the differences between these constructs are emphasized in the article.

3.3.1 A model of trust

Mayer et al. [5] specify two important parties; the party that trusts (trustor) and the party that is trusted (trustee) while building the organizational focused model. Even though trust is not mentioned in an organizational setting in this report but rather a retail setting, the mentioned model will still be used due to it being referred to in multiple other articles related to web application and trust. The trustee will be the online retailer while the user/customer will be the trustor.

The model below is based on the following factors: trust, risk, and outcomes. Trust for a trustor towards the trustee is based on the following perceived factors; the trustee’s ability, benevolence and integrity which are described below as the trustee’s characteristics [5]. Furthermore, the propensity to trust does also impact the trustor’s decision to trust and could be the described as the general willingness to trust other parties. This willingness generally depends on prior experiences, personality, and cultural background. As stated in figure 2, risk also plays a major role in a relationship where the presence of trust is needed for it to function [5].

Figure 2: A model of trust describing the connection between trust, risk and outcomes [5].

There are multiple approaches to understanding to which extent a party is trusting another party [5]. One of them is to analyse the trustor’s propensity to trust another party, something that also is brought up in the article Reconceptualizing the determinants of risk behaviour [23]. Another approach is to observe the

(14)

9

characteristics of the trustee [5]. Since it is possible to change the characteristics of the online retailer, the latter approach will be the one taken in this report.

3.3.2 Characteristics of the trustee

Mayer et al. [5] analyse the characteristics of trustee’s mentioned in several other articles regarding trust. The most frequently used characteristics was collected and the three most common are ability, benevolence, and integrity.

According to the authors, ability is the overall competence, skills and characteristics of a party that enable it to have influence within a domain [5]. It is worth emphasizing that it is domain-specific. While a party could have great ability within a specific domain it could be close to non-existent in others [24] [5]. For example, while an employee might have great ability within a certain technical domain, the employee may never be entrusted to close out a big deal since the technical ability do not imply that the person has a great negotiation ability [5]. Mayer et. Al [5] defines benevolence as to which extent a trustee wants to do good by the trustor, without taking the self-interest of the trustee into account. Furthermore, the authors are proposing that if there is benevolence between the trustor and trustee, it implies that the trustee has specific connection with the trustor. Lastly the authors bring up the connection between integrity and trust. This connection is based on the trustee standing by principles that is found acceptable by the trustor.

To summarize these three characteristics mentioned by Mayer et. Al [5]:

” Trust for a trustee will be a function of the trustee’s perceived ability, benevolence, and integrity and of the trustor’s propensity to trust.”

While these three characteristics may vary independently of each other it does not mean that they are unrelated [5]. Furthermore, the authors claim that even if the trustee has managed to develop some of these characteristics well, it is enough that one of the other parameters is low in order to significantly reduce the trust of the trustor. The significance of each characteristic of the trustee differ depending on which state the relationship is in. Early in the relationship before the development of any significant benevolence, integrity will have the greatest impact [5]. Since an online retailer web application was developed in this study and the first impression is critical for these kinds of applications according to Roy et al. [6] integrity was an important parameter to consider.

3.3.3 Cause of trust

In the article Measuring Online Trust of Websites: Credibility, Perceived Ease of Use, and Risk [25] it is shown that honesty and expertise intertwined. The information that the user interpreted as good was information that was perceived as both honest and that it came from a source with expertise. The authors state that if any of these parameters were low, the value of the information decreased significantly or completely lost its value.

3.3.4 Risk

In the article Reconceptualizing the determinants of risk behaviour [23] risk is explained as a characteristic of decision defined from the uncertainty of potential significance or disappointing result of the specific choice. Three key dimensions are defined by the authors. Firstly, the outcome uncertainty which is the lack of knowledge about the specific case. It is divided into two factors, the uncertainty about what outcomes that are possible and then the uncertainty about the probability of these different outcomes. Secondly, the outcome expectations propose that risk could both be the expectation of a negative outcome as well as a positive one. This is based on that it is not the expected outcome itself that causes the risk to increase but rather to which degree the outcome would be disappointing to the decision maker. Lastly, the outcome

(15)

10

potential takes the best and the worst outcomes into consideration. These three factors affect the perception of the risk and plays a major role in deciding whether a risk is worth taking [23].

3.3.5 Designing for trust

According to the article An overview of online trust: Concepts, elements, and implications [26] there are four main frameworks stated for trust-inducing features of design that improves the perceived trust from a consumer viewpoint:

Graphic design

In terms of graphic design, three dimensional images covering in total at least half of the screen was preferred by consumers. Furthermore, images should be of high quality to enhance the perceived ability of the business. A cool toned interface colour scheme, with a “moderate pastel” main colour with low brightness and a symmetrical execution was also found to have a positive impact on consumer trust.

Structure design

This feature addresses the organization and structure of elements on the interface. Simplicity and clear design with few irrelevant elements reduce consumers perceived risk of wasting time. In addition, ease of navigation and consistency were mentioned as important.

Content design

The informational content available on the website were also shown to affect the perceived trust. Consumers should, from images and text displayed, be able to understand the purpose of the website.

Social-cue design

In this feature it is stated that an interface can appeal to the social-cues of the consumer by the nature of images, text, and ability to direct contact, for example an online chat function, in order to create a more personal experience for the consumer.

In a study about trust in cyber-banking-system from the article Designing towards emotional usability in customer interfaces-trustworthiness of cyber-banking system interfaces [27], some conclusions that could be related to online retailers were drawn. The authors claim that main clipart and the main colour layout of the web application has the most significant influence over the perceived trust by the user . A web application without clipart resulted, for a customer, in a growing feeling of unreliability. The conclusion about the colours was that a cold faint pastel colour infuses the most trust for the user.

In a study analysing the connections between colours and words, Hynes discovered that the colour blue was associated with words like trust, reliable, stable, and protective [28]. Another colour associated to these adjectives was brown. The study was investigating colours connected to corporate logos and the corporation’s description.

In the article An experimental investigation of the influence of website emotional design features on trust in unfamiliar online vendors [29], it is found that perceived visual appeal significantly influences both perceived ease of use and trust. Worth noticing from this study is that in the male subject model both the visual appeal and ease of use had significant influence, however, in the female subject model the visual appeal was the only significant factor.

According to the article Assessing dimensions of perceived visual aesthetics of web sites [30], the aesthetic of a website is divided into two different factors. The classical aesthetics corresponds to visual clarity and represent what traditional aesthetics is, it is dependent on these design attributes: Aesthetic, pleasant, clean,

(16)

11

clear, and symmetrical. The expressive aesthetics are focused on the following attributes: creative, using special effects, original, sophisticated, and fascinating. According to the authors, a good design should strive to balance these two factors. The correlation between the classical aesthetics and the expressive aesthetics reflects the visual beauty that they have in common, and e ven though they have a strong correlation they are still distinct.

Lavie et al. [30] found that both aesthetics factors had a significant correlation with pleasure and were at about the same level. Furthermore, the authors claim that usability and classical aesthetic have a significantly larger correlation than usability and expressive aesthetic. In the article a linkage between usability and aesthetics is proposed due to the classical aesthetics’ being strongly related to perceived usability.

3.3.6 The importance of trust in an online setting

In the article Exploring Factors that affect usefulness, ease of use, trust and pu rchase intention in the online environment [10] the authors found that a lack of trust heavily affected the customers intention to purchase. The authors claimed that in regular stores the most trust influencing factor is the salesperson’s likeability and experience but in an online setting the salesperson is replaced by functionalities such as frequently asked questions and search functionality. In the results of the article they noted that perceived service was a major factor affecting trust and that low trust had a significant negative impact on intention to purchase.

3.4 Method Theory

To complete the project, a variety of methods were used. Prototype development, evaluation and measuring methods, testing methods and market analysis tools were all used at different stages of the project. The theoretical frameworks for these are described in this section.

3.4.1 Prototype

In the article High-Fidelity or Low-Fidelity, Paper or Computer? Choosing Attributes when Testing Web Prototypes [31] a prototype is defined as “a working model built to develop and test design ideas”. The authors describe that the purpose of a prototype is to examine content, aesthetics, and interaction techniques from different users such as designers and clients. They can be made using more than one technique, moving closer to the final product as the design progresses to completion. The authors explain that the benefits of using a prototype is to discover and fix usability problems at an early stage of the design, before substantial resources are invested. Prototypes more similar to the final product are called “high-fidelity” and those with less similarities “low-fidelity”. The former has the same interaction techniques and appearance as the final product but is more expensive and time-consuming compared to a low-fidelity prototype which on the other hand have big advantages in terms of cost and ease of iteration [31].

3.4.2 Evaluating usability

ISO’s definition of usability [8] implies that to be able to properly evaluate usability there needs to be metrics that measure effectiveness, efficiency, and satisfaction.

3.4.3 Measuring effectiveness and efficiency

Based on the previously mentioned definition of effectiveness the following formula was constructed. Equation 1: Measure effectiveness

𝐸𝑓𝑓𝑒𝑐𝑡𝑖𝑣𝑒𝑛𝑒𝑠𝑠 =𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑡𝑎𝑠𝑘𝑠 𝑐𝑜𝑚𝑝𝑙𝑒𝑡𝑒𝑑 𝑠𝑢𝑐𝑐𝑒𝑠𝑠𝑓𝑢𝑙𝑙𝑦

𝑇𝑜𝑡𝑎𝑙 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑡𝑎𝑠𝑘𝑠 𝑢𝑛𝑑𝑒𝑟𝑡𝑎𝑘𝑒𝑛 × 100%

(17)

12

Efficiency is the second factor in usability. According to the article Current practice in measuring usability: Challenges to usability studies and research [32] measuring efficiency by measuring task completion time is the most common way of estimating efficiency. On the basis of this article and the defintion of efficiency the following formula was constructed to measure efficiency:

Equation 2: Measure time based efficiency

𝑇𝑖𝑚𝑒 𝐵𝑎𝑠𝑒𝑑 𝐸𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑐𝑦 = ∑ ∑ 𝑛𝑖𝑗 𝑡𝑖𝑗 𝑁 𝑖=1 𝑅 𝑗=1 𝑁𝑅 Where:

N = The total number of tasks (goals) R = The number of users

nij = The result of task i by user j; if the user successfully completes the task, then nij=1, if not, then nij=0

tij = The time spent by user j to complete task i. If the task is not succesfully completed, then time is measured until the moment user quits the task

3.4.4 Measuring satisfaction

According to the article SUS – A quick and dirty usability scale [33] System Usability Scale (SUS) is a simple scale that is used after a respondent has used the system that is being measured. This measurement is done by having the respondent take a stand on ten different statements concerning the system:

1. I think that I would like to use this system frequently 2. I found the system unnecessarily complex

3. I thought the system was easy to use

4. I think that I would need the support of a technical person to be able to use this system 5. I found the various functions in this system were well integrated

6. I thought there was too much inconsistency in this system

7. I would imagine that most people would learn to use this system very quickly 8. I found the system very cumbersome to use

9. I felt very confident using the system

10. I needed to learn a lot of things before I could get going with this system

The article states that the statements are constructed according to the Likert Scale where participants will answer on a scale one to five, where one means “strongly disagree” and five means “strongly agree”. This scale will be used to measure satisfaction in this report. When the questionnaires are collected after each iteration test a total SUS score for that specific test is finalized by [33]:

• subtracting one from each odd numbered question.

• subtract the score from five for each even numbered question.

• summarize the calculated scores and multiply by 2.5 to receive the final result. This gives a SUS result between 0-100 [33].

A study on mean SUS results when observing different applications that included over 3000 applications showed that the average SUS result was 70 [34]. The study also made an adjective scale that they found corresponded well with the numerical scale.

(18)

13

Figure 3: SUS scale [34].

3.4.5 Measuring trust

A common way to measure trust today rely on the Likert scale [35] [36] [37]. The problem with this is that a questionnaire using the Likert scaling must be adapted to fit the web application it is testing on. Therefore, the article Measuring Trust: The Simpler the Better? [38] sought to find a generalization of a questionnaire for measuring trust.

Table 1:Adjectives allocated to the three different dimensions of trust

Benevolence Integrity Competence

Benevolent Malicious Honest Dishonest Competent Incompetent

Cordial Rude Sincere Insincere Skillful Unskilled

Caring Uncaring Trustworthy Untrustworthy Knowledgeable Clueless

The likert scale let the subjects rate how well a statement describes their perception of the web application in question from “strongly agree” to “strongly disagree” on a five-point scale [38]. The authors of Measuring Trust: The Simpler the Better? [38] found nine adjectives allocated to three identified dimensions of trust and their antonyms which was put in the edges of the scaling. This way, the results can be compared across different web applications of different kinds, since the questions are not application-specific.

A value from five to one, five representing the positive adjective in the pair, is given by the test participant [38]. After all the adjective pairs are given values, the mean value of all pairs is calculated to be the overall trust score. The convergent validity of this scale and the Likert scale was found to be sufficient [38]. A higher trust score indicates a higher perceived trust by the test participants.

3.4.6 Test groups

To determine the number of subjects in a test group, the authors of the article The role played by perceived usability, satisfaction and consumer trust on web application loyalty [39] claim that a formula for the proportion of unique problems found can be used.

Equation 3: Proportion of unique problems found

(19)

14

Where p is the mean problem discovery rate, and n is the number of subjects.

Nielsen et al. [40] found that p was approximately 0.3 by calculating the average of p from a large number of independent reports.

In the article Beyond the five user assumption: Benefits of increased sample sizes in usability testing [41] it is concluded that there is a great variance in the amount of errors found when using five test participants. The standard variance when testing on five subjects is 9.3 but is greatly diminished when increasing the amount of test participants. The article Streamlining the Design Process: Running Fewer Subjects [42] concludes that approximately 80% of usability problems are detected with four to five subjects. Additional test participants are less likely to reveal new information.

According to Ungvarsky [43], subjects tend to perform tasks differently when observed. This phenomenon is referred to as the observer expectancy effect by the author. In an environment where the user is being monitored, a cognitive bias of the user can be generated, that can create increased focus or negligence. It is also a possibility that the test subject tries to please the observer, thus generating bad data [43].

3.4.7 Usability methods in agile development

According to the authors of Usability engineering methods for software developers [44], changes to the design can be hard to implement and turn out costly, if testing or usability methods are not applied until the end of the design process. Holzinger [44] explains that usability methods can be divided into two general categories: inspection and testing.

Inspection methods makes comparisons between an interface and already established standards, to find deviations and potential usability issues [44]. The author explains that the most common such method is called heuristic method which involves having usability specialists making independent evaluations of the interface and whether it follows usability standards.

Test methods puts the emphasis on the end-users and tries to determine problems with regards to their interactions with the system [44]. The author claims that a common such method is thinking aloud, this method has users interacting with the system and expressing thoughts verbally while doing so. It creates an understanding for how the system is perceived and understood by intended end -users, thus revealing potential usability issues. Holzinger [44] further explains that other common test methods are questionnaires and interviews. An advantage of using such methods is that they make studying more subjective user experiences of the user possible, such experiences that otherwise can be hard to objectively measure. Holzinger [44] mentions in his article that:

“Testing with end users is the most fundamental usability method and is in some sense indispensable.”

(20)

15

4 Methodology

The methodology section describes extensively how the theories lifted in section 3.4 were used to help develop the application and obtain relevant information. The study was executed using an iterative approach in order to be able to systematically record the impacts of changes in a controlled environment. This also led to the possibility of incorporating test results into the development in order to further enhance the usability of the application.

4.1 Development environment

Python version 3.8 was the programming-language of choice for the backend of the web application together with the web framework Flask, version 1.1.1, for the tools, libraries and technologies required to create a web application. A database developed in Flask was used to store all the data needed such as; user information, product selection and the orders. The orders include previous and current orders, but not the client’s shopping basket since this is stored using cookies. For the front-end HTML, JavaScript and CSS was used with the frameworks Bootstrap, version 4.4.1, and jQuery, version 3.4.1, to implement the web design.

4.2 Application development phase

During the application development, the focus was to develop and test a functioning e-commerce application selling snus. This phase was divided into three iterations, all with planning, development and testing phases.

4.2.1 Iterations

The development process was split into three iterations, each starting with planning, followed by development and concluded with testing. The first iteration’s planning phase was based on the marketing plan, found in appendix 9.8 (Marketing plan), and theory how to enhance usability. This resulted in a prototype that was used as a foundation for the web application. Furthermore, a back-end architecture model was constructed, according to section 4.2.3, in order to enable initial user functionalities. The later iterations’ planning phase was also based the theories for enhancing usability, but also on test results from previous iterations, and were applied to the Adobe XD prototype before being implemented into the actual application. Changes to the database were implemented directly into the application during the development phase instead of updating the backend architecture model. Each iteration was concluded by a testing phase to determine the effects of the changes made during the iteration. Iteration three’s development phase was concluded with extensive bug testing.

4.2.2 Prototype

A prototype was created to get a guideline on how the visual part of the e -shop should be designed. This prototype was primarily based on the theory of how a web application should be designed to enhance usability according to the theories lifted by Hynes [26], Wang and Emurian [28] and Nielsen [19]. It was also inspired by the common trends in both snus e-shops and e-shops in general.

The program used for the prototype was Adobe XD in order to rapidly create and iterate different layouts so an internal discussion could take place resulting in a final design. This was made before the coding of the web application had begun since changes in this stage would be made with a fraction of the costs compared to later in the project according to Walker et al. [31]. The first draft was primarily based on existing literature and consisted of a start page, a product catalogue and a log-in screen. The second draft consisted of the same functionality but with a theme change to conform to the existing theory and make it more presentable. A heuristic evaluation of the usability was conducted by an art director who is a web development specialist at a digital agency based in Stockholm, in accordance with Holzinger [44]. The prototype was sent to the specialist who performed a visual inspection to verify if the prototype followed

(21)

16

their company standards for how a website should be designed with regards to usability. The specialist listed deviations from the company standards in a pdf that was returned after the evaluation. The evaluation was used as basis for further improvement. The results from the evaluation can be found in appendix 9.6 (Heuristic evaluation).

4.2.3 Back-end planning

To ensure that data relevant to the application was stored effectively, a mapping over the necessary data was made. This mapping was then used to create an Enhanced Entity Relationship (EER)-model of the database. The EER-model was then mapped into a Relational Model (RM). The Relational Model was extended by including the datatypes of the different attributes of each entity.

Routes that enabled user functionalities to connect to the database were specified by stating which HTTP methods for RESTful API (POST, PUT, GET, DELETE) were needed and what they should perform when called upon, for each route.

4.2.4 Testing

At the end of each iteration, the application was tested and evaluated by test groups. How this was done is described below.

Test groups

Test groups were formed according to Turner et al. [39], with p = 0.3 and n = 5 giving the result proportion of unique problems found ≈ 83%. As stated by Virzi [42], using more than five subjects would with a high probability not reveal much additional information since the first five subjects are probable to find 83% of usability problems. To clarify, every test group consisted of five participants with expected proportion of unique problems found = 83%.

The test results of the different iterations of the project were compared in order to draw conclusions about the continuously updated usability and its potential effect on trust. For every test session, new groups of five participants were assembled in order to eliminate the learnability factor. When testing, each participant in the test group was given tasks to complete according to appendix 9.3 (Testing Tasks Form). Thinking aloud method was used during the testing session with a test leader, physically present, monitoring and taking notes. The participants’ completion rate and completion time was measured while completing the tasks. When the thinking aloud method part of the test was completed, the participant was given a questionnaire found in appendix 9.2 (Iteration Test Questionnaire) with questions regarding their satisfaction and perceived trust.

Questionnaire

The questionnaire found in appendix 9.2 (Iteration Test Questionnaire) was based on the theory of measuring trust presented by Reiser and Bernhard [38] and the theory of the SUS-scale presented by Brooke [33]. The same questionnaire was used for all prototypes in order to see trends in the answers with regards to the changes from previous testing iterations.

Using test results to enhance usability

After the testing was completed, SUS and trust scores were calculated from the values collected through the questionnaire while effectiveness and efficiency of the tasks in the task list were calculated from the completion rate and time recorded during the testing as described by Hornbaek [32]. The calculated SUS and trust scores were then converted into averages representing the iteration in order to be able to compare the results from each iteration to see if the improved usability had any effect on the perceived trust.

(22)

17

Since there was no definition of what is a high or a low trust score using the scale described by Bangor & Miller [34], only the differences in results between the iterations were used to evaluate the change in perceived trust.

From each test participant, comments from thinking aloud were gathered and these comments were compared in order to find any critical or reoccurring problems that the test participants encountered while testing the web application. If two or more test participants had the same opinion about a specific part of the web application, this was enhanced based on their comments so that the next group of test participants would not experience the same problem.

The comments as well as the ratings from the questionnaire were the foundation upon which the continued development of the web application was based on.

4.3 Concluding phase

When all testing had been completed at the end of iteration 3, a concluding phase was initiated. In the concluding phase the results from the different iterations were compared and analysed with support from the theory presented, and conclusions were drawn on this basis. The method utilized was also analysed to try to gain some understanding of how the chosen method may have impacted what results were achieved along with the conclusions drawn, and how things potentially could have been done differently. This was all concluded with an attempt to answer the research question at a reasonable level of certainty.

(23)

18

5 Results

This section presents the results of the completed iterations. The results of an iteration can be broken down into the results of the product development and the results from the user tests. The results from the user tests are presented through the usability and trust scales, and comments recorded during the tests were used as a basis for further development.

5.1 Iteration 1

The first iteration resulted in a primary version of the web application which is presented below. The results from the performed user tests are also presented below in terms of usability and trust measurements. The test participants during this iteration were between the ages 20-24 and used the internet on a daily basis.

5.1.1 Development

A prototype was developed, and the result can be seen in appendix 9.5 (Initial prototype). A heuristic evaluation of the usability of the prototype was conducted. The results of this evaluation be found in appendix 9.6 (Heuristic evaluation) and were taken into consideration when further developing the prototype. The final prototype can be found in appendix 9.7 (Final prototype).

The focus of the first iteration was to get a sufficiently working web application apart from some features such as payment API and product reviews. Order history was also not yet implemented. The goal was to include only the most important f eatures to have a foundation for the complete build of the e-shop. This includes the index page, product page, splash-screen, about-us, FAQ, contact, login, register and checkout. A navigation bar was used for easier navigation between the pages. This was fixed in the top of the web application and included a search bar that still was not working.

Figure 4: Product catalogue page during Iteration 1.

Splash-screen

Upon entering the e-shop for the first time, the user is required to confirm their age. This is done with a page consisting of the Dosan logo, a header 18 år eller äldre? (18 years of age or older?), information about why this question is being asked, a green symbol and a red symbol. The green symbol would send the user to the web application, whereas the red symbol directs the user to a page informing them that Dosan only

(24)

19

is available to consumers older than 18. Beneath the symbols, in a smaller font, is information informing the user that this web application uses cookies.

Navigation bar

All the pages on the web application inherited a navigation bar. The navigation bar can redirect to the following pages in order: index page, product catalogue, about-us, FAQ, contact-us, register, login and the checkout page. Also, a search functionality was implemented in the navigation bar which meant that users could search for a specific product.

Index page

The index page had a banner slideshow showing different offers that Dosan had. It cycles between three pictures with a timer of five seconds. Three big circles below indicated for the user which banner was currently showing. In addition to the time-cycled slideshow, users could utilize a button to alter between the same offers. However, this caused a bug which caused the banner to act unproperly. Below the slider, information about the negative consequences of using Tobacco was displayed for legal reasons. Lastly, a few ads could be found at the bottom of the page with information regarding the offering of free shipping for purchases exceeding a sum of 799 kr. In order to enhance usability, this page was deliberately designed in a minimalistic way which was in line with theory lifted by Nielsen [19].

Product catalogue page

There existed a filter bar at the top of the product catalogue page which could be utilized ungainly since an additional click had to be done after choosing what criteria you want to sort on, the Filtrera (filter) button. Also, as a consequence of the wide product range, the filter dropdowns covered a part of the catalogue. The first iteration of the product cards included two buttons for purchasing a product; one called Köp dosa (buy single) for buying a single product and one called Köp stock (buy bundle) for buying a pack of ten units. There were different colours on the fonts for the price, indicating that it should be a discount when buying ten at a time. However, the price was the same.

Product view page

All the products on the product catalogue page redirected to their specific product view page where more information could be held. The product view page displays a specific product by showing a picture of the product alongside a descriptive text. Users could change the desired quantity of the product in either single units or bundles of ten, where methods had their respective input box. Change of quantity could be made either by increasing or decreasing by single units or by entering a quantity through the keyboard into the respective input boxes. Further down, there existed a dysfunctional button called Lägg till i varukorg (add to basket).

About-us and FAQ pages

These pages were made with the purpose to enable the customer to find information about the company Dosan and thus feel safe to buy from the e-shop. About-us consisted of the logo along with the headers Vilka är vi? (who are we?), with information about the company, and Om snus från ett hälsoperspektiv (about snus from a health perspective), discussing the health perspectives of tobacco products. FAQ consisted of a couple of frequently asked questions. At this point there was no styling applied to these pages.

Contact page

This page contained contact information to Dosan with e-mail and phone number as well as a contact form. Above the form there was text implying that all sorts of feedback were appreciated, even complaints.

(25)

20

Login and Register pages

The login and register page had identical styling, with the input boxes telling them apart. A high-resolution picture of Snus covered the left side of the page, and input form could be seen on the right side. The form on the registration page included input boxes requiring name, email, address, city, postal code and a desired password. The form on the login page included two input boxes requiring email and password. Additionally, the login page also had an “remember me” box, which was not yet working.

Checkout page

The checkout page was divided into three sections where the first started from the header Din Varukorg (your basket). Here, the products currently in the user’s basket are listed in rows with names, amount and price. In the middle of the row, there are clickable plus and minus symbols to change the amount of a specific product. There is also a trash can symbol to the far right if the user wants to completely remove the product. Below the rows, the total price of the entire basket was listed. However, this function had bugs since the total price did not initially display the correct sum. The second section, Leveransinformation (shipping information), contained input forms for the recipient’s information. In the last section, the buyer had two options for the shipping. Either through a normal letter service which takes 2-4 days, but is free, or purchase the express delivery for 40 SEK. The express delivery only takes 1-2 days.

(26)

21

5.1.2 Concurrent Thinking Aloud Protocol

This section will summarize the comments of the test participants during th e first iteration. Some of the tasks tested functionality that was not yet implemented, these comments have not been taken into account.

Table 2: Comments by test participants during iteration 1

Participant Design Functionality Miscellaneous

A - Desires to be able to

choose filter by clicking on text as well as checkboxes. - Struggles to understand if products have been added

to the basket. B - Colour scheme on

button for ‘save changes’ on profile information gives the

impression of being disabled. - Banner on home page

appears inconsistent in terms of colour scheme

and font size.

- Struggles to understand if products have been added

to the basket. - Struggles to purchase several products as the page

scrolls to the top when clicking ‘add to basket’.

- Noticed a bug with the total sum of the basket.

- Would prefer if products were shown in

alphabetical order.

C - Found several of the buttons and checkboxes

too small. - Found the layout for

‘FAQ’ page to be disorganized.

to the basket. - Desires the ability to sort the products according to a criterium. The test user had to search through all the products to find a specific

one.

- Felt confident about the meaning and functions of

icons used.

D - Desires hover

functions on buttons. - Found some buttons

too big.

- Found the layout for ‘FAQ’ page to be

disorganized.

- Desires to be able to choose filter by clicking on text as well as checkboxes. - Desires to write quantity instead of having to increase

by one

- Desires automatic log in on registration. - Desires to be able to press start page banners and be redirected to the products.

- Did not find ‘FAQ’. - Noted that there were no password restraints.

- Did not find order history.

E - Noted inconsistency in font family between

pages.

to the basket.

- Overall lacking feedback on multiple functionalities.

- Could not find the specific product requested

(27)

22

5.1.3 System Usability Scale

The average SUS score of the primary version of the web application was 74. According to the SUS scale, this score qualifies as ‘good’. The standard deviation of the participants different SUS scores was 6.5. This yields in a coefficient of variation of 8%. The average rating of each question in the questionnaire was taken into consideration for the next iteration.

Table 3: Average question rating SUS questionnaire iteration 1.

Question Average rating

I think that I would like to use this system frequently 2.0

I found the system unnecessarily complex 1.2

I thought the system was easy to use 4.2

I think that I would need the support of a technical person to be able to use this system 1.0 I found the various functions in this system were well integrated 3.0 I thought there was too much inconsistency in this system 2.0 I would imagine that most people would learn to use this system very quickly 4.8

I found the system very awkward to use 2.8

I felt very confident using the system 4.2

I needed to learn a lot of things before I could get going with the system 1.6

Table 4: Average SUS Score iteration 1.

Participant SUS score

A 75.0

B 65.0

C 70.0

D 80.0

E 80.0

Average SUS score 74.0

Standard deviation 6.5

Coefficient of variation 0.08

5.1.4 Efficiency and Effectiveness

During the first iteration participators of the user tests failed to complete some of the tasks, resulting in an effectiveness of 79%. This was mainly because some functions were not yet fully implemented, such as reviews of the products nor the order history. The other measurement used is efficiency. This shows that the participants were able to successfully complete 0.07 tasks per second with a standard deviation of 0.34.

5.1.5 Trust Score

The average trust score of the web application was 3.76 during the first iteration. The standard deviation of the participants different scores was 0.40 which results in a coefficient of variation at 9%, a value indicating low differences among the scores. The trust scores were calculated by using the average rating of all questions in the trust questionnaire for each participant.

(28)

23

Table 5: Average question rating Trust questionnaire iteration 1.

Question: “Which adjective best matches with your experience of the web application?” Average rating

Benevolent – Malicious 4.2 Cordial – Rude 4.2 Caring – Uncaring 4.0 Honest – Dishonest 4.6 Sincere – Insincere 3.6 Trustworthy - Untrustworthy 3.0 Competent – Incompetent 3.2 Skilful – Unskilled 3.4 Knowledgeable – Clueless 3.6

Table 6: Average Trust Score iteration 1.

Participant Trust score

A 3.67

B 3.22

C 4.11

D 3.67

E 4.11

Average Trust score 3.76

Standard deviation 0.40

Coefficient of variation 0.09

5.2 Iteration 2

The results of the second iteration are presented below. This iteration resulted in a further developed web application with regards to the previous web application’s testing results. The test participants during this iteration were between the ages 20-29 and used the internet on a daily basis.

5.2.1 Development

Since most of the web application’s necessary functionality was implemented during the first iteration the majority of the changes made during this iteration focused on the usability aspects of the web application. This included changes in most of the pages of the web shop such as adding hovering functions to all buttons and changing different font-families used to a single font-family. However, some functionalities were added in line with the user tests from the previous iteration. This includes implementations such as addition of footer, form validation during checkout, prevention of a checkout with an empty basket et cetera.

Index page

There were no major changes that were implemented on the index page. However, it was noted during the user tests from the previous iteration that the banner slideshow which was displaying different deals was not functioning properly. The slideshow was altering between different offers too fast and inconsistently.

(29)

24

This bug was fixed during this iteration. Moreover, a click functionality was added to the banner so that the user would be taken to the specific product in the offer upon a click.

Product catalogue page

The product catalogue page was mainly improved by added functionality. The filter was improved by allowing users to filter by clicking on text as well as the checkbox next to the text. Furthermore, a sorting functionality of the products was added which allows users to view products according to the chosen sort method. Products can be sorted by popularity, price, brand, and name. During the user tests it was found that several of the participants desired feedback when attempting to add a product to their basket. In order to accommodate this, a simple message was displayed underneath the basket icon informing the user when a product had been added to the basket. Moreover, users struggled to add several products to the basket as the web page scrolled to the top once a product was added to the basket. This scroll functionality was removed in line with the participant’s desires. Also, a zoom functionality was added which resulted in larger icons and product pictures when hovering over them.

Figure 5: Picture of the product catalogue during iteration 2.

Product view page

The product view page underwent minor changes during the second iteration. The option of adding a bundle to the basket was removed. Moreover, an additional picture was added for each product meaning that the product view page displayed two pictures of a product in a slideshow.

About-us and FAQ

During the second iteration pictures were added to these pages. The picture on the about-us page presented Dosan’s history and the prediction of Dosan’s future timeline. These pictures provided complementary information to the about-us page and the FAQ page respectively. The structure on the FAQ page was changed by hiding the answers to the different questions. Instead all questions were represented as large buttons which displayed the answer to the question once clicked.

(30)

25

Contact page

The layout of the contact page was changed by separating the contact form and the text above. Also, a picture of a person in a call was added to the page.

Registration- and login page

The registration page was improved by automatically logging in the user upon completed registration. Also, responsive alert banners were added to the forms which reacted on wrong user inputs.

Checkout

The checkout page was developed by preventing a checkout with an empty basket. This was done by changing the content of the checkout page and thus preventing a user from checking out if the basket was empty. Also, a validation functionality was added to the checkout form which prevented checkouts with empty or incorrectly filled input fields. The user test results implied that the buttons and checkboxes on the checkout page were too small and thus they were enlarged during this iteration.

(31)

26

Navigation bar and footer

A hover functionality was added to the icons on the navigation bar.

During this iteration a footer was added to the origin template which meant that all other pages inherited the footer. The footer contained information about Dosan and redirected users to different service pages such as contact us. All the icons on the footer reacted on hover by changing colour and increasing in size.

Does Enhanced Usability Increase Trust? : A case study on a Swedish snus e-commerce application