Do you have any left?

(1)

HALMSTAD

Digital Design and Innovation, 180 hp

Do you have any left?

Three triggers to persuade into sustainable behaviour within e-grocery

Informatics 15 hp

Halmstad, 2021-06-02

Kayleigh Bishop Källberg, Zephyr Orlando Vasquez

Crabtree

(2)

Design and Innovation – at Halmstad University. We want to thank our supervisor Esbjörn Ebbesson. The commitment to your students has been and continues to be impeccable. At every roadblock we experienced throughout this study, you have provided us with tools to overcome them, as well as providing us with insights that we had never considered ourselves.

It is very apparent to us that you want to help your students with an engagement that we regard as a rarity. Our sincerest thanks to you Esbjörn!

We also want to thank all who participated in this research – whether as respondent, opponent or other supervisor. You helped us make this research possible, for that we are extremely grateful.

Halmstad, 2021

Kayleigh Bishop Källberg

Zephyr Orlando Vasquez Crabtree

(3)

Abstract

The concern about the rising food waste has drawn the attention of policy makers around the world, from France banning supermarkets throwing away unused food, to a more global perspective, with the UN aiming to reduce food waste by 50% by 2025 (Hinckley, 2018;

Valencia, 2016). As the concern about food waste increases - so does the concern about what role e-grocery will play in it. By reason of the decreasing sense of ownership of the product, it is likely that food waste will increase (Illyuk, 2018). It has been demonstrated that

persuasive design is an efficient tool to influence users’ habits within e-grocery, but currently it is primarily used to influence purchases and subsequently waste more (Chu et al., 2014).

We believed that e-grocery showed potential to also be able to counteract the food waste with the help of persuasive design. This research therefore aimed to investigate if persuasive design could be used for more sustainable purposes; meaning to decrease food waste instead of inciting overconsumption. Using current e-grocery triggers from Chu et al. (2014), this research has repurposed them to pursue a more sustainable approach. The triggers (default personalisation, reminder, suggestion & expertise-credibility) performed within a design- oriented research. The aim was to use the prototype in affiliation with criteria, that would prove their effectiveness through critical perspectives of the literature.

A study was set up with 12 respondents, in which they were asked to navigate the prototype with tasks. After navigating the prototype, they were asked questions based of the criteria.

Analysing and evaluating the empiric data with the criteria helped to conclude three triggers that could be useful within e-grocery to reduce food waste.

Keywords: E-grocery, Persuasive design, Triggers, Food waste, Ethics

(4)

1 Introduction

Food waste is on the rise. It is estimated that one third of all the food produced globally is wasted (Lim et al., 2017). This estimation only measures food that is meant for human consumption (Gustavsson et al., 2011). Although studies can distinguish that food waste is not isolated to one individual actor in the supply chain - it also shows that significant losses lie within the hands of consumers in domestic settings (Illyuk, 2018). The concern about the rise of food waste has drawn the attention of policy makers from around the world, such as in France where supermarkets have banned discarding unused food, to a more global

perspective from the UN, aiming to reduce food waste 50% by 2025 (Hinckley, 2018;

Valencia, 2016). At the same time as food waste increases - so does online grocery shopping (also known as e-grocery). In the U.S., a total of 3% of the population currently uses e- commerce to shop for groceries; a number that is estimated to rise to by 13% annually (Illyuk, 2018). Within the E.U., about 25% of customers now buy their groceries online, a number that is estimated to rise to 70% by 2029 (Saskia et al., 2016).

E-grocery refers to the sale or purchase of groceries being transacted online instead of visiting the traditional brick-and-mortar stores (Buldeo Rai, 2021). The customers can have the choice of either home delivery or pick-up which makes e-grocery more accessible to the customer (Pujani & Nazir, 2010; Saskia et al., 2016). Research has shown when customers that use e-grocery services leads to a reduction of their perception of ownership. The reduction of ownership stems from the customer uses less of an effort when purchasing online than instore (Illyuk, 2018). For example, when a customer goes to a traditional store to buy onions, they put in more of an effort to pick the one they deem perfect. When this is not possible to do in an online setting, customers feel less responsibility of the products outcome which can then lead to it becoming food waste (Illyuk, 2018). National Academies of

Sciences, Engineering, and Medicine [NASEM] (2020) suggest that interpersonal factors (such as income, attitudes, knowledge, and relationships) effect individuals' attitude towards handling of food. Moreover, NASEM (2020) suggests that the use of persuasive design [PD], with affiliated triggers, can be helpful as a strategy to lower consumer food waste.

Persuasive design is a strategy that designers can implement into a service to influence human behaviour toward a certain action (Cialdinin,1993). For instance, there are several e- commerce sites that use persuasive design strategies to generate more profits by influencing people to buy more (Chu et al., 2014; IJsselsteijn et al., 2006). These strategies are also known as triggers. There exists an array of different types of triggers that differentiate between each other by how they are implemented and how they influence people (Chu et al., 2014). For example, the reminder trigger is used to draw attention from the user to a certain action, such as registering an account for a service (Fogg, 2003). For our context of food waste, an e-grocery site can use the reminder trigger to enable the user to recall what they have previously bought by telling them “You bought this last week double check if you have any left, so you don’t buy unnecessarily”. With the use of persuasive design triggers, we can influence human behaviour to understand how they can reduce their food waste to meet the UN goal of minimizing that waste.

Even though persuasive design triggers are more commonly used by e-commerce sites to influence their customers to purchase more products (Chu et al., 2014; IJsselsteijn et al., 2006), we are not going in blindly to use these triggers for the opposite effect. We are aware of PD triggers, how they operate and of their limitations. For example, repeatedly reminding a user on how they can be more sustainable can lead to frustration, annoyance or even both (Fogg, 2003). There are principles that the designer can follow to ensure that the persuasion

(7)

is ethical (Berdichevsky & Neuenschwander, 1999; Fogg, 2003). For example, the persuasive strategy must not misinform the user to achieve its goal. Before fully implementing

persuasive triggers during the evaluation stage the designer can use several different criteria to understand their effectiveness (Cabana et al., 2019; Fogg, 2009; Oinas-Kukkonen &

Harjumaa, 2009; Thomas et al., 2019). Currently we do not yet know how successful using the triggers for the opposite effect other than intended is going to be - or what kind of results it may have.

In our study, we will research and try to answer the question “How can persuasive design be useful in e-grocery to reduce food waste?”.

Purpose

PD triggers are more commonly used to influence people buy more (Chu et al., 2014). In this study we used the persuasive triggers for the opposite effect to make customer reflect and buy less. We have two main purposes for this study, first, aim to explore the usefulness of

persuasive design triggers within an e-grocery setting to reduce food waste, secondly, to develop guidelines (presented in a framework with example criteria and implementation) for future designers to follow when implementing triggers for the opposite effect than commonly done. Since PD can be both ethical and unethical, researching with an ethical scope has also become a purpose of this study.

Limitations

We have made three limitations in this research. The first limitation concerns the triggers.

Chu et al. (2014) suggests fifteen triggers that is currently used within e-grocery (see attachment 1). We chose four triggers that had potential and only researched these four further. The chosen four triggers were deemed relevant, ethically appropriate and had the most potential for the opposite effect.

The second limitation concerned our population. The population was limited to individuals over the age of 18 with experience of e-grocery to get ethical and accurate results. The age- limitation was based upon the fact that PD is deemed unethical for underage users.

Lastly the prototype was limited to the use of ICA Maxi e-grocery webpage to counteract bias. Would we have used more than one e-grocery site, likely we would not be able to distinguish if respondents were expressing their experience objectively or not. For instance, imagine if we tested the triggers with both ICA Maxi and Coop. Respondent #1 liked ICA Maxi but respondent #2 preferred Coop. It is safe to assume that respondent #1 would express more interest and in favour for ICA Maxi, and respondent #2 for Coop. Therefore, either we had to implement the triggers into all e-grocery services or limit ourselves to one. We decided to test the triggers within one e-grocery in order to keep the research unbiased.

(8)

2 Literature review

The following section we present related literature that highlights relevant concepts, theories, and challenges within the domain. It will discourse topics such as PD, e-grocery, ethics and how the three relate to one and another.

E-grocery and food waste

The food supply chain highlights consist of five different stages of a products lifespan from the agricultural to consumption stages (Gustavsson et al., 2011). These five stages are agricultural production, post-harvest handling and storage, processing, distribution, and consumption (Gustavsson et al., 2011). Food waste can come at any point of the supply chain but in this paper, we are going to be focusing on the last stage - consumption. Food waste is more commonly attributed to the consumption stage as it produces the largest amount of waste (Parfitt et al., 2010). It should be also noted that, food waste is only measured for products that are meant for human consumption (Gustavsson et al., 2011).

Over the previous decades there have been changes in consumption patterns and behaviours (Buldeo Rai, 2021). These changes have been particularly driven by the digitalization and incremental improvements to the accessibility of digital domains (Buldeo Rai, 2021). One digital domain that has been impacted is e-grocery. E-grocery refers to the sale or purchase of groceries being transacted online (Buldeo Rai, 2021). The U.K. developed and established the service in the domain at the beginning of this decade, making it a starting point for e-grocery in all of the E.U. (Saskia et al., 2016). Subsequently other European countries followed the U.K. in their pursuit of the e-grocery market. Currently, about 25% of customers now buy their groceries online in the UK; a number that is estimated to rise to as much as 70% by 2029 (Saskia et al., 2016).

E-grocery sales are not only rising within the UK. In South Korea, it has been established that almost a fifth of the grocery market now consists of online transactions (Svensk Digital Handel, 2018). In the 2018 research, published by Svensk Digital Handel indicated that one in eighth Swedes considers e-grocery their primary source of purchasing food. A proportion of the Swedes that remain customers of in-store purchasing of groceries explain that this is habitual (Svensk Digital Handel, 2018). Others explain that they like the freedom of being able to choose the groceries themselves, keeping them from purchasing through e-grocery websites. However, e-grocery retailers use digital tools to change behaviours and attract new customers (Svensk Digital Handel, 2018).

Persuasive design

Place yourself in the perspective of the Swedes that have the habit of in-store grocery purchasing. As prior mentioned - these are potential customers who will need persuasion to start using e-grocery (Svensk Digital Handel, 2018). For instance, Svensk Digital Handel (2018) suggest the variables of special offers and discounts that are given to online users can be used to attract more customers. These two variables can convince individuals into

becoming customers – which is exactly the e-grocery retailers’ intent when making offers specifically for online user only (Svensk Digital Handel, 2018). This phenomenon is known as PD and aims to intentionally shape, reinforce or change how individuals think, feel or behave concerning an issue, object or action (Fogg, 2003). These businesses are clearly and intentionally trying to reinforce a new behaviour (shopping groceries online) by persuading (entitling discount or/and offer) so that you think about an action (i.e., your thoughts about e- grocery).

(9)

As many decisions are made in online environments, persuasion does not always occur with intent (Fogg, 2003). To put this in perspective; if the retailer chose instead to give non- customers offers and discounts as a way to counteract food waste, their intent is no longer to recruit. Their intent is to make use of the groceries that would otherwise be wasted. If the retailer happens to recruit in the process, it is unintentional. What has been described is one way that reflect how much persuasion is a part of human life, as well as part of information systems (Fogg et al., 2017). All, information systems have the potential to be considered persuasive (Hamari et al., 2014). What makes PD systems different from other information systems is that it has a predefined intended outcome (Berdichevsky & Neuenschwander, 1999).

Triggers

One way to use PD as part of a predefined intended outcome is by using triggers. This is a method used by designers to get a certain reaction from the user within a product or a service.

The persuasion is made by guiding users into desirable behaviour without force. The triggers ensure that the desirable behaviour is portrayed into the most attractive choice for the user.

This is made possible as available options are designed (the way information is presented, how the interface is structured) in a way that can exert a subconscious influence on the outcome (Weinmann et al., 2016); for example, “what is chosen often depends upon how the choice is presented” (Johnson et al., 2012, p. 488). For instance, let us say that an e-grocery retailer managed to recruit you as a customer. Once you create an account, the information system informs you, the user, that the chosen password lacks strength. Consequently, you are suggested to change to a stronger password. In all likelihood you will change the password, and the trigger has been successful.

There are two types of triggers: internal triggers and external triggers. An internal trigger is something that is manifested automatically within our minds (Winter, 2019). It works because it is connected to a system where our emotions are in charge (Winter, 2019). You can’t hear it, see it or touch it (Winter, 2019). For instance, a growling stomach, when

hungry, is an internal trigger (Fogg, 2009). External triggers are affiliated with external inputs of information. The information tells the user what to do next through persuasion, sometimes transparent relating to desired outcome and sometimes not (Winter, 2019). For instance, special offers and discounts, to recruit new customers, would be an external trigger. External triggers, as already mentioned are a way to direct individuals into a predefined intended outcome.

Evaluating triggers

Fogg (2009) acknowledges that there are three key factors to consider for triggers to be effective; being noticeable, being associated with target behaviour and occurring at the right time. Although Fogg (2009) is not alone in having criteria for triggers to be effective. Thomas et al. (2019) also suggest that triggers need to be motivating, should reinforce reflection and have the potential to change users’ attitudes. Cabana et al. (2019) also suggests that the trigger needs to be shown at the appropriate frequency. To be able to design this criterion and subsequently evaluate them, Oinas-Kukkonen and Harjumaa (2009) suggest putting them into a framework with examples of implementation.

To concretise what the framework could look like, the Swedes will once again be

exemplified. Swedish e-grocery retailers intended to recruit offline customers (the customers that traditionally would make their grocery purchases in-store) by giving offers and discounts (see table 1). The first cell explains that the trigger is Offers and Discounts. Followed by the

(10)

criteria is the example requirement and then implementation, suggesting how and in what context it could be used.

Table 1. Criteria framework.

Trigger Example Criteria Example Implementation

Offers and Discounts System should entitle virtual offers and discounts for users in order to give credit for performing the target behaviour.

Discounts that are entitled to the customers first purchase at an e-grocery site.

Triggers and e-grocery

As previously stated, e-grocery retailers use PD to persuade individuals into becoming customers. If the intended outcome of the persuasion is successful (making you into a

customer) the PD redirects you into a new intended outcome. The retailer now must persuade you to purchase more in order to be considered successful (Chu et al., 2014). Currently, there are multiple ways that the e-grocery retailers persuade customers to purchase more. Take scarcity for instance. Scarcity urges customers to buy products by illustrating that there are limited quantities of a product (Chu et al., 2014; Weinschenk, 2009). It adds value to the trigger as it makes the customer believe that said product will be hard to acquire in the future.

It drives customers to purchase, not out of need, but to avoid the feeling of loss (Caraban et al., 2019). The feeling of fear can for example, make customers pursue unsustainable purchases and overconsume (Caraban et al., 2019; Illyuk, 2018; NASEM, 2020).

As the number of e-grocery sites are expected to multiply in the near future this allows us to introduce persuasive triggers more frequently, thanks to the accessibility of smart devices (Fogg, 2003; Fogg et al., 2007; Illyuk, 2018). This accessibility allows customers to be persuaded at any time and place, not just within store boundaries.

Triggers against food-waste

As mentioned above, triggers are frequently used as a tool to persuade customers into buying and consuming more (Chu et al., 2014; IJsselsteijn et al., 2006). As of now, triggers are not commonly used to achieve the opposite effect; that is, making customers reflect and buy less.

It is evident that triggers do show the potential to be opposite and instead used for sustainable purposes. To determine whether this is the case or not, a predefined number of triggers will be researched.

Below, the chosen triggers for further research are presented. They are selected from the fifteen triggers (see attachment 1) mentioned by Chu et al. (2014). The majority of the fifteen refer to implications of e-grocery interfaces, which is inapplicable for this. Surface credibility for instance, implies that the overall look of a website has an impact on customers confidence for it to be believable and be seen as competent. The trigger example refers to having

uncluttered and professional designs, which operates beyond the point of this research (Chu et al., 2014). We therefore choose to exclude this from our work.

Other alterations were considered ethically inappropriate to reverse into a sustainability trigger, one being social proof. This trigger refers to persuasion, by proclaiming other customers’ line of action within a website. This can be used to show what other customers bought, or to show customer feedback on product pages (Chu et al., 2014). Hence, the trigger

(11)

is based upon customers own opinions. Reversing the trigger into becoming more sustainable would subsequently mean to try to persuade customers into claiming an opinion that is not their own. From an ethical viewpoint it would seem immoral to do such a thing. Ethics are discussed further in 2.5.

Remaining were four triggers which showed potential to be opposite (from having the intention to sell more, to having the intention of being more sustainable). The four triggers were default personalisation, reminder, suggestion, and attractiveness (see attachment 1).

Some were modified as Chu et al. (2014) suggest combinations are more successful than when using each one individually.

Default Personalisation

Implementing triggers on an e-grocery site becomes cheaper and faster compared to those in- store experience as there will be several tools to analyse and track an individual (Berger et al., 2020). The analysed data could be used to personalise the interface in different ways. For instance, implementations could be last viewed products or member preferences. It is the idea of recognition that is implemented. In turn this fosters credibility and has an improved ability to persuade its users into the predefined, intended outcome of the e-grocery site (Fogg, 2003;

Fogg et al., 2007).

The research is combined with the compliance of a default option. This because Mont et al.

(2014) suggest that people choose the passage of least resistance and therefore would comply with the default. For instance, countries where the default of organ donation is consent, participation is greater as opposed to countries where the inhabitants must actively enrol.

Mont et al. (2014) suggest the option of default to be simplified as it helps consumers make informed decisions. This trigger intends to reinforce an act of compliance.

Reminder

Reminder is a trigger that is used for support of certain target behaviour (Chu et al., 2014).

For instance, exposing users to the information as to how much they would have to spend for an upgrade of their membership, thus encouraging the user to buy more (Chu et al., 2014).

For instance, some e-grocery sites use membership points that are provided whenever the customer purchases products. If the customer reaches enough membership points, they are rewarded with a discount – but only if they have purchased enough products. The reminder as a trigger has the intent of altering its user’s behaviour.

Suggestion

Chu et al. (2014) describe suggestion as a persuasive tool to present a message to guide the user to a behaviour. It could, for instance, be a visualization in the users’ shopping cart showing what other people are buying, stating People Who Bought This Item Also Bought…

(Chu et al., 2014, p. 484). Consequently, users are more likely to buy things out of impulse, hence overconsuming (Chu et al., 2014). Furthermore, Fogg (2003) describes suggestion as pervasive tool built upon users pre-existing motivation. For Instance, Oinas-Kukkonen and Harjumaa (2009) note that if the goal is to eat healthier, a system can suggest eating fruit as a snack, rather than candy. This type of trigger has the intent of reinforcing a behaviour.

Expertise-credibility

Expertise-credibility is described by Chu et al. (2014) is a trigger used to draw the user’s attention to subsequently establish a favourable impression for the target outcome. In

(12)

addition, Fogg (2003) explains that certain professions can seem more trustworthy than others. For instance, companies that hire famous spokespersons or models, do it for the credibility – again this kind of trigger has the intent of altering the attitudes of its users.

Ethics of Persuasive design

Ethics within PD is different depending on whom is asked. Some will argue that persuasion has a link to indoctrination, coercion and manipulation, which are considered unethical (Fogg, 2003). Others will point out that persuasion is necessary as a part of sustaining democracy (Fogg, 2003). Here we can see that persuasion can be both ethical and unethical.

Kight and Gram-Hansen (2019) argue that persuasion is therefore an ethical deliberation, making persuasion an ethical matter that is constantly distributed between persuader and persuaded. Furthermore, Fogg (2003) argues that there are different actors taking place in the matter and this has created a framework to distinguish ethical responsibilities. The framework identifies matters such as intent and predictability as two key actors in determining where the ethical responsibility lies (see attachment 2). Using this framework, the designer is invited to think critically about the ethics behind the chosen persuasion (Berdichevsky &

Neuenschwander, 1999; Kight & Gram-Hansen, 2019).

In some cases, PD is considered unexceptionally unethical. For instance, Fogg (2003) suggests that it is unethical to expose PD to vulnerable groups such as children. Despite this, it is estimated that only two percent of children's applications and websites do not contain any triggers, most of them created to be indistinguishable from the website or application itself (Nairn & Dew, 2007). Nairn and Dew (2007) as well as Fogg (2003) argue that this is unethical as it exploits children, suggesting that site owners reconsider their policies.

Considering that the PD is not always intentional, it is hard to know whether site owners are even aware. Although as e-grocery is aimed at the supply of products to adults, this specific ethical issue is irrelevant to this research.

Ethical Principles

In an extension to the ethical framework presented by Fogg (2003), Berdichevsky and Neuenschwander (1999) present eight principles that can be used in PD to ensure that it is ethical (see table 2). These principles can also be regarded as risk factors, meaning the more you violate the rules, ultimately there will exist the greater possibility for unethical behaviour (Berdichevsky & Neuenschwander, 1999). These eight ethical principles were referred to whenever ethical matters were discovered.

(13)

Table 2. Ethical Principles. (Berdichevsky & Neuenschwander, 1999).

Ethical principles

1 The intended outcome of any persuasive technology should never be one that would be deemed unethical if the persuasion were undertaken without the technology or if the outcome occurred independently of persuasion.

5 Persuasive technologies relaying personal information about a user to a third party must be closely scrutinized for privacy concerns.

2 The motivations behind the creation of a persuasive technology should never be such that they would be deemed unethical if they led to more traditional persuasion.

6 The creators of a persuasive technology should disclose their motivations, methods, and intended outcomes, except when such disclosure would significantly undermine an otherwise ethical goal.

3 The creators of a persuasive technology must consider, contend with, and assume responsibility for all reasonably predictable outcomes of its use.

7 Persuasive technologies must not misinform in order to achieve their persuasive end.

4 The creators of a persuasive technology must ensure that it regards the privacy of users with at least as much respect as they regard their own privacy.

8 The Golden Rule of Persuasion

The creators of a persuasive technology should never seek to persuade a person or persons of something they themselves would not consent to be persuaded to do.

(14)

3 Method

In this chapter of the study, we describe our methodology and related strategies for data collection. After describing strategies, the design of the prototype is presented. The

description of the prototype is followed by the strategy to analyse and then evaluate it. Lastly, we describe how the pilot-study was performed. First and foremost, ethics of research are conveyed.

Ethics of Research

Research must be conducted ethically (Vetenskapsrådet, 2017). It is the researcher’s

responsibility to ensure that the practice of ethics is incorporated within the research, so that it is beneficial for both participants as well as collateral participants (Vetenskapsrådet, 2017).

The Swedish research council has developed eight general rules researchers refer to ensure ethical research.

Table 3. General rules of ethical research.

General rules

1 You shall tell the truth about your research.

5 You shall not make unauthorized use of the research results of others.

2 You shall consciously review and report the basic premises of your studies.

6 You shall keep your research organized, for example through documentation and filing.

3 You shall openly account for your methods and results.

7 You shall strive to conduct your research without doing harm to people, animals or the environment.

4 You shall openly account for your commercial interests and other associations.

8 You shall be fair in your judgement of others’ research

Approach

The approach was to use design to orient the research. Design research is a formative method to test and refine educational designs with the help of theoretical principles originating from prior research (Collins et al., 2004). The intention with the design was to make progressive refinements throughout iterations until bugs are worked out, although that is not the exclusive reasoning for using design research. The method also aimed to address theoretical questions as well as issues (Laurel, 2003). The questions and issues were addressed by observing the social and contextual variables interactions with cognitive variables. For example, the overall objective was to observe how the chosen design(s) performs in practice.

Our overall objective was to modify current e-grocery triggers (the ones that usually are implemented to persuade customers into purchasing more) into sustainable performances. We chose four triggers, from the literature by Chu et al., (2014), that we believed to have

potential to opposite from the current ecologically unsustainable use. We wanted to put said triggers into a design that could help us research the theoretical question “How can

persuasive design be useful in e-grocery to reduce food waste?”. By letting respondents (the population of our context) navigate a design where the triggers were implemented, we could observe how they performed in practice. The replies from our respondents would then be analysed and evaluated. The evaluation was conducted by putting the answer in relation to criteria (see table 4) that was collected by multiple sources (all criterion in which were created specifically for PD).

(15)

The research was divided into four steps - literature review, prototype, evaluation and pilot- study. The literature review was performed to obtain a knowledge base. The knowledge base provided us with an understanding of PD, e-grocery, ethics and how they currently operate.

Subsequently, that understanding led to the discovery of the four triggers with the potential to be opposite (default personalisation, reminder, suggestion & expertise credibility). Through literature, we gained an understanding as to how the triggers could be designed and

evaluated. Because of the understanding we developed, the triggers were designed and implemented into the prototype. The prototype was initially tested in a pilot-study to ensure that faults were discovered and adjusted before initiating the evaluation-step of the research.

The evaluation embodied the prototype being tested by respondents for data-collection. The respondents were also asked a number of specific questions (see attachment 3). The empiric data was analysed and re-analysed in relation to criteria (see table 4).

Table 4. Example criteria.

Triggers Example Criteria Reference

All Trigger is noticeable. Fogg (2009)

All Trigger is associated with target behaviour. Fogg (2009)

All Timing is appropriate. Fogg (2009)

All Frequency was appropriate. Cabana et al. (2019)

All The trigger was motivating/engaging. Thomas, Masthoff & Oren (2019), Némery et al. (2011)

All The trigger prompt reflection. Thomas, Masthoff & Oren (2019) NASAM (2020)

All The trigger has potential to change attitudes. Thomas, Masthoff & Oren (2019)

All The trigger was not invasive. Fogg (2003)

All The trigger was important. Thomas, Masthoff & Oren (2019)

Default Personalisation

The trigger was personally relevant. Thomas, Masthoff & Oren (2019)

Default Personalisation

The trigger was personally accurate. Thomas, Masthoff & Oren (2019)

(16)

Literature review

As an initial step we conducted a literature review. The literature review is crucial step to research other knowledgeable sources. The review concerned our subject, methods, and ethics. We read and critically evaluated the literature to immerse ourselves in the topics of PD and e-grocery.

According to rule six of the eight general rules, presented by Vetenskapsrådet (2017), it is important to keep the research organised. In order to do this, we created a framework (see table 5) that illustrated the title and description of each data input that we found valuable for our research. This helped us keep notes, find relevant keywords and other relevant literature referred to (see table 6).

Table 5. Framework for literature review.

Date Title Reference Link Origin Keyword(s) Loss/When

When the article was found

The title of the Article

The reference of the Article

The link to the Article (if there was one)

What database it was collected from (or article)

Keywords presented within the article.

Loss of Article (From

irrelevance etc.) and when.

Table 6. Example of how framework for literature review was used.

Date Title Reference Link Origin Keyword(s) Loss/When

4/2/2021 Taking a look at different types of e- commerc e

Nemat, R.

(2011). Taking a look at different types of e-

commerce.

Citeseerx.ist.p su.edu.

https://citeseer x.ist.psu.edu/v iewdoc/downl oad?doi=10.1.

1.684.6401&r ep=rep1&type

=pdf

https://citesee rx.ist.psu.edu/

viewdoc/dow nload?doi=10.

1.1.684.6401

&rep=rep1&t ype=pdf

Semantic Scholar

e-commerce e-commerce types business

4/2/2021 Majority of author’s referencing in text, originate from Wikipedia.

To find relevant literature for the topic (PD & e-grocery), general keywords such as

Persuasive design, HCI, Sustainability, Food waste and Online grocery shopping were used in several different databases. The primary databases that were used to find relevant literature were Scopus, ScienceDirect and SpringerLink. Google scholar and Microsoft Academic were also used to find articles that were relevant on different databases. Once relevant literature was discovered, it was reviewed by looking at the title, abstract, language, publication date and the literature’s references. When reviewing the literature new key words were found. The new key words that were identified were e-commerce, e-grocery, and persuasive technology.

These new keywords were later used together with the previous keywords to specify our search results, which gave us new literature. As a consequence, when reviewing the literature references, repeat references were looked up to yield substantially more information of PD within HCI. It fostered a sense of how PD was conceptualized and concretised.

With the data from the framework previously created, we were able to put together our PRISMA (see attachment 4). A PRISMA is a way to illustrate reported evidence-based items that is used within the systematic review (Page et al., 2021). In our PRISMA you can see

(17)

what databases we used, how many results we obtained, at what stage in the process which articles were lost and how many we ended up using in our research.

Prototype

The purpose of the prototype was to explore how PD could be used as a tool to combat food waste. A prototype can be used to visualise the ideas that were generated after reviewing the literature. The idea was to put the four chosen triggers (default personalisation, reminder, suggestion & expertise-credibility) into a prototype. The fundamental, visual representation of the triggers was designed according to the examples Chu et al., (2014) presented of the triggers. For instance, the trigger reminder is described:

was used to remind user of or call user’s attention to certain target behavior.

For example, when a user login, the system will display how much the user should spend more to get a membership upgrade. Another example was to repeatedly display certain information, such as a new service announcement or the information about a marketing campaign, on the screen to catch users’

attention“ (Chu et al., 2014, p. 484).

The trigger was therefore designed to remind the users by indicating what they had bought previous in the session followed by the question “Do you have any left?”.

A web-based prototype was used as a representation of the triggers. ICA Maxi was chosen as the representative e-grocery website as a limitation. The ICA Maxi interface was

incorporated into a visual prototyping tool (the program Figma was used) with the triggers of interest. The triggers were designed in alignment with ICA Maxi’s current interface.

Deliberately all elements were created with the same fonts, colour, and shading. The triggers were placed within a directory that were assigned to a scenario. What specific pages the triggers were placed within the directory was based upon what was appropriate to the context.

For instance, the trigger reminder was placed in affiliation to where the customer would choose the quantity of the product.

After implementing the triggers, the strategy was to validate and make changes accordingly.

That way, potential necessary changes could be implemented easily and again evaluated. The prototype was improved one time.

Evaluation Population

Our population resides within the consumption part of the food supply chain (Gustavsson et al., 2011) (see figure 1). The population consisted of individuals over the age of 18, with experience with e-grocery. Partly because e-grocery customers must be of age and partly because PD is considered unethical for underage individuals (Fogg, 2003; Nairn & Dew, 2007).

Figure 1. Illustration of the food supply chain.

Agricultural Production

Post-harvest handling and

storage Processing Distribution Consumption

(18)

Users’ personal experiences have an impact on information processing (Oinas-Kukkonen &

Harjumaa, 2009). It was therefore important to ensure respondents interest and experience within the topic. The population was set up in two iterations – pilot study and design study.

The individuals chosen for the pilot-study were chosen because we knew they had experience with e-grocery. We called them respondent X and Y (see table 7). Variables such as age and gender were taken into consideration as these things generally have an impact on the outcome of the research (Trost & Hultåker, 2012).

Table 7. Description of pilot-study population.

Respondent Age Gender Time Comment

X 20-29 Man 39:00

Y 20-29 Woman 31:00

For the design study, we retrieved contact information from individuals that had

communicated their information in a survey about e-grocery and food waste (see table 8).

The respondents were contacted by e-mail, stating its purpose and asking if they were

interested to be participants in the study. Out of these individuals, we chose ten individuals to represent the population. Once again, variables such as age and gender were taken into consideration as these things generally have an impact on the outcome of the research (Trost

& Hultåker, 2012).

Table 8. Description of the research population.

Respondent Age Gender Time Comment

1 20-29 Woman 17:49

2 30-39 Woman 17:05

3 30-39 Woman 26:24

4 40-49 Woman 18:47

5 50+ Woman 20:24

6 50+ Woman 17:49

7 20-29 Man 17:34

8 20-29 Man 58:12

9 20-29 Man 18:37

10 20-29 Man 26:32

11 50+ Woman 09:29 Non-Completion

12 50+ Woman 25:47 Non-Completion

Two respondents (respondent #11 & # 12) ended up being non-completion. Respondent #11 had misinterpreted the objective and had not purchased food through e-grocery, whilst respondent #12 failed to understand the tasks and how to navigate the prototype according to the scenario they were provided. Therefore, the data retrieved from these respondents was excluded from the research.

As age and names of the respondents were initially used, respondents were assured their personal information was going to be anonymized through de-identifying, an ethical

intervention suggested by Vetenskapsrådet (2017). As every precaution needs to be taken in order to ensure respondents integrity, all videos were transcribed with names transduced into numbers and subsequently deleted to cut ties with all personal information.

(19)

Data collection

Our data collection started with a pilot study. A pilot study is conducted the same way as the design research eventually would be (with respondents navigating a prototype, analysis and evaluation). The pilot study was conducted to observe the social and contextual variables interaction with cognitive variables – i.e., how one trigger design did in relation to the criteria (see table 8). For instance, it was unclear where the trigger suggestion would be noticeable (which stems from the criteria trigger is noticeable) within the prototype, so it was

implemented twice. One position in the prototype was proven to be noticeable and the other was not.

Once the pilot-study had been concluded, we could start the design study. The design study was set up with video to ensure that all gestures and expressions were taken notice of. This was important as it can suggest how certain aspects of the prototype should be designed (Daly-Jones et al., 1998). The respondents were informed of their rights as suggested by Vetenskapsrådet (2019) to ensure ethical procedure. In other words, the respondents were made aware of the purpose of the study, that it would be respectful of their integrity and that they could discontinue their participation at any time. Respondents were also asked if they consented to share their screen.

Once respondents had been made aware of their rights, they were provided with a scenario giving them a concrete description of the tasks they were about to be assigned, creating engagement (Haynes et al., 2004) It also helped remove some of the vagueness the respondents might have felt, by grounding it in the user experience (Haynes et al., 2004).

When the scenario had been carried out, the respondents were asked to navigate the

prototype. The navigation was administrated with four tasks (see figure 2). The tasks guided the respondents through a desired navigation, in which the four triggers were all

implemented.

Figure 2. Tasks respondents were given when asked to navigate the prototype.

After the navigation, respondents were asked questions based of the criteria. The idea was to use the criteria affiliated with respective trigger – exposing the respondents to the triggers and asking questions relevant to the criteria afterwards. We asked questions such as “Did you see any messages?” and “What was your reaction to them [messages]?”. That way we could see how each trigger related to the criteria without having to make the respondents be the ones to do the analysing.

Analysis & criteria

After the respondents had navigated the prototype and answered our questions, we needed to analyse the empirical data. A thematic analysis was chosen because of its eligibility to analyse qualitative empirical data (Bryman, 2012). We carefully read the transcripts and thereafter reconstructed it into data. The data was thereafter put into an index for each trigger.

The indexes acted as central point to relate the criteria with themes and subthemes (see table 9). For instance, if a respondent was asked “Did you see any messages?” and responded with a description or experience of a trigger, we knew that the trigger was noticeable (which is the first criteria for the triggers – see table 4). Subsequently, it could also indicate other things

Log into your ICA Maxi

account

Buy specifically

two carrots Proceed to

checkout Complete your

order

(20)

such as if the respondents found the triggers invasive (which stemmed from the 8^th criteria, the trigger was not invasive - see table 4). This provided us with a theoretical understanding of PD, e-grocery and food waste helped us with our theoretical contribution.

Table 9. Example of index for criteria, themes and subthemes.

Criteria Respondent #1 Respondent #2

Frequency was appropriate Explained that the trigger needs to surprise the customer, without scaring them away.

Explained that they were okay with it as long as it only happened once.

Method Discussion

The respondents X and Y gave us feedback (in the pilot-study) in which resulted in improvements to the prototype. The trigger that was mainly discussed was expertise

credibility. The trigger was originally implemented with a video placed on the homepage of ICA Maxi. The representative expert, Paul Svensson, was chosen and implemented (see figure 3). He was chosen as representative because of his solidarity to the subject of

counteraction of food waste. He has, for instance, created a book that is called Vegetables A- Ö where he enlightens the readers of his best tips to counteract food waste.

Respondent X was forthright and displayed dismissal of the video. They said that they would not watch the video. When respondent X was asked if the right expert could change their mind, they said that nothing could persuade that outcome.

“I wouldn’t watch the video. There’s nothing that could change that.” – Respondent X Respondent Y somewhat resonated alike respondent X, by stating that the design would not appeal to them. Although they proceeded to explain the appeal was dependent on who the expert was.

“If the video showed a celebrity chef, I really take an interest in, I’d probably watch.

Especially if it auto played” – Respondent Y

It was concluded that although Paul Svenson’s solidarity to the subject is pristine, he lacked fame – both respondents failed to recognize him. The pilot-study therefore concluded with the assessment that the trigger required improvement. Edward Blom (a gastronome and chef) was instead selected to feature as he is more well known. He was also presented in a picture as opposed to a video, as a result of the dismissal of respondents X and Y (see figure 4).

(21)

Figure 3. Trigger expertise-credibility before improvement.

Figure 4. Trigger expertise-credibility post improvement.

(22)

4 Design study

In this chapter of study, the problem we are studying, is addressed. After addressing the problem, the prototype introduced the triggers, how they were implemented and how they were evaluated.

Problem Definition

As humans, we are currently creating havoc with our planetary health because of

unsustainable, wasteful use of resources (UN, 2021). It is suggested that countries implement strategies to intervene in critical stages of the food chain, as an initiative to reduce

unnecessary food loss and waste. Research shows that strategies need to be implemented at the consumption stage (see figure 1). One of the stages in which we believe food waste occurs and can be influenced, is at consumption stage - when individuals are purchasing their food.

E-grocery as a service has been growing and is estimated to increase within coming years (Saskia et al., 2016). E-grocery is also estimated to increase food waste as online purchases have been proven to decrease consumers feeling of psychological ownership (Illyuk, 2018) This stems from customers feeling less responsibility of the products as less effort is put in to choose the products (as the customers can not feel nor smell the product) (Illyuk, 2018).

NASEM (2020) suggests that PD could be used to counter said food waste. Although many of the persuasive systems within retail aim to encourage customers to purchase more – which leads to unreflective behaviours and impulse buying (IJsselsteijn et al., 2006; NASEM, 2020). PD then becomes a variable of the likelihood of food waste (Chu et al., 2014;

NASEM, 2020). Unlike using PD to incite purchases – there are currently few strategies for e-grocery retailers to counter this food waste. The idea was to use PD for the opposite effect – to make customers reflect of their purchasing habits. We did so by reworking four common triggers within e-grocery (default personalisation, reminder, suggestion & expertise-

credibility).

Design

In this chapter of the thesis, the vision and specification of the four triggers are introduced further. The motivation of how they have been reworked for the opposite effect is included.

Both a written and picturised visualisation is presented.

Default Personalisation

The default-personalisation trigger originated from a combination of the Chu et al. (2014) trigger personalisation and Mont et al. (2014) suggestion for the compliance of default. Chu et al. (2014) describe personalisation as a mean to provide information or a service for a specific individual and Mont et al. (2014) describe default as a way to provide automatic choices that are better suited for the user. Although the two triggers have previously been used separately, the triggers have been combined in this research because of their potential together. The idea was that the system would use digital traces and data to present a default personalisation. Consequently, personalisation could by default sort the groceries preferred by user – one of the most important concerns costumers have. In the prototype, vegetarianism was exemplified, but the idea was that filtering changes according to personal need. For instance, if you are gluten intolerant, the system could filter accordingly. Hence, the idea was for it to act universally. The target behaviour of the user subsequently becomes supported in making more informed decisions and consequently buying less unnecessary groceries.

Although it is noteworthy, that the filtering did not intend to withdraw any types of options

(23)

for the user. The idea was for the user to still have the freedom to buy anything from the selection but having the filtered products prioritised.

The personalisation is visualised by a pop-up welcoming the user back with name and associated sorting upon login (see figure 5 & 6). It informs the user of the default

implementation that was made and upon what premise. The user thereafter can either agree to the implementation or navigate to their user page to change the default. This is the decision where our intent lies, the intent of reinforcing compliance. As Mont et al. (2014) describes, in situations where the user needs to make a decision such as this, they will agree as they don’t want to take the extra time to change it within their user page.

The pop-up would not be used every time the user was to log in, because the personalised prompt itself has been proven to be more effective to change behaviours than the generic and periodical reminder (Caraban et al., 2019).

Figure 5. Default personalisation. When logged in to their account the customer receives the pop up stating that their selection has been filtered based on previous knowledge.

Figure 6. Default personalisation. Close-up of the implementation.

Reminder

The idea was to use a trigger to remind the users by indicating what they had bought previous in the session followed by the question “Do you have any left?” (see figure 7 & 8) assisting the users to reflect on food waste when shopping.

The idea was to grab the user’s attention when they added groceries to their cart. To do this, we presented the trigger as a pop-up that appeared when inserting the desired amount of goods. If the purchase included groceries that had previously been bought, the system proceeded by asking the user to check that they did not have any left. The user is thus faced with decision-making – making the user rethink their planned actions and consequently altering their behaviour (Fogg et al., 2007). This impaction would hopefully result in the user inspecting what groceries that they currently ha on hand.

(24)

Figure 7. Reminder. The highlighted box on the right reminds the customer on what they have previously purchased.

Figure 8. Reminder. Close-up of the implementation.

Suggestion

The trigger suggestions aim to encourage re-use and motivate the user to a more sustainable handling of groceries. Consequently, the trigger sends a sustainable message rather than a persuasive one trying to encourage the user to buy more products.

The dilemma lies with timing; it is critical for suggestions to arrive at the right time for the user (Fogg, 2003). If the suggestion arrives after the appropriate timing, the goal behaviour will not occur, even with high motivation and ability (Fogg, 2009).

The trigger was implemented in the area where the user inspects the list of groceries they have put in their basket before finalising their purchase. To make use of having the user’s attention, the message is aligned next to the list of groceries. The message could, for instance, suggest to not throw away the vegetable scraps but instead use them to create a stock, or how to store groceries for the longest shelf-life possible (see figure 9 & 10).

Figure 9. Suggestion. On the right-hand side, the customer receives suggestions on how they can store their products for longer time.

Figure 10. Suggestion. Close-up of the implementation.

Expertise-credibility

The Expertise-credibility trigger originated from a combination of the Chu et al. (2014) trigger attractiveness and the Fogg (2003) trigger credibility. Chu et al. (2014) describes attractiveness as using an attractive model or a famous spokesperson to make your site more appealing for the user. Credibility is explained by Fogg (2003) as something that is perceived trustworthy which can be achieved by using someone with a certain stature, such as a judge.

The idea of expertise-credibility was to use a professional chef to inform customers on how to become more sustainable in the kitchen and with their products. Taking into account that customers might feel like the information that the expert provides is more trustworthy than

(25)

someone who is not (Fogg, 2003). The idea was that if an expert provides sustainable

messages such as encouragement to purchase groceries with short shelf-life, it is more likely that users would purchase those groceries. Consequently, the trigger sends a sustainable message potentially altering the attitude of the user. The trigger was exemplified as an ad on the front-page (see figure 11). The intention is for the expert to capture the user’s attention.

Once the expert has captured the user’s attention, they can become encouraged as to how to become more sustainable.

Figure 11. Expertise credibility. The e-grocery site uses an expert to encourage the user to buy short shelf-life products.

(26)

5 Evaluation

The evaluation was performed with a total of 12 interviews in which 2 of them were non- completion (because of failure to complete the objective that were provided to them, by us).

From the 10 interviews that were completed we received different results from each

respondent. In this section of the research, the evaluation is presented. It illustrated how each trigger performed to relevant criteria (see table 4). The criteria were collected from prior literature such, as from Fogg (2009), that acknowledges that there are three key factors to consider for triggers to be effective; being noticeable, being associated with target behaviour and occurring at the right time. Although all triggers were put in relation to all criteria, the most relevant ones are presented below.

Default Personalisation

The personalisation trigger is used by e-grocery services to assist their customers to find the right products by providing them with a new default. For example, an e-grocery service acknowledges a user’s preference by analysing what they purchase. If the customer only buys vegetarian products the e-grocery service filters out everything that is not vegetarian. The trigger was implemented as a pop-up that appeared right after the user log in. It welcomed the user back and informed of the default filter.

Trigger is noticeable

All the respondents noticed the trigger, but it was not a part of their goal, which resulted in them continuing with their shopping, as respondent #3 stated. Some of the respondents such as respondent #1 didn’t like that the store filtered out what they could and could not see in the store, something that suggests that the respondents’ marketplace metacognition relates to their freedom of choice.

Respondent #3: "I’m just closing this down. I’m just buying carrots "

Respondent #1: “When the pop-up appeared, I didn’t want it to filter my food”

Some of the other respondents, such as respondent #4, stated that they were happy that the store was aware of what they had previously shopped for and filtered based on their previous choices.

Respondent #4: "I was happy to be welcomed. And I was happy that they had recognised that I had been eating a lot of vegetarian food"

The trigger was personally relevant

As the store filters the selection based on the user preference the results are personally relevant. Respondent #9 and #7 stated that they believed that it was relevant to them as the scenario told them that they eat a vegetarian diet.

Respondent #9: "I eat a vegetarian diet"

Respondent #7: "Generally, absolutely yes. And I noticed that there was more vegetarian food on my homepage than what I usually have on ICA Maxi"

But on the other hand, respondent #10 was unsure if the filter was relevant to them.

(27)

Respondent #10: "I guess"

The trigger was not invasive

As the pop-up only comes up when the site realises the user’s preference the respondents didn’t find the trigger invasive. Overall, the respondents didn’t find the trigger invasive.

Respondent #7: "It was very comfortable. Nothing bothered me if I put it that way"

Respondent #10 mostly agreed with the rest. But they stated that if the message only came up once instead of every time they logged in, they would not see it as invasive. Hence, if the trigger were be used more than once, it could be perceived as invasive.

Respondent #10: "No, I don’t think the first one [personalisation] wouldn’t be as bad if it only happened once"

Reminder

The reminder trigger is used by e-grocery services to enable the customers to recall what they have previously bought to eliminate purchasing products that they already own. For example, if the customers had bought a sack of carrots the week before the trigger reminds them so that they do not over consume. The trigger was implemented as a pop-up to interfere with a sustainable message and had the intent of potentially altering its user’s behaviour.

Overall, the respondents noticed the trigger. Respondent #10 indicated recognition by stating that they found the trigger annoying as it was just another click to complete their purchase.

Respondent #10: "Whatever, that kind of annoyed me."

Some of the respondents such as respondent #4 stated that they do actually need more carrots which could mean that they knew that they had no more left from their previous purchase.

With said statement it is safe to assume that respondent #4 also noticed the trigger.

Respondent #4: " Now it tells me if I’m sure that I want to buy more carrots, but I eat a lot of carrots, so I want more carrots."

Much like respondent #4, respondent #2 showed acknowledgement by stating that they appreciated that the store could recognize what they had bought previously.

Respondent #2: "I appreciated that it could recognise that I bought carrots last time"

The trigger was motivating/engaging.

Most of the respondents found the trigger motivating as it could make them aware of what they had previously bought. Respondent #1 stated that they think that it will make people more aware.

Respondent #1: “I believe it’s a good sign, it makes people more aware”

(28)

But at the same time respondent #10 and #7 stated that it’s a good idea but it has a negative emotion attached to the message. Respondent #10 stated that the message wasn’t motivating but it was more shaming them for wanting to buy carrots.

Respondent #10: "Yeah I defiantly think that is a good idea but now its shaming me for wanting to buy two fruit and greenery, which I do not like."

Respondent #7: "Could be positive, could also be negative. Should ICA be telling me this?

But the conception is good. I believe that it will be more beneficial than the other way around"

The trigger was not invasive

Respondent #2 among others, did not find the trigger invasive but more of a friendly reminder of what they had previously bought.

Respondent #2: "No. I regard it as like a friendly reminder"

Respondent #10 and #4 on the other hand found the message invasive when they were shopping. Furthermore, respondent #10 stated that they saw the message negatively as it questioned their ability to know what they wanted to buy. They suggested that the message should be more passive.

Respondent #10: "This window in particular makes me more annoyed, I guess. I don’t like that it questions my ability to know what I need."

Respondent #4: “The one with that you bought carrots last week was a little invasive, [...].”

Ultimately, the majority of the respondents found the trigger non-invasive, but some felt the very opposite.

Suggestion

The suggestion trigger was used to encourage, motivate and inform the user of how to be more sustainable with their products. For example, when the customer are looking in their shopping cart, they receive a suggestion on how to properly store the product to get a longer shelf-life. The trigger was implemented beside the shopping-list.

One of the key factors according to Fogg (2003) was for the trigger to be noticeable. Overall, the respondents seemed to acknowledge the presence of the trigger. Respondent #1

acknowledged the trigger by expressing appreciation for the message. In the same way as respondent #1, respondent #4 praised the message but would like a prolongation stating the shelf-life date.

Respondent #4 “I really like the tips [granska page] but I would like to see how long time i have to eat them for example try to eat the carrots within two weeks. When we get the food bags it usually says on them that think about doing the recipe this week as the ingredients have short life span.”

Respondent #1 ”I like the message about how you properly store carrots”

(29)

Respondent #10 on the other hand did not initially acknowledge the message within the trigger. Although the respondent implied that the message would be acknowledged in a real- life scenario of purchasing groceries or if the placement was rearranged.

Respondent #10 "Not the first time, I was more looking at this (the red box that says “hem- leverans”) I would have probably notice this if I was actually purchasing it, it also could be because it was on the right of the screen maybe I didn’t know it at first."

The trigger has potential to change attitudes.

Another criteria was for the trigger to have the potential of changing the attitude of the user.

Respondent #1 implied that the objective of the trigger had come across according to the target behaviour – stating that its insinuated encouragement being to review their habitual consumption. Respondent #4 implied similarly by stating how easy such a thing would be to do.

Respondent #1 ”It informed me that I have to be careful of how much I waste and consume”

Respondent #4 “I have never stored carrots in water it’s a super easy thing to do.”

Although, respondent #1 and respondent #4 respectively, only implied that their attitudes could potentially change, respondent #3 provided the research with a straightforward answer, claiming that it probably would change their attitude. Respondent #3 also emphasized that they recognized the value of such message - something that the majority of respondents sided with.

Respondent #3 "Yes, I believe so"

The trigger prompt reflection.

Some respondents expressed indifference toward whether it made them reflect upon the trigger or not, but a few respondents expressed that it made them reflect on their behaviours.

Respondent #7 "These kind of messages makes me reflect more"

Respondent #9 "Reflect about better food habits. Think about the environment you know "

Expertise-credibility

The expertise-credibility trigger was used to inform customers on how they can be more sustainable by using an expert such as a professional chef. Customers might feel like the information that the expert provides them with is more trustworthy than someone who’s not.

It was implemented as a add-like feature on the homepage with Edward Blom as the expert.

The majority of the respondents did not acknowledge the existence of expertise-credibility.

Respondent #5 argues that the pre-existing interface from ICA Maxi’s frontpage has an overwhelming number of pictures, making the potential for noticeability of the implemented trigger less likely. The reaction of respondent #3 was aligned with respondent #5, with both stating that all the advertisement overwhelmed them.

Do you have any left?

Digital Design and Innovation, 180 hp