Thanks to the widespread use of smartphones, many customers especially in Japan rely on using electronic coupons with their smartphone applications. However, in subscription-based electronic coupon services that are widely used, some customers do not take full advantage of their electronic coupons, as they forget they have subscribed to the service. In this paper, we study how to implement persuasive system design (PSD) model to encourage electronic coupon use for various services. To be able to create persuasive techniques, we conducted in-depth interviews to extract clues on how to achieve suitable methods. The interviews indicated that we rst need to improve the application's general user interface to ful ll the postulates presented in the PSD model and common usability goals. We also found that PSD model has several system qualities we can take advantage of such as dialogue and social support to create more persuasive features in the system. For example using praise, rewards, reminders, and 'nudge' to encourage term-limited coupon use. We also compare the system features to the PSD model with examples while pointing out that the model lacks development that often uses cycles instead of linear development and is partially vague in describing system credibility and direct and indirect persuasion implementations. Last, we shortly discuss ethical implications of using persuasive technology and behavioural change methods in a commercial application.
The widespread use of electronic coupons in Japan enables purchase of items and services with discounted price or getting free products. The coupons in general are ubiquitous and are o ered by many companies. Many customers use multiple membership cards (T-point1, WAN2, Aeon3, etc.), or use coupon 1 http://tsite.jp/ 2 https://www.oizumifoods.co.jp/premium/ 3 https://www.aeon.co.jp/point/waonpoint/index.html applications for speci c stores, such as McDonald's4 or Lawson5 convenience stores. This saves money in a constant and convenient way. The reason they are so widely used in Japan comes from changes of the current society according to a web article6. Japan has high population density and compared to other countries, competition is more erce, as they o er good price-performance from shop to shop to the users. Customers are also more price conscious because of the long-stagnant domestic economy and want to save more money by any means available. Smartphone applications also play a signi cant role in encouraging customer's coupon use behaviour, since they can search, show and bene t from them. According to a report7, the number of coupons has increased at the same rate as the amount of smartphones in the market.
However, while coupon numbers increase, not all users utilize them. According 8, nearly 40% do not use their electronic coupons, because there is too large variety available, which makes it di cult to make use of them when wanted. The abundance leads to ood of information. This is a challenge to companies providing a subscription fee-based coupon service model. When the customers cannot use their coupons or forget they have subscribed to a coupon service, they might feel they are paying for nothing and this might degrade their customer service satisfaction. Therefore, we need to develop an approach using persuasive technology to encourage customers to use their coupons, so that they feel satis ed with the service. KDDI(au)9, a telecommunication company that o ers telecommunication services, has a similar challenge with their Smartpass coupon service10 . A customer can buy their subscription and get discounts for example on foods and drinks at various stores, or to watch movies in a theatre. However, many customers forget they have subscribed to the service. Thus, we aim to develop persuasive techniques to encourage its use. For the above-mentioned objective, we have rst conducted in-depth interviews where we inquired reasons why customers do or do not utilize their Smartpass coupons.
Many other studies[
Our major contribution is that we are rst to propose encouraging coupon use by utilizing Persuasive Technology. Speci cally, the interviews provided clues towards using suitable persuasion. Second, we looked into how PSD model's system qualities would encourage coupon use in meaningful ways. Lastly, we compare our system features to the PSD model with examples, and argue how the model lacks focus when actual development often uses cycles instead of linear approach. While PSD is good for our system, it is still partially vague in 4 http://www.mcdonalds.co.jp/ 5 http://lawson.jp/en/ 6 https://www.nytimes.com/2011/02/14/business/media/14groupon.html 7 http://www.ilovecouponmonth.com/statistics/ 8 https://whatjapanthinks.com/2010/05/02/mobile-phone-discount-coupons-injapan/ 9 http://www.kddi.com/english/, KDDI refers to a telecommunication company and au is their mobilephone communication services brand name 10 https://www.au.com/entertainment/smartpass/ how system credibility is created in practice and how to implement direct and indirect persuasion. To the best of our knowledge, there are very few papers working towards the above mentioned topics. 2 2.1
Persuasion and/or behaviour change are closely related topics, which can be
used in various ways and in elds to change a person's state of mind, attitudes
or behaviour. For example, they can be used to guide a person to change into
healthier life-style by dieting, reinforce existing willingness to quit smoking, or
form a totally new behaviour. To enable these changes, persuasive technology
(PT), persuasive system design (PSD) and behaviour change support system
(BCSS) methods have emerged. In Oinas-Kukkonen's [
While the use of persuasion and behaviour change techniques can o er
advantages in changing a customers behaviour towards using applications or services,
at the same time ethical questions need to be considered, which might be linked
to forcing a user to do or buy something they do not want to. There have been
papers discussing how to approach ethical issues gathered in Karppinen et al. [
However, in the case of Smartpass coupon service, introduced in more detail in section 3.1, the customer is a willing participant as they have subscribed to the service to get the coupons. What they need is a support system that 1) reinforces the use of the service 2) reminds the existence of the coupons 3) encourages other bene ts for the user by utilizing e.g. health improvements via walking or exercises, and 4) is tailored for the users' speci c needs. Our initial assumption is that behaviour change in the case of electronic coupon use is not coercive or deceptive in nature and can be utilized.
To create a persuasive system, Oinas-Kukkonen [
PSD presents seven guiding postulates to design, as seen in Table 1, or for evaluating a system design, as is also our case. While on the surface, postulates 6 (unobtrusive) and 7 (ease-of-use, useful) are common usability goals for any system, being able to persuade also requires a system that is pleasant. This leads to user acceptance , and for the acceptance of future behaviour or attitude change as it is easier to achieve with usability. Some users found it frustrating or di cult to nd information about the coupons or had problems locating coupons they needed, suggests postulates 2 (organised and consistent info), 6 and 7 should be implemented in the next iteration of the application to reach basic usability levels for the software. We can also utilize various direct and indirect routes (postulate 3) to persuade users within the application, while making sure persuasion is done in an open manner (postulate 5). PSD guidelines, as seen in Table 2, o ers four main categories of system qualities: 1. Primary task support, 2. Dialogue support, 3. System credibility and 4. Social support, each including seven sub-principles for reference. These o er design improvements and basics that focus on di erent parts of a system and the user, so that the end product is assisting in the main tasks of the user, encourages use with di erent dialogues, is presented in a credible format and also focuses on social aspects for the user. The end goal is to make use of the sub-principles to create more e ective persuasive system features.
Smartpass is a subscription-based coupon service. Fig 1 illustrates the ow of starting and using of this service. First, a user subscribes to the service for a monthly fee and gets the right to use any coupons distributed in the Smartpass application (Smartpass App.) separately installed in a smartphone. With the coupons, they receive discounts and items for free at grocery stores, restaurants, movie theaters, and other locations. Smartpass is o ered in two types of subscriptions: Smartpass-premium and Smartpass, where the only di erence is that users can get more discounts and items when using the former.
In our survey, most users joined the service when they purchased a new phone or when they upgraded their old smartphones. In Japan, users often do Qualities Principles Primary task Reduction. Reduce complex behaviour into simple tasks support Tunneling. Guiding a user creates changes to do persuasion Tailoring. Tailoring to needs, usage context and others to a group Personalisation. Personalised content/services is good for persuasion Self-monitoring. Keeping track supports user achieving their goals Simulation. Simulating provides cause and e ect linking behaviour
Rehearsal. Practicing enables change in behaviour/attitude Dialogue support Praise. Using praise can make users open to persuasion Rewards. Rewarding target behaviour is great for persuasion Reminders. Reminding target behaviour a ects achieving goals Suggestion. Fitting suggestions is good for persuading Similarity. A system that remind users themselves are persuasive Liking. Visually attractive system is more likely persuasive
Social role. A system in a social role is more likely persuasive System credibility Trustworthiness. Creates increase in persuasion support Expertise. A system viewed as competent a ects persuasion Surface credibility. Initial impressions a ect credibility Real-world feel. System highlighting creators has more credibility Authority. System's roles of authority enhance persuasion Third-party endorsements. Well-known sources boosts credibility
Veri ability. Content accuracy should be veri able Social support Social learning. Target behaviour motivation by observing others Social comparison. Motivation by comparing to others performances Normative in uence. Peer pressure increases adopting behaviour Social facilitation. Performing with others increases change Cooperation. Motivation to change from natural cooperation drive Competition. Motivation from natural competition drive
Recognition. Public recognition increases target behaviour change this exchange in yearly or biyearly intervals, while often forgetting about the subscription service later on even though they pay a monthly fee. This is a challenge for the service. In order to correct this, KDDI has held campaigns like \Santaro-no-hi" three times a month, where customers were able to get special foods or items at stores, while also advertising the campaign on TV and in print. However, the e ects were only temporary and limited in encouraging customers to make full use of Smartpass coupons. 3.2
Fig 2 illustrates the ow of our research. In \In-depth interview", we conducted a questionnaire survey to extract reasons why customers do or do not utilize their Smartpass coupons. In \Plan persuasive techniques", we will develop techniques to encourage customers to take advantage of the coupons using Persuasive Technology and behaviour change approaches. In \System deployment", we will develop a smartphone application system that takes advantage of the techniques
KDDI customer signs up to purchase Smartpass service, when he/she buys phone.
Select the coupon in Smartpass App.
Exchange the coupon to items at store
Customers make use of coupons.
persuasive technique
One bottle of Coke
FREE of CHARGE!! developed in the previous phase. Lastly, we will evaluate the persuasive techniques and the systems in \Evaluation". In this paper, we focus only on the \Indepth interview" parts of the ow where the interviews were done as follows. First, we conducted a screening questionnaire to exclude unsuitable subjects, and second we gathered the subjects for in-depth interviews. 3.3
Method We had a questionnaire to extract individuals whose frequency of using Smartpass service was low. It consisted of the following items: { Demographic information: gender, age, occupation, marital status, children, income, etc. { Information on subject's phone model, hours of use, used functions, and their service provider { The use of au services: term, uses Smartpass (yes/no), frequency of Smartpass application use, and the use of Smartpass services { Lifestyle: the use of coupons besides Smartpass, frequency of eating out and the names of the restaurants, frequency of using leisure and other services Data collection We conducted a web-based questionnaire between 02.10.2018 { 05.10.2018 for users registered to Rakuten Insight11. The subjects were asked to answer the questionnaire via email or by pushing noti cations from the Rakuten App. Users answering the questions, earned Rakuten super points, which they could use as money in Rakuten services and an in an Electric Commerce website. 8,000 subjects answered the questionnaire ranging from 18 to 69 years of age. Screening results From the 8,000 we screened users suitable for in-depth interviews with the following steps. Empirically, chose subjects who ful lled: { Frequently-used phone type: smartphone { Frequently-used service provider: KDDI(au) { Smartpass application status in the user's phone: installed { Median frequency of coupon use other than Smartpass was more than 2.0 { Subjects who have used any point or discount cards. { The use hours of their smartphone in a day was more than an hour. { The number of used smartphone functions, such as calling, internet, and email was more than three.
In the end we arrived at 12 subjects by choosing them by their individual scores, using au service, loyalty scores, and frequency of using Smartpass apps, so that they would all be di erent with each other. Average age was 41.25. Table 3 shows the pro les of the 12 subjects for the in-depth interviews.
We conducted semi-structured interviews between 17.10.2018 { 22.10.2018 for the 12 subjects extracted from the screening process. The objective was to nd out the di culties and barriers for the customers to using their Smartpass coupons. The interviews lasted on average for 90 minutes and the questions focused on the following topics: { How often they used the Smartpass App., how they launch it, in which situation and activities they used it, and if there has been changes in how often they use it, and if they use the Premium service or not { How often or rarely they used the coupons { How often they use other coupons besides Smartpass either electronic or paper ones, and what are the di erences from Smartpass coupons { Frequency of using restaurants and other food related services, and if they took advantage of Smartpass coupons in these services, and if not, to explain reasons why. { Discussion on barriers and limitation when using the Smartpass application, improvement ideas, and discussion on what they thought about privacy issues our suggested new features 4 4.1
From the answers, we found out that users do not always launch the Smartpass App. from the main icon. Four subjects did it from within \au Service Top". A 11 https://member.insight.rakuten.co.jp/ F F C C F F C F u u o o u u o u ll-t ll-t P tn tn P ll-t ll-t trn ll-t P im im ra ra ra a im im a O im tS r
c c r c o e e -t t t -t e e t th e ud fe em em itm em em itm em em em re em en iss p p e p p e p p p s p t on lo lo r lo lo r lo lo lo lo y y y y y y y y e e e e e e e e e e e e e e e e A i A iP iP n A A i A h d h d h h dA ihP ihP dn dn hP dn eon n P n r o r o o r o o r r o r o n o n n o n n o o n o t id e id e e id e e id id e id yp o o o o o o o o o Y ev ev 1{ ev ev ev 2{ ev ev ev 1{ ev ea r r 2 r r r 4 r r r 2 r
r 4 4 4 4 4 4 4 4 4 s
P e P m le e n e e ea ay on on on eek tonh eek cy e o e e k th k k m th th th l e s s o n t h 8 8 1 1 le 8 In 0 0 6 2 6 6 ,0 o ,0 s 0 6 c 0 0 0 0 0 0 0 v 0 s 0 0 o 0 0 0 0 0 re 0 t 0 m
h { { { 1 1 { { { { { 14 {1 an ,1 (
{ e 8 4 8 8 1 , 8 J ,00 ,00 00 00 00 00 ,20 00 ,0 0 0
0 0 P 2 2 0 0 Y 0 0
0 0 0 ) Y Y Y N Y Y Y Y Y N Y N ra se se se o se se se se se o se o r i e d h Y N N N N N N Y Y N Y N il se o o o o o o se se o se o d
M C au service top App.
Notification view
Smartpass App. separate application, even though they also had the Smartpass App installed. \au Service Top" is a portal application for au customers, where they can check their current telecommunication fee, apply for changing their contract and for example read news online.
The portal application also has in-app noti cations, that show the availability of Smartpass coupons, which leads users to launch the application from within, instead of launching the app separately. 4.2
Ten subjects answered that they had trouble nding information on their target coupons when discussing their coupon frequency. They had trouble looking up suitable coupons from the application, while one commented having trouble getting information on the prices or items from the coupon banners. As a summary from the answers, we found out that the user interface design of the application is poor, as it was hard to nd or access preferred target coupons. The design of an easy to use interface for the user is a fundamental postulate of PSD and HCI, so before conducting a new evaluation, we need to improve the application. 4.3
From discussions on the di erences of coupon use between Smartpass and others, eight subjects answered they had trouble using coupons, because they could only be used during speci c times and they had a habit of forgetting to use them on time. These comments imply that persuasion with for example a reminder noti cation or using others like a 'nudge' style behaviour change features from the PSD model would help inform the user of an impending expiration date or otherwise encourage the use.
5.1
In this section we discuss some examples and improvements which we can create derived from the guidelines of persuasive system design.
Persuasion context We looked into using the PSD model from its persuasion
context of "Intent - Event - Strategy" for the current coupon system. The
intent, is about recognizing the persuader and the type of change we are looking
to achieve (attitude, behaviour change or both). In the case of Smartpass, the
persuader is the developer of the software, and the related parties o ering
services for it. The intended change is to have customer's behaviour towards the
use of the application changed. We can also argue that we are trying to change
the indi erent attitudes towards an unused application in some cases. The event,
is the use-context, where the system has a problem where customers forget to
use it, have trouble nding the correct information, or can not e ectively use
when needed. Mostly these use-context events are the characteristics of a
persuasive system, as the coupon application events enable change in behaviour. Most
important factor is to understand the user's goals, which would be to increase
coupon use to save money, or as an example, to use the application to exercising
via visiting coupon locations far away. According to Oinas-Kukkonen[
In this section, we present our views on some ethical concerns when using persuasive techniques for our system. Smartpass customers want to use coupon services, so at rst glance encouraging its use via Persuasive technology would not be unethical. However, conditioning and surveillance would not be acceptable, while also being open about the persuasion and o ering users a change to opt-out if needed, are essential.
In the interviews we also inquired privacy issue concerns with the subjects and from the answers almost all of the subjects did not care about their privacy related to gathering personal data. Some answered that they would agree, because they would be able to utilize their coupons better from the more accurate and tailored information.
In [
In addition, Karppinen et al. [