Going real time in water conservation – the Singapore experience

As the shower is a key water use area in households, accounting for as much as 30% of the water used, a collaborative project was undertaken between PUB and NUS (National University of Singapore) in 2015/16 to explore methods of leverage on smart shower devices to modify user behaviour towards greater water conservation. Motivated by, and building on, a previous study in Switzerland, 530 residential houses in Singapore were fitted with smart shower devices and water use data recorded automatically over four to six months. The data collected from more than 300,000 showers showed that Singaporeans use 20 litres of water per shower, on average.

The households were divided randomly, on the basis solely of location, into seven groups (see Table 1), comprising one control and six treatment groups. All treatment groups were provided with real-time consumption information, while five groups were further assigned a water consumption target to achieve per shower. The targets ranged from 10 to 35 litres per shower, corresponding to ambitious, moderate and easy goals. During showers, the smart shower device showed the volume of water used, as well as indicating ‘very good’, ‘OK’ or ‘too much’, depending on the water consumption goal and the volume of water used. There was also an accompanying animation showing a polar bear that ‘melted away’ into the background should the water consumption goal be exceeded.

It was concluded that water consumption was reduced by 2.13 L or 10% per shower when the smart shower devices provided real-time feedback on consumption. It was also found that the goal set for the household played a crucial role. The goal should neither be too ambitious (less than 10 litres per shower) nor too lenient (30 litres or more), to best motivate households to reduce water consumption during showers.

In Singapore, a moderately ambitious goal of 15 litres per shower appeared to be optimal, relative to average consumption of 20 litres. Subsequent monitoring showed that the conservation effect was also stable over time. The study also showed a positive water conservation spill-over effect into other household water-using activities (beyond showering).

In comparison with other methods of providing users with water consumption information – e.g., monthly water bills – where the information is usually too coarse (i.e. aggregated at household level) and out of date (i.e. not current consumption), smart shower devices enable the provision of real-time feedback targeting a specific activity and relate more directly to the user. With ∼30% of daily water use concentrated in a short time of only 5 minutes or so, smart shower devices are a technological means of catching and modifying the user's attention towards water conservation very quickly.

Given these encouraging findings, PUB embarked on its Smart Shower Programme in 2017, to install smart shower devices in some 10,000 new homes under a demonstration project over the next 3 years, to enable further study and validation of their conservation effects during showers. PUB has appointed two companies, Amphiro AG, and Smart and Blue (Hydrao), to supply and deliver the devices (see Table 2):

Smart shower device installation started in 2018. A survey/interview will be conducted for a selected group of participating households to collect their socio-demographic data, which will then be matched with the consumption data to gain better understanding.

PUB hope that the use of these smart shower devices will motivate behavioural change and complement traditional policy tools, in the wider effort to manage water demand and achieve Singapore's long-term water conservation goals.

If the demonstration project results are promising, PUB will work with the relevant agencies to study the feasibility of rolling out the Smart Shower Programme on a larger scale.

PUB collaborated with SUEZ using their long-range smart metering and smart water technologies in a two-year pilot project. Smart meters fitted with VHF transmitters were deployed to 525 residential households (five residential apartment blocks) in Punggol in March 2016. Using the data collected, a mobile application called ‘WaterGoWhere’ (Figure 1) was created to motivate residents and increase their awareness of water use. A six-month field trial, which ended in December 2017, was conducted to assess acceptability and feasibility.

The WaterGoWhere app uses smart metering, analytics and an innovative gamification concept to engage the residents with respect to water conservation (Figure 2). Besides leak notification and usage monitoring, users also receive daily, weekly and monthly challenges to help them reduce water consumption.

Using the data collected from AMI (Advanced Metering Infrastructure), the programme:

• Provided residents with daily and hourly information regarding water use patterns through an innovative mobile application, helping to raise awareness of water wastage and how to improve consumption management.

• Provided PUB with AMI maintenance and operational information, as well as metering data computed by in-house Data Analytics tools (forecasts, patterns, savings, residual volumes, etc.). With these data, PUB can customise engagement strategies to help customers conserve water better.

The app provided a gamified, engaging and rewarding experience for residents. By using smart meters and data analytics, the system can detect household consumption patterns and present customised challenges to each household. The algorithms also detected the peak consumption periods during the day or week, and sent high consumption alarms or challenges at specific times.

The use of leak alerts in the app also enabled customers to become aware of leaks more quickly. They could detect leaks on their own using the information provided, which also reduced the manpower required to visit households and ask them to fix leaks. Leak volumes have also fallen significantly since the programme's introduction.

The app also rewards residents with points and level status for participation in water saving challenges, application login and water use reduction efforts. The top teams and players win rewards, and a grand prize at the end of the programme. The points serve as a virtual currency, to raise residents' awareness of their water consumption. Voluntary submission of data by residents is also encouraged via gamification. Apart from the online group, an offline group (i.e. households who did not sign up for the apps account via their smart phones or website) is also being engaged actively via monthly reports. Offline group members tend to be those without smart phones and unable to speak English.

Preliminary study has shown that:

• 52% of households were interested in the programme and downloaded the mobile app – this unexpectedly favourable result arose from the customer engagement strategies used

• 34% of households (who have signed up to an account) viewed their water usage and participated in challenges on a regular basis (at least once a week)

• There is positive feedback from residents (Figure 3) – e.g., their comments on Facebook and email – even including requests to join the programme received from residents of other estates.

• Engagement via app can deal with ‘tough customers with leaks’ where traditional methods have failed.

• Total water savings in the five selected blocks were 5%, with leak alarms and customer engagement efforts contributing 6.9 litres saved per capita per day.

Customer segmentation and profiling based on the data collected from residents through the app and AMI have also helped improve understanding of the factors contributing to customers' behaviour towards water wastage and saving.

The aim of the app design was to allow customers to communicate actively and interact with the mobile device, while learning about their water consumption, so that they would login again for new information.

The prototype interface design was based on industry best practice, and feedback from employees in both SUEZ and PUB. It was then presented to focus groups to gather feedback for further enhancement. User interface development has yielded some key learning points:

• The colours chosen relate to the subject. Blue represents water and white provides visual resting space.

• To guide users forward, all interactive buttons are green or pink. These colours are reserved specifically and only for CTAs (calls-to-action), enabling users to navigate the app easily.

• Complex, data-heavy, scientific graphs are simplified to make the app friendly for users as young as 10-years old.

Some 267 households signed up for an online account by the end of November 2017 and 52% were successfully recruited as online users. A large proportion of the online users, 60%, were recruited through door-to-door engagement, while 40% signed up after receiving invitation letters or monthly reports.

Analysis of the data shows that the majority of customers are between 30 and 39 years old, in line with Punggol's age demographic. On the other hand, 45% of those actively engaged with the app (by logging in once a week) were over 50 years old. This shows that the programme can appeal to an older generation, despite Punggol being a relatively young estate. However, the opposite could happen on a mature estate (i.e., the app may appeal more to a younger generation there).

Online customers were also segmented into four different types based on analysis of their login frequency, water savings and participation in challenges. All online customers showed interest in the programme by winning various challenges in the app and 96% of them had won points at least once because of a fall in weekly water use at home.

• Devoted Players (55%) participated in all challenges and reduced their water use as a result. They have played the challenges and realised how effective they were in water saving. Equally, they consistently get what they perceive as a great experience, triggering them to login again, and have progressively higher expectations of the programme, as they earn points and rewards. They also achieve the greatest per capita reductions in household usage, making them the ‘best in class’ customers needed to advance and improve the programme. The focus is to continue to engage and reward them for their water saving efforts. They were gained as customers after thorough engagement and consistent customer service.

• Delighted Players (31%) participated in water saving challenges, but not the weekly quiz challenges or Facebook discussions. The water saving challenges were sufficient for them and they felt no need to devote more time on quizzes or discussion. Such people are thought to prefer just to use the app for self-monitoring and are not interested in rewards, although they could also be affected by lack of knowledge of how to play the quizzes or participate in discussion. A solution to provide more in-depth engagement with the app might need developing. The most active engagement period for these customers is in the morning – useful information for targeting content update, triggering challenges and avoiding server maintenance.

• Detached Players (4%) login to participate in quiz and monthly water saving challenges, but do not win any points for weekly water saving. They are thus detached from the programme objectives. They seem only to want to play specific challenges to win points but not to put in effort to reduce their usage. They have won monthly water saving challenges due to the specific instructions to reduce usage over a certain period but have not achieved a family water reduction. For example, a player's monthly water saving challenge was to reduce usage on Sunday, so instead of doing the laundry on Sundays it was done on other days, just to win points. This is the smallest group of players.

• Dormant Players (10%) logged in only once, during registration, and not thereafter. There was no household reduction in water use. Even though they did not login after the first time, they received leak alerts and monthly usage reports consistently by email for notification. However, these did not persuade them to save water. They were generally in older age groups, perhaps because of the challenge in accepting new technology or inability to comprehend the purpose of the programme. The obstacles for these customers need to be understood to determine whether they arose from difficulties logging into the app, insufficient information in the monthly report or disinterest in playing the challenges.

Customers who did not sign up for an online account continue to receive monthly reports with information on monthly use, leaks and high usage information. Each report includes a call-to-action – ‘sign up to an online account to view more usage details and participate in challenges to win vouchers’. The challenge in converting offline customers to online was their low occupancy rate. They are rarely at home, so it proved difficult to promote the programme through door to door engagement or monthly reports. However, 60 offline customers were successfully converted to online through the monthly reports.

Information about the effectiveness of the reports towards water saving was inconclusive due to limitations in feedback from offline customers, but the reports do seem to be an effective channel for promoting the app.

Feedback was collected from customers during face-to-face meetings, by emails, Facebook discussion and events. Several app features were deemed effective by participants in the programme:

• Neighbour and block comparisons (customers like to compare their usage with those around them)

• Positive messages, rewards and encouragement (customers are motivated by positive reinforcement)

• Daily usage trend information (with hourly and weekly information)

• Monthly face-to-face meetings to help customers with challenges in using the app and/or saving water

• Monthly reports sent out to offline customers

Building on the success of the WhereGoWater pilot trial, the programme will be launched in a second estate, Yuhua, in mid-2018. The households in Punggol will also continue to access the app. (This AMI programme is readily adaptable to the energy and waste management sectors.)

Based on reports and block observations, it is noted that Yuhua has a relatively elderly population, which makes it very different from the Punggol estate. The population is also understood to be predominantly Chinese. Therefore, changes will be made to ensure that the programme is effective in achieving its goals in Yuhua, including:

• Digital Clinics – classes on how to use the WaterGoWhere app, with sessions also available in Chinese.

• WaterGoWhere Ambassadors – ambassadors (RC Volunteers) will be recruited to aid engagement

• Home visits – home visits will be made to customers needing help using the app. These give the team the opportunity to talk face-to-face, and build trust in the product and services.

Before the launch of the Smart Shower and AMI WaterGoWhere programmes, residents were not generally aware of their water usage or ways to reduce it. These knowledge gaps were addressed by providing both real-time shower and household usage data, and water saving tips.

AMI has also enabled the collection of richer household information for a customised experience. Using visually appealing graphs and charts to represent complex data has also helped to engage residents to understand their usage better. The programme's innovative and unique feature was its provision of water usage data in a gaming context that includes a ‘fun and reward’ element. This has not been tested in previous research in Singapore. By playing various challenges and monitoring their own water use, residents were equipped with better knowledge on how to change their behaviour to manage home water resources effectively.

The engagement with the 30 to 39 age group was expected as it aligned with Punggol's general demographic. However, there was unexpectedly good engagement with the over-50 s, too.

The challenge remains for user retention in using the smart shower devices and the mobile app, and how to keep users engaged and ensure repeat usage.

In conclusion, this trial proved that the programmes have been a successful medium in changing customers' behavioural attitudes and habits. For AMI, much planning in other trial aspects also helped its success – e.g., the customer engagement efforts, and thorough planning and innovation of the programme. This programme has also demonstrated the effectiveness of using gaming to encourage efficient water use in customers' households.