Review of Survey Data Investigating Customer Preferences Affecting Demand for Clean Water
Background
Last week we completed behavioural surveys of two villages from our recent projects. One purpose of this survey was to try to understand why communities are not consuming water in the volume we anticipated. To this end, we selected a (relatively) high consumption village (Mandwanga) and a low consumption village (Nyundo 1). There relative performance vs. forecast (at 8 litres per person per day) is seen below:
Both of these involved significant infrastructure investment to bring clean and safe water where previously community members from both villages had to travel many miles in the dry seasons for water, which was often not clean and safe.
Summary
This article describes how our hypotheses (described here) around rainwater substitution and resistance to behavioural change have affected consumption at our distribution points. While we factored in a 40% reduction in consumption during the rainy season and a ramp-up period to anticipate behavioural change, the results have still disappointed.
However, the good news is that the ease of use and smartcard charging associated with prepaid metres is not a barrier. Low demand seems to stem from other causes:
the taste of water in Nyundo 1 is considered ‘bad’ and described as ‘salty’, affecting its use for drinking;
there is a general appreciation of the benefits of consuming water from the taps (instead of stored rainwater, for example);
but there seems to be a gap between conceptual understanding of marketing messages and action taken as a result.
We will try to find a technical solution to the salty water and we have developed a marketing strategy to address the gap between message understanding and action. More below.
Interested readers can review the data themselves here.
Hypotheses Examined
Rainwater preference and seasonal patterns
The data confirmed seasonal consumption variations coinciding with the rainy season.
This was not a surprise and our forecasting has always assumed a 40% drop in consumption during the rainy season.
Convenience of collection and the fact that it is free were the main reasons for this. In addition, in a discovery that will be examined more below, preference for the taste of rainwater over water from the taps was prominent in Nyundo 1 village (Cohort B in the graph below right) chosen for its low consumption) but not really in Mandwanga.
Limited access to smartcard sellers
The data show that relative difficulty in charging smartcards to use at prepaid metres is not a big difference between a high use village (Mandwanga - Cohort A) and a low-use village (Nyundo 1 - Cohord B):
However, our experience on the ground and informal feedback does make us believe that distance to top up a smartcard can be a barrier to use. Indeed respondents to question Q7, ‘How easy is it to charge?’ may not have been including the distance to travel in their consideration of ease.
This view that proximity to a tap reduces rainwater substitution finds support at the extreme end, where nearly 60% believe they would not continue to collect and store rainwater if they had a tap in their household.
We will therefore work with the CBWSO (who assigns agents) to increase the number of agents in the villages with the aim that these can be within 10 minutes’ walk for the majority in the communities.
Smartcard ownership is high and people value the flexibility and convenience of 24/7 access to water:
We will explore ways to increase smartcard ownership in Nyundo 1 but we do not think that this is a significant factor in the low take up.
Discoveries
Taste as a factor
The graphs below show unequivocally that the salty taste of the water in Nyundo 1 village (Cohort B) is a reason for lower water consumption than in Mandwanga village (Cohort A).
The first two graphs above show how that Nyundo 1 respondents describe the taste of the water as bad and salty. The third graph show respondents’ opinion that this taste stops people from using more water. And the fourth graph shows that water from the public tap is used markedly less for drinking than for other uses.
We are exploring a technical solution to this problem in Nyundo 1.
We will also conduct another round of water quality tests which we expect to show that the taste is not related to water quality. This can feed into our marketing, where we can compare test results between tap water, stored rainwater and other alternatives that are used. We can then show the communities in direct and intuitive ways that using tapped water for drinking is much better from a health standpoint (which our data confirm is a priority for community members).
Marketing Opportunity
Till now, we have included messaging through intermediaries (local leaders) at the start of our interventions. However, changing people’s behaviour is a longer-term process that requires continual reiteration and reframing of the messages.
The data tell us that this approach has not been effective. A little less than half respondents received messaging about using water for more than drinking and even fewer received training about water needs (the first two graph below). Of those who received the marketing, less than half took action (L15 below) despite respondents being fairly positive about the content of the messages (L12).
This could present us with an opportunity to rethink our marketing strategy. Our marketing channels are clearly not effective and the messages did not inspire action to an acceptable degree. While respondents seem to understand the health risks of rainwater and the advantages of tap water (even the low users in Cohort B in the graphs below), this knowledge does not seem to affect behaviour.
Marketing Messages - What we should say
Marketing ideas include:
making the health consequences of consuming water from sources other than the tap obvious and startling (see above in relation to taste vs. water quality);
reframing cost as protection, not payment - “You’re not buying water. You’re protecting your family’s strength.” (cf. graph B47 above and respondents’ priority of family’s health);
We will start to incorporate in our savings group marketing income generating activities that have water as an input. There are many tried and tested examples of these and it is just a question of helping community members grasp these opportunities. Examples of small businesses that have been effective include: motorcycle cleaning services; brick-making and vegetable gardening.
Marketing Channels - how we should convey the message
We will focus our efforts on reaching customers in two ways:
60% of all respondents agreed to accept marketing on their mobile phones (which is 90% of those who own a mobile phone). This is a good cohort with which to experiment. We will develop a marketing strategy of short and impactful messages to be sent to mobile phones by SMS. These messages will be tailored to the season and what we know of the priorities of people at different times.
Every hamlet has a Community Health Worker (CHW). This is a new initiative from the Government. These CHWs can act as a sales force for clean water. Their roles involve gathering community members for events (plays, music etc) that are designed as means to convey messages. We will work with these CHWs more closely to align marketing materials and messaging on an ongoing basis.
Savings Groups
What will we do and how will we know if we have been successful?
SMS messaging - we will randomly select 50% (across Mandwanga and Nyundo 1) and over three months we will feed them SMS messages. Every week we will ask all respondents who agreed to accept marketing (and whose cell numbers we have) how much water they have consumed that week. We will track the difference between the two cohorts over three weeks and see if the cohort that has received messages has increased their consumption over time.
CHWs - this will be harder to test and is likely to be more qualitative. We will speak to CHWs and find out how they currently assess their success and use that as a starting point.
Savings Groups - number of new businesses started.
Conclusion
We are excited about what the results of the survey has shown us, revealing new issues and confirming or contradicting hypotheses. We will put our findings into effect over the coming months and share progress.
