Skip to main content

Monitoring and messages for faster maintenance

Published on: 28/10/2016

When something is broken, we might shout for help and then, if we are lucky, get it fixed. Our kids try this and sometimes they get a response. Something similar happens in water supply.

When a pump breaks or something else goes wrong with a rural water supply system, the WASH committee (WASHCO) running the scheme need to cry for help if they can't fix it themselves or find the necessary spares. They will most likely call or send a messenger to the woreda water office who may be able help. If needed, in turn they will call on the further help of the zone or region for a serious repair.

There is room to improve how such messages travel. And there is room to improve the response that a community gets to its cry for help. Sometimes there are gaps in capacities or the supply chain of spares, and then a repair may take months. The good news is that a lot of organisations are innovating to address these kind of problems. Often they are NGOs. Many are collaborating with government to take their innovations to scale. This needs a major effort given ambitious government targets. Under the second Growth and Transformation Plan (GTPII), the target is to reduce non-functional systems to 7% over the next five years, says Tamene Hailu who leads on rural water supply for the Ministry of Water, Irrigation and Energy. 

Non-functional water scheme

The company SweetSense, a spin-off from Portland State University, is working with the USAID-funded Lowland WASH Activity to see how sensors fitted on water supply schemes might help. The first of their sensors have just been deployed on relatively remote water schemes in pastoralist Afar. The sensors measure power, and share information over mobile or satellite networks. With the right tinkering and calibration, the power readings can tell you a lot of things: Whether or not the pump is working normally at any particular time, the amount water pumped during a day, and even whether the pump is likely to be about to break-down.

In Rwanda, SweetSense have used their sensors to compare the effectiveness of different ways of fixing rural water supply schemes, comparing business-as-usual repairs with a circuit rider model and a data-driven 'ambulance' model. Crucially, we don't incentivize maintenance, according to Evan Thomas from SweetSense. We don't put our money where our mouth is. We are much more interested in spending money on new things than fixing old ones. This problem needs much more than sensors to fix. The sensors are just a tool, says Evan. They cost roughly a 1000 USD a piece to install and then a few more hundreds to maintain. That investment is only worthwhile if you do something with the data. The most likely requirement will be to spend more money on maintenance. But in the longer term, that's a good investment of course, and the sensors can actually help you show how good an investment it is.

Presenters

Such sensors can also be used to shed light on other challenges in WASH. A lot of our policy advice is based on impact evaluations, based on indicators and questions that turn out to be rather uninformative or unreliable. SweetSense have used sensors, for example, inside water filters to identify the differences between what people say they do, and what they really do. This is important since water filters don't deliver their health benefits unless they are used nearly all the time, and something similar is true with the use of latrines. Sensors can also be used as part of payment systems. SweetSense work in other countries is linking payments to families to the actual use of cook stoves.

Even during the recent emergency triggered by drought and floods, UNICEF and its WASH Cluster partners have been innovating. With the help of Akvo and their software tool FLOW on smartphones, they developed and piloted improved monitoring of key emergency WASH indicators to help plan and improve the efficiency of the humanitarian response. Interestingly, the indicators included quantitative service delivery indicators like the number of people receiving 5 litres of drinking water per capita per day. Regular sector monitoring of water supply doesn't yet include service delivery related indicators like this, with no real attempt made yet to estimate the quantities (or qualities) received.

During the emergency, twice-weekly monitoring organised by specialised Information Management Officers and NGO partners was used to direct trucking of water to where it was most needed. A lot of care went into ensuring the quality of data collection: rubbish in, rubbish out, we were reminded by Luuk Diphoorn from Akvo. There was a quick response to the messages sent (in this case to a planning team in Addis). However, there was not an attempt to repair broken systems as part of the response, and apparently this was not possible given institutional constraints. Maybe that is something to revisit as part of future disaster preparedness. And in the future, it is also intended to link such efforts to the main sector databases as these are improved. That could save money.

To tackle the problem of water systems that don't work in Tigray, the Relief Society of Tigray (REST) has developed its Wahis Mai Maintenance Program with the help of the NGO Charity: Water and working closely with government. The target is for 93% of projects to be functional at any time, and broken water schemes to be repaired within 30 days says Zelealem Fisseha from REST. The program works at scale covering 30 woredas and 4,704 water schemes to date. When a scheme has a problem, messages can be sent in different ways: by sensors installed on 2,900 schemes (these are units directly fitted on handpumps measuirng water flow), via free calls to the office in Mekele, a direct call to the local Wahis Mai technician on their mobile phone or through data collected (also using by Akvo FLOW) during regular monitoring visits by technicians.

What is unique about this Tigray program is that water technicians have been assigned at the Tabiya/Kebele level. This is the lowest administrative unit and in most regions, apart from Tigray, there is no full-time water staffing at this level. At the woreda (district) level there are water woredas officers as in other regions, but a further difference in the programme is that at cluster level Wahis Mai units have been put in place to handle repairs that are beyond the capacity of the Kebele or woredas level government staff. The functionality rate for these schemes is already 93.8 percent. A key question is how to extend such systems to the whole region. At the moment, it only covers REST/Charity: Water schemes, even if these are rather numerous.

Participants at the Seminar

The innovation on monitoring-messages-maintenance in Ethiopia is encouraging. And there should be even more. We hope and expect to see more NGOs and more regional and local governments developing their systems and responses over the coming years as part of the GTPII effort. Some of the challenges that have been identified in doing this at scale are how to build the private sector in Ethiopia as part of the response, how to support frontline WASHCOs better, how to use the open data movement and share data based on agreed standards, and crucially, how to build those financial models to make mending things more attractive.

This blog post is based on the discussions at a WASH learning seminar held on 18 October 2016 at the Getfam Hotel, Addis Ababa. The seminar was organised by IRC in collaboration with the National WASH Coordination Office, the Ministry of Water, Irrigation and Energy, the USAID-funded Lowland WASH Activity, Charity: Water/ REST, and UNICEF. Presentations by MoWIE, REST and UNICEF/Akvo can be downloaded here. The presentation by SweetSense is available at the SweetSense website.

Disclaimer

At IRC we have strong opinions and we value honest and frank discussion, so you won't be surprised to hear that not all the opinions on this site represent our official policy.