An enumerator with Ugandan Water Project collects a water sample in a school outside of Kampala, Uganda

Co-founder at mWater

Solstice is an operating system for governance in low-resource regions. As what gets measured gets done, Solstice approaches the task of eradicating global poverty with the ability to measure better and more meaningful data.

Six years ago, our small team piloted an approach to support infrastructure in developing countries. We began with a platform focused on the water and sanitation sector, called mWater. Local governments, utilities, non-profit organizations, health workers, and researchers all used our mobile app to map water sources and sanitation facilities and then monitor them over time with mobile surveys. We started with a small contract from UN Habitat in Mwanza, Tanzania and grew to three East African countries that first year. The next year, with the backing of two large NGOs, Water.org and WaterAid, we expanded to 27 countries. The following years, we had an organic userbase that grew to 54 countries in 2014, then 73 in 2015, and 100 in 2016. Today, we have a user base that spans 177 countries.

What began as a small pilot has become a scaled success. Each month, mWater users map and update about 150,000 sites and submit .over 400,000 surveys. Now, with Solstice, we are expanding to other sectors within aid beginning with health, education, agriculture, and other humanitarian topics adjacent to water. What was an infrastructure management platform for water and sanitation is growing into an operating system for sustainable governance.

The first datapoint humans measured and tracked over time to better understand their world and their health and safety was the solstice. Twice a year, in each hemisphere, the sun reaches its lowest or highest point in the sky, and its fewest or most hours of shining. That day is called the solstice. Humans tracked this to know where they were in the year, when the growing season would return, and how many cycles of the sun’s ebb and flow had passed over time. We name our new platform Solstice to honor this global human tradition of using data for better understanding of ourselves and those around us.

It began with water

We began with a focus on water because it is, in our professional experience, the most important cornerstone to making progress in every sector of aid. Unsafe water is the second leading cause of death for children worldwide. And yet, those who do not die often survive with lifelong disability from physical and mental stunting. Malnutrition’s single greatest cause, even more so than food supply shocks, is unsafe water. Children who suffer repeated bouts of diarrheal disease from infancy through childhood acquire a condition called enteropathy in which the intestines are so inflamed from infection that they cannot absorb nutrients. Children mentally and/or physically stunted are disabled for life. They are less able to learn in school. They are less able to do physical and mental work as adults. Women who suffer stunting under age two are more likely to suffer complications and death from pregnancy and birth. The stunted body is more vulnerable to all causes of disease, including malaria, tuberculosis, and HIV.

The Millennium Development goals set a global target for water, as well as other measures of international development in 1990. The water goal, which was subtarget C on goal 7, focused on eliminating the burden of the long walk for water. It aimed to double the world’s access to nearby improved water sources by 2015. It was one of the few Millennium Development Goals that was reached as aid agencies and governments around the world created millions of improved water sources. Unfortunately, during this time, the incidence of diarrheal disease did not significantly decrease. The world’s burden from unsafe water persisted because people were not dying from walking for water, they were dying from fecal contamination in their water.

Two workers for The Water Trust in Uganda map water sources and track their safety and function with mobile surveys.

It is a lesson in the management rule: what gets measured gets done. The target was made to be achievable because the global community did not believe measuring water quality was possible in 1990. Expensive labs and highly educated staff to run them were not available in the regions with unsafe water. However, by 2015, remote water testing methods were developed that made it financially and logistically possible to test water anywhere. The management mistake was righted with the new and improved set of global goals, the Sustainable Development Goals, established in 2015. The SDGs aim to achieve global access to “safe and sustainably managed water systems” by 2035.

mWater was designed for this task. By mapping the infrastructure of water sources and water systems, they became manageable at the local level with mobile surveys. The data collected was digital and stored in a relational database that made it comparable across time and geography, and most importantly, across NGOs and governments.

Previously, maps of water sources and surveys about them were kept on paper and, even if entered into Excel files, locked away on computer hard drives. mWater’s open access database was designed for safe collaboration. Organizations could label water sources they managed and make them openly viewable to the governments and other organizations around them. The advantage to keeping data private was once donor competition. Organizations and governments used to keep data private for fear of losing funding in a fiercely competitive donor market. However, in recent years, the donor pressure shifted toward open data and all NGOs and governments are encouraged, if not mandated to make their data public.

It began with the survey

In low-resource countries, surveys are the engine of governance. The farthest reach out from the government to the communities is the health worker. This person may be specialized in maternal health or vaccines or agriculture and checks on families in her or his catchment regularly to monitor how they are doing and administer education and other support. They do their job by conducting a survey. The survey may instruct the health worker to check the children’s weight, to see if the newborn’s breastfeeding latch is correct, to ask if the parents are saving one chicken egg a week for each child to ensure protein in their diets.

A government health office wall is filled with posters of compiled surveys on child health in Tigray, Ethiopia

The health workers conduct these surveys every day, typically on paper, and turn them into a supervisor who then compiles a report to summarize all the surveys turned in every month or so. These supervisors traditionally kept the compilation tracking on their walls on large paper with charts of key metrics. Regularly, often yearly but sometimes more often, they would report the full compilation to the central government.

The Veterans Administration data storage room in Arlington, Virginia. We use this image on our opening slide for trainings: paper is where data goes to die.

For the 73 years of the international humanitarian movement, this is how the United Nations, The World Bank, and all NGOs have instructed governments to set up. For 74 years, the surveys and compilations were conducted on paper and piled on desks and the work was often done very well but poverty persisted because paper is where data goes to die.

The first advance that mWater, and now Solstice offer is to move from paper to digital data collection. Enumerators ranging from health workers to NGO representatives to local researchers can now conduct their surveys on a mobile phone or tablet in the Surveyor app, which syncs through cloud computing with the data management portal, making data visible in real time to managers, government overseers, funders, and citizens all over the world. This has the revolutionary impact of making the data collected immediately actionable. With a simple survey, the collective work of the private, public, and NGO sectors can work together like a mesh network to amplify the findings and agendas of all stakeholders. The ultimate result is a catalyzing of the pace and achievements of everyone’s efforts toward achieving sustainably managed water and health systems.

Management begins with data

The process of collecting data no longer results in piles of paper on desks and walls of charts that are up to a year old. Instead, it is collected, collated, analyzed, and responded to in real time. Data collection becomes triage. Progress toward long-term targets can be monitored alongside red flag issues that need immediate attention. Government policy making becomes a data-driven activity that can be nimble to changing conditions.

Up until now, data collection mostly happened under an umbrella of “M&E”, or monitoring and evaluation. Data collection for M&E was demand driven by donors, who would not agree to fund or re-fund projects without the provision of recipients’ compliance with collected data toward their projected outputs. There are two weaknesses in this model. First, outputs are collected, analyzed, and reported at the end of a “project,” an aid effort encapsulated by a start and end. Errors in planning and adjustments for unforeseen events are not possible mid-stream here. Second, a project’s success or failure is solely measured by the donor’s decision to re-fund. This is a perverse economic model, in which the recipient of the product has nothing to do with the judgement of the success of the product.

An allegory for outputs-driven, M&E-focused aid is a hypothetical market that sells apples. If the apples turn very fast and cannot be eaten but rather thrown out, the customer would be unhappy with the product and stop purchasing from the market. The market would lose its source of funds to continue operating. If the market received funding based on its provision of apples, rather than customers’ satisfaction and returned business, it would not matter to the market whether the apples were enjoyed by the users. It would report to the funder how many apples were distributed at the close of its business cycle. In a healthy economic model, the funding output depends on a closed loop from recipient to product provider. The donor’s traditional measurement of aid success has had the impact of breaking that loop, and the outcome of ineffective aid.

Ongoing data collection focuses on the big picture of governance outside of a project model. It incorporates a “metrics that matter” focus that gives way to a management practice built on end-users and realtime measures of impact and problems that need to be addressed. It allows for a micro-focus on data and a meta-focus on data that the old fashioned baseline/endline approach precluded out of limitations on time, cost, and feasibility.

Next Billion Technology

The Surveyor app is a native app on the Android and iOS platforms, and a web app that works in a computer’s, phone’s, or tablet’s internet browser. The data management Portal is accessible on any browser. This technology has been built as a “Next Billion Technology.” In theory, the first billion people to embrace the internet in their personal and work lives were the easiest to onboard because they had plenty of access to high-quality devices and well-functioning internet. The next billion require more focus on spanning a wider range of quality in both devices (mobiles and computers) and internet.

A government stakeholder meeting about data collection needs in Cambodia.

As a Next Billion Technology, this platform is designed for more rugged internet conditions. It comprises an app for mobile data collection, called Surveyor, and a central data management Portal, where all surveys are designed and deployed and data coming it can be cleaned, organized, analyzed, and understood with dashboards, maps, spreadsheets, pivot tables, and more. The Surveyor app works on and offline, syncs with the cloud as soon as it has a good enough connection, and always looks for that connection as long as it is open. The Portal does require an internet connection but can handle the intermittencies in connectivity and fluxes in broadband strength of signal that are common for much of the world.

The design also maintains a low level of required fluency for interacting with the platform. Already translated into 9 languages (soon to be 12), the app makes use of symbols and images to reduce the reading fluency required for collecting data. For technical fluency, the app borrows many design features of Facebook Lite and Whatsapp, the world’s two most used apps. This means that any users new to data mobile collection but familiar with one of these apps can quickly begin working; and users completely new to smartphones or tablets can be encouraged to learn by playing with apps that are fun and interesting to them. We typically onboard enumerators new to smartphones within 1.5 days and train enumerators to conduct sound data collection with rigorous ethical and methodological adherence in another 1.5 days.

An mWater training conducted by The Water Project in Kakamega, Kenya.

The Portal has a bit higher threshold for fluency required. Generally, most features can be used with the same level of knowledge required to use Google Docs, also among the most common platforms used in low-resource regions’ governments and NGOs. The first generation of digital survey technology on the market required coding to design surveys and manage their deployments, but this platform is designed in a “What you see is what you get” interface, like Google docs, forms, and spreadsheets, allowing users to work with no knowledge of coding at all. This reduces the cost in terms of personnel needed to manage with digital data. The portal has many templates built in for designing surveys, analyzing results, and creating visualizations of the data collected so that users can duplicate a best practice approach with a click of one button and not need advanced experience or education in data analytics or data visualizations to begin collecting data and understanding it once results come in.

Web 2.0 for International Development

The Web 2.0 movement was a revolution that began with the internet. Around the time of 1997, blogs were born, and with them, the internet went from a one-way communication platform (authority → recipient) to a two-way communication platform in which comments and chat backs and, soon, social networks meant all flows of information were at worst two way, and most commonly multi-directional. By the year 2000, the phenomenon was hitting its tipping point and it initiated a series of revolutions in all other sectors of the economy and social life. The revolution had four major influences: flatter (as in, fewer structures of hierarchy between boss and lowest personnel), porous (as in, flows of information now easily moved in and out of an organization’s or company’s walls, leaner (scale does not require capital), and agile (quick adjustments to circumstance equal survival, if not advancement against competitors).

The revolution is perhaps easiest to see in the media industry. After Web 2.0, every city’s newspaper does not need a movie reviewer to cover the same movie the whole country just received at a local theater. Only one or two great reviewers are needed and a relationship for republishing them nation-wide. And a layperson not paid to review movies by any newspaper might publish their own review that rivals the quality of the professionals. Forgetting the sadness of firing small town movie reviewers, it is an efficiency for the industry to save costs and maximize scale. The public benefits from more choices, more involvement, and lower costs.

Other industries hit early by the Web 2.0 revolution were education, banking, and commerce. Finally, this has come to the aid industry. Aid is perhaps the one industry that should welcome the cut in personnel and operating costs, because, if the aid industry is living true to its ideals, it should be welcoming its own exit strategy. Aid uniquely operates in an economy where should is a pressure point. Aid should want to go out of business if local governance become so efficient and independently managed that outside aid is no longer needed for development. This is the foundational goal mWater embraced from its beginnings: put ourselves out of business. Solstice helps end aid by creating management ecosystems where aid is no longer needed.

mWater’s business strategy is borrowed from the tech industry to take advantage of Web 2.0 and the efficiency revolution. First, we charge for services at the top of the user chain and make the product free to the end users. In technology, the value of a platform increases with more eyeballs seeing it. The free end user model maximizes our potential to have the most possible users. Just as end users do not pay for Google or Facebook, but rather those who benefit from them using the product pay for it, we structured mWater in such a way as to charge the beneficiaries of a widely used, broadly applicable platform and left it free to the end user.

Kiki Tazkiya from Water.org conducts a household microloan survey in Indonesia.

In this case, our primary clients are large NGOs and multilateral institutions (UN divisions, WHO, and development banks) that would have to pay experts to build in-house, bespoke software solutions to manage their data. We encourage them to invest in this platform, paying us to build what they need, and we write every contract to say that what we build will be made free to the world. This is possible because the cost in technology is in research and design; once a new feature is built, it can scale at no cost.

This open access model and free to the end user cost structure has three disruptive impacts on the aid industry. First, it solves the dependency cycle crisis in aid that starts when a large entity partners with a local community or government to initiate a new approach or strategy. If it is based on paying for the platform the work is building from, then after the NGO or multilateral leaves the local entity is left with a need to borrow money or find additional aid to keep the activity going. If the work undertaken is managed with Solstice, there is no ongoing operating cost.

Second, it lowers the barrier for low-resource entities to try out approaches to management on their own, without outside involvement of aid. It makes possible the opportunity to experiment, to play with the platform and fail without cost. This has resulted in locally created and easily sustained approaches to management of new initiatives. In many cases, locally-based users have been the ones to introduce our platform to larger institutions, suggesting Solstice to them when planning new work.

Finally, this platform is a data management software built to make collaboration natural and safe. It means no aid organization has a reason to own or protect public data from the local government and citizens they serve. It is, in effect, an operating system for governance that aid can plug into and out of without disrupting the government’s management strategy. A central theme in all of the Sustainable Development Goals is the target of sustainably managed systems. In the case of creating strong government systems, sustainable means the work can continue without the influence of the aid industry. We designed Solstice to always build work with the exit strategy in mind. It is the operating system for sustainably managed governance.

Advances in data for governance

The aid industry has been managed from the paradigm of academic departments within a university. The sectors that have grown in these past 70 years mirror the silos of research: water, maternal health, child health, agriculture, infectious disease, etc. It often means that a worker in one sector of aid can spend their entire careers without interaction in another sector. Donors have encouraged the silo effect with narrow missions that do not bleed out of their focus. The problem is sustainably managed governments do not work in silos.

The advances of technology since about 2009, specifically, relational databases, mean that collecting data does not need to follow the silo structure. Data collected ostensibly for one purpose can be made available at the indicator level, if not at the level of raw data, to any other sector. This means they can all work together. Solstice’s Global Indicator Library is a Wikipedia-like repository of indicators with well-researched question sets that are wired up to them. Using indicators to design data collection means any one actor can collect data that is comparable with any other actor using that indicator. The end result is surveys focused on individual need, but that reflect the larger condition of a community, regions, or country beyond the individual actor.

A government field worker conducts a survey at a borehole in Malawi

Take an example: a maternal health worker can collect data from a family on the Surveyor app focussing on their sector, but be triggered to ask questions and collect data about agricultural stability if a separate branch of the government’s health workers focused on crop outcomes collect data indicating a new risk is present. In traditional aid management, it takes an average of two years to pick up on a food shock so bad that children are taken out of school to support agricultural work. With cloud-based data, systems communicate with systems naturally. Algorithms can create evolutions in data collection without humans slowing down the process. Governments can support families long before things get so bad that they take their children out of school.

Another example: a recent study found that local health clinics have, on average, among the lowest rates of access to safe water than any other entity in communities. This is very dangerous because, among other reasons, unsafe water and sanitation conditions in health facilities contribute to maternal mortality. This problem is a legacy of silo-based aid — that the outcomes measured for health clinics were typically health-related and water-related outcomes were only measured by water-focused aid, which of course did not know about health. The two sectors did not share conferences, publications, or employers, so no shared data collection was planned. Now, if both sectors are using Solstice, they do not have to work together in the real world to collaborate in the cloud. Safe water in health facilities is one of the fastest growing targets in the Solstice platform, where the WHO has built a stand-alone interface for health facilities to easily measure indicators that integrate health and safe water. What gets measured gets done, and making new ways to measure outside of silos means managing outside of silos is getting more done.

The unwritten goal

The Sustainable Development Goals, like the Millennium Development Goals before them were created to focus the work of independent actors worldwide on achieving the most important benchmarks that are thought to lead to the fastest and most efficient path toward eliminating global poverty and the suffering associated with it. Solstice catalyzes this effort by establishing an operating system for this work that makes data comparable and meaningful. It makes collaboration natural. It makes policy-making with data-driven evidence possible for managers and authorities in low-resource regions. It makes the independent work of many small enterprises come together to a larger product than any one actor could have achieved alone.

mWater, the non-profit organization that created Solstice, is not an aid agency, but rather an enterprise built to improve and support the work of aid, communities, and governments around the world. NGOs, multilaterals, and governments are now firmly on board. In 2017, we doubled the user base it took four years to create. We hope for a future where we can also work with donors. We believe we make possible the ultimate, if unwritten goal: the end of aid. It is for this goal that we created Solstice.

Thanks to John Feighery.

Introducing Solstice: a technology platform built to eradicate global poverty