Many governments make decisions based on poor-quality evidence

Government officials frequently make important decisions which impact the livelihood and wellbeing of their constituents. Despite this, many officials are unable to base their decisions in evidence.

In our experience, officials understand the benefits of evidence-based decision making yet lack good systems to collect that evidence. For instance, a government official from a West African State recently explained to us how he has no system to understand what is happening in the school he can see from his third floor office building, let alone schools in the far reaches of the state. Without good information on schools, this official is unable to pinpoint the challenges, implement solutions, or effectively hold people accountable.


Digital data systems are cheaper, faster and more accurate than traditional methods

Due to recent advancements in technology, digital data and monitoring systems are the obvious choice for governments. Driven by the low cost of scaling, digital data systems are relatively cheap to rollout. This is especially so when considering the cost of the monitoring system relative to the cost of providing the service which needs monitoring. For example the annual cost of delivering an immunization program can exceed one hundred million dollars, the cost of a new data and monitoring system is merely a fraction of this amount.

Not only are digital systems cheaper, significant value comes from their ability to rapidly collect highly accurate and validated data.

Digital data systems have contributed to rapid transformations

Some countries, including Peru and Minas Gerais (Brazil) have developed digital data and monitoring systems which give officials and managers close to real-time insight into their systems.

In Peru, independent monitors visited a large sample of schools monthly to collect data on basic inputs such as infrastructure and student and faculty attendance, which was made publicly available online. This system contributed to improvements in head teacher attendance from 85% to 95% in less than two years.

 Percentage of students in classrooms by province in Peru , November 2017

Percentage of students in classrooms by province in Peru , November 2017

In Minas Gerais, monitors collected data on the performance of each region, school and student. The data was compiled in an online database which improved accountability and supported decision making, contributing to a 37 percentage point increase in 8th grade reading levels in four years.

Collecting the right data is critical

The most effective data systems are developed from the ground up; designed to solve specific challenges. For example in Punjab, Pakistan, the government had no way of knowing whether their immunisation outreach team was actually going into the field and immunising children. To solve this the government, with technical support from Acasus, designed a system which required individual vaccinators to report their GPS location on a low-cost smartphone each time they performed a vaccination. The data from the phones meant that the government knew in real time exactly which communities were being visited and which were not. This system contributed to an increase in immunisation coverage of more than 35 percentage points in less than three years.

The right data to collect depends on the problem that needs to be solved. Different problems will require different insights. High level data can be useful to understand system-level insights, for instance population trends. However this level of data provides limited support for day-to-day management.

To support implementation, highly-frequent granular data is required to uncover the deep insights required to manage better. This level of data helps government officials to better understand the effects of new initiatives and identify ‘bright spots’ and poor performance.


The power of granular data systems

Granular data systems provide officials with the insights they need to make better decisions and improve the livelihood and wellbeing of their constituents.

Granular data systems provide insight into:

  1. Cause and effect: granular data allows officials to see the impact of events in the system. This allows officials to test new programmes and initiatives, quickly and accurately understand their impact, and decide whether to carry on, modify the strategy, or stop.
  2. Bright spots: most systems have areas of high performance somewhere, the challenge is finding and understanding them. Granular data systems allow officials to identify these areas. Once identified, officials can learn from them, and spread the best practices to other parts of the system.
  3. Poor performance: most systems have areas which are yet to achieve their full potential. Granular data systems allow officials to identify these challenge areas. Once identified, support can be targeted to understand and address the challenges.

About the Author(s) 

Will Anderson is the lead researcher for public-sector projects at Acasus.