The big data opportunity in the South African public sector – Zwelijongile Gwebityala, Richard Migwalla and Johann Myburgh

South African organisations have truly embraced the global boom in big data. Many have invested significantly to build big data capabilities that are already delivering value in terms of revenue growth, risk avoidance and efficiency. Banks, retailers and insurance companies are leveraging big data to implement ever more accurate digital marketing campaigns. When a consumer receives an advertisement SMS or a pop-up in their banking app or on a website, it is very likely that a “Next Best Product” model has informed the decision of the organisation to send the communication to offer that product. Recently, a few of the large banks in the country have announced that they are investing heavily in big data and hiring armies of number-savvy data scientists to staff their analytics departments. This is a strong indication of how bullish many large organisations in the country are about the potential value that big data can deliver.

The buzz and excitement in big data is unfortunately almost exclusively located in the private sector. In government, we have seen very little activity targeted at leveraging big data. This is understandable given the many challenges that beset all spheres of government, as well as the seemingly high price-tag required to truly derive value from big data. However, we would argue that investing in big data could unlock considerable value in the public sector and address the very challenges that prevent government from exploring the big data opportunity. We also argue that leveraging big data in the public sector is not as daunting as it first appears.

In our work in the public sector we have come across several easily implementable quick-win big data use cases that promise tremendous revenue and efficiency potential for the government. For instance, we assisted a mid-sized municipality to uncover almost R240 million in additional annual revenue by leveraging big data tools – including geospatial mapping, data wrangling and basic estimation models – to more accurately pinpoint business and residential consumers who were not paying their fair share of municipal rates and taxes. That municipality is now in the process of closing gaps in their revenue processes by getting those identified entities to pay their fair share.

In another case, we worked with a government entity to help them better plan the routes of their large fleet of service delivery vehicles. The improved route planning had the impact of greatly enhancing service delivery as the vehicles were able to get through their volumes of work in a short space of time. In addition, it resulted in massive savings in fuel and improved driver morale.

These use cases are just the ones that we have been involved with. There are ample opportunities in any government department and entity that one can contemplate. For example, procurement spend analytics could help the government greatly improve their negotiating power and thereby reduce procurement costs across departments; health analytics could greatly enhance the speed of detecting outbreaks and save the government a great deal of money in treatment; fraud analytics could help a number of departments to identify governance breaches more easily; big data tax audit models could help SARS address tax under-filing by auditing all tax payers instead of only a small sample; the list goes on.

Indeed, we see several global governments leverage big data even in developing countries. For example, the government of Burundi uses open data provided by both private and public sector health practitioners to track and advise on the country’s health-related indicators. In Columbia, where subsistence farming is the primary source of income and means of livelihood for farmers, climate data is collected, aggregated and processed by its International Centre for Tropical Agriculture, with predictive models applied to assist local farmers to better navigate volatile weather patterns.

The journey towards deploying big data does not have to be a daunting one. It also does not need to be one characterised by large upfront investments in systems and people. In fact, we strongly believe that the best approach to deploying data analytics is to take a gradual approach that delivers incremental results that enable future investments in big data to be self-funding. As a start, each government entity could select only one high-impact use case and devote resources to deliver on it. The selection of the right use case will prove the value of big data, but also address immediate and high-impact opportunities.

However, to begin the journey towards deploying big data, government must address one major stumbling block, and that is data availability. Government services across many departments tend to be paper-based, or where data is digitised, it is found in disparate systems with no golden source of data. This paucity of data renders the notion of big data almost impossible. This stumbling block is however not insurmountable. Advances in optical character recognition enable the fast digitisation of paper-based data; failing which, deploying armies of data capturers is also an option.

There is a reason local and global large organisations are making significant investments in big data and it would be a shame for the South African government to miss out.

Zwelijongile is a managing director at Ntiyiso Business Consulting, where Richard Migwalla is an Associate Partner and Johann Myburgh a Senior Data Scientist.

Read OTHER Articles

We would love to hear from you.