A robust analysis of available data is critical for the development of policies and strategies. However, many of our public sector clients lack the capacity to undertake this analysis internally and bringing in external consultants is a luxury few can afford. Where donors provide funding for analytical projects, the solution can be difficult for internal teams to sustain on an on-going basis. This is why one of our key project performance metrics is the degree to which an institution can build upon our work in our absence. Achieving this self-sufficiency, for us, signifies a successful engagement where we not only delivered value, but also left the client better equipped than before. To achieve this, we embed a philosophy of continuous learning and skills development throughout the duration of our projects – especially skills relating to data analytics. In this blog, Hanjo Odendaal, 71point4’s chief data scientist shares some of his insights and experiences on building internal analytics capacity.

Leveraging administrative and internal data is essential for driving informed decisions and delivering efficient, citizen-centric public services. It helps increase confidence and promotes better communication when governments need to build consensus and respond quickly to a situation.

Yet despite these clear benefits, public institutions in many countries have struggled to either develop and/or adopt advanced analytics capabilities. While there are notable exceptions – for example, in the healthcare sector where analytics is instrumental in predicting disease outbreaks and optimizing resource allocation – by and large, public institutions face significant challenges when it comes to developing analytics capabilities in-house and leveraging data. In some cases, these challenges arise from inadequate data infrastructure and limited funding. But in many cases, the primary barriers relate to people.

The first challenge relates to lack of data literacy throughout the organization. Often management lack a clear understanding of the potential of data analytics, and executives may not be aware of the technology and skills that are needed to produce the desired results. This creates a disconnect between business expectations and data team capabilities. This disconnect can lead to unrealistic expectations which in turn result in undue pressure on poorly equipped data teams and ultimately, a high probability of failure when it comes to data related projects.

This brings us to the second challenge hindering success: limited skills and competence of data teams. Many institutions struggle to assemble a team with the right mix of skills and expertise in data analytics, machine learning, data engineering and software design. This is partly due to a lack of awareness about the current data maturity in the organization, the importance of engineering skills to deliver quality input for analysis and the limited availability of experienced professionals in this field more broadly.

If it was easy to hire good engineers, data scientists and analysts, why do large organizations like Netflix and Spotify struggle to find talent? Despite their strong brands, sizeable budgets and extensive interviewing processes they still have bugs in their code.

With regard to skills and competence, quality does not scale easily. A good engineer/data scientist can do the job of twenty mediocre employees. A large but mediocre data team might never achieve success because it does not have the expertise to bridge the gap between sandbox and production. In contrast, talented data teams can solve the extraordinarily difficult problems that arise in the ‘real world’, outside the neatly curated setup found in most online courses or university classrooms. The ability to solve these idiosyncratic problems becomes key in determining whether a ‘data-driven’ project fails or succeeds. Rather than going big, we believe institutions should start small when embarking on their data journeys. Assemble a core team of experts. This dream team could include a principal data scientist with a strong background in engineering, machine learning, software design and strong business skills supported by a data scientist and data analyst. The role of the principal is to transfer knowledge, break through technical hurdles, mentor juniors and be the facilitator between the business and the data analytics team.

To attract and retain top talent, public institutions must offer competitive compensation packages that match industry standards. We have been involved in many interviews and recruiting processes that fail to secure the right candidates because fixed pay scales in the public sector are simply uncompetitive. After months of searching, public institutions might identify the right candidate, but the candidate rejects the offer, and the institution is left to pick up the private sector’s leftovers. It’s time public organizations rethink their compensation models for specialized teams to attract and keep the best and brightest in data analytics. 

In summary

Data analytics has the potential to revolutionize policymaking and public service delivery in Africa. As more public institutions start to recognize this, it is likely that there will be increased investment in building internal capacity. This investment must take its lead from a well-articulated vision and coherent strategy to integrate data analytics into the decision-making framework of the organization.

This will require that executives;

(1) develop a strong understanding of the analytical potential of the data collected and maintained by the institution,

(2) undertake an honest assessment of their internal data team’s skills and capabilities,

(3) invest in human-capital alongside fit-for-purpose infrastructure that act as strong catalysts for the data strategy, and

(4) create a sound monitoring and evaluation framework that balances accountability on the one hand with individual growth and innovation on the other.

It is imperative for public institutions to adopt a long-term vision of integrating data analytics within their organizational framework. The successful implementation of such a data-driven strategy (although not always easy) will almost certainly guarantee success in navigating the complexities of the modern-day data-rich environment. By doing so, public institutions will become more effective in achieving their ultimate purpose – designing and implementing citizen-centric policies.