By Ankit Mittal, Senior Project Manager, REI Systems, Inc.
Business Intelligence (BI) is designed to answer a simple question: How is the organization doing? It takes into account both what has been done in the past , as well as current operations and future aspirations. While the question may be simple, getting to the answer(s) can be difficult. Most Federal agencies are already collecting enormous amounts of data, but linking current with historical data is not always easy. Furthermore, the timeframes represented by the word “current” can vary: Is it one month? One week? One minute (real time)? Who has access to it and what can they do with it? Finally, agencies are seeking to improve their raw data analytics capabilities with advancements in artificial intelligence, machine learning, and so on.
This task is even more challenging with respect to grants, because some data comes from the grant-making agency, while other data comes from a variety of grant receiving entities. Given the focus on results-oriented accountability for grants under the President’s Management Agenda, CAP goal 8, agencies are increasingly determining they need to modernize their BI or analytics capability. They are faced with some key questions: Where do they begin? How do they frame what needs to happen? How far can they go? In this article, we’re going to take a look at the main drivers for BI initiatives, a model for self-assessment, and a framework agencies can use to design a modern analytics platform (note that several suggestions here are drawn from experience developing and using HHS/HRSA’s award winning New Data Analytics Platform).
Common Data Analytics Problems
When it comes to data analytics, agencies commonly face four problems:
In a typical IT shop, when funding and personnel are constrained, analytics don’t always get priority. When a few specialized resources do exist, staff are dependent on them to get data or generate insights and are therefore unable to move the needle on their own. This can have important implications for end-of year reporting, increasing the needed turnaround time.
Another common problem is that while rich sources of data exist, they are usually spread out across multiple source systems. Aggregation of this data is a manual, slow process that increases turnaround time for generating insights. As applications and media channels diversify, the number of data sources also increases, compounding the problem. The resulting manual data analysis and report generation tasks compound the effort required to find actionable value in the data.
Programs that seek to improve their outcomes or expand their effectiveness are driven by guidelines, concepts, or theories. Therefore, they need an effective mechanism to compare agency mission with program outcomes and organizations' performance or to generate trends and compare outcomes between one component or grantee and another. This application of data-driven insights to policy necessitates a strong BI backbone.
Another area with which many agencies struggle is risk management. When the data exists in silos, risk identification is also conducted in those silos. For example, if one program has identified a grantee as a high risk for poor compliance or a lack of financial controls, the data about that grantee or the corrective action taken is not available to other programs that might award another grant to the same entity. Further, the finance functions at the grant-making agency may have important information, like specific grantee organization contacts marked for exclusion. This information is difficult to distribute or share with the multiple grant issuing program offices. Such challenges represent a barrier to combatting fraud, abuse, and waste activities at the agency level.
Addressing the Problem: Self Assessment
The first step towards modernizing BI is to conduct a current state self-assessment across several dimensions. One model that has worked well is the Federal Data Maturity Model developed by the US Department of Commerce’s National Technical Information Service (https://www.ntis.gov/). This model helps identify current capabilities and conceptualize where to head in the long term. It also provides a common language to advance solutions and best practices amongst agencies.
Building a Data Strategy
To build a roadmap for BI modernization, organizations should first build a robust data strategy which brings together six components that operate effectively within a data-driven culture.
Together, these components form a roadmap to help the organization modernize its BI, while working together to help create a mindset shift and influence the data-driven culture of the agency.