**15. Summary**

The euphoria created by the advent and exponential evolution of predictive analytics seems to have left many stakeholders in awe. From every level of business to different institutional categories, the best and optimal performance seems to be in sight with no establishment being left behind.

### *Enhancing Program Management with Predictive Analytics Algorithms (PAAs) DOI: http://dx.doi.org/10.5772/intechopen.98758*

While the positive outcomes achieved so far continue to escalate, institutions at large need to take one step backwards to do some stocktaking. This process involves asking critical and provocative questions, including: Are we doing the right thing? How evidence-based are our strategies? Are they sustainable? How reliable are our data sets? Is client data adequately protected from potential cybercriminals? Have all the ethical concerns been adequately addressed? What is the gold standard?

If PAAs' dynamics are any indication, the learning curve is bound to be long, steep, and daunting. One major reason for this possibility is the growing complexities of managing data and the institutions involved in processing them. There is also the challenge of establishing a diverse team of experts involved in developing problem solutions. Members of such a complementary group serve as an invaluable backbone to any potential success. The problems are complex, ranging from good quality data to the nuances that accompany risks and assumptions of selecting and applying the appropriate algorithms.

As already indicated elsewhere in this chapter, good quality data is *sine qua non* to any successful analysis (quantitative and qualitative). Mark Twain's characterization of lies, "lies, damned lies and statistics," should always serve as a compelling reminder that the information generated from data through the machine learning (ML) process is only as useful as the quality of data used. Having and using the appropriate and reliable piece of information is a catalyst for making informed decisions. PAAs are no exception! ML processes continue to gauge significant amounts of data. This data is transformed through the ML process to predictive outcomes (information) used in making informed decisions. ML's propensities to process big data sets have made cloud computing an inevitable requirement. The arrival of quantum computers (QC) has made the transformation process faster, reliable, and more efficient. These QCs, which have miniaturized the binary digit (bit), have moved computing to a higher level. According to an IBM definition, "Quantum computers, on the other hand, are based on qubits, which operate according to two key principles of quantum physics: superposition and entanglement. Superposition means that each qubit can represent both a 1 and a 0 at the same time." Access to good quality data is one way of optimizing the utilization of these QCs.

In one of my series of lectures given to graduate students at the University of the West Indies in Kingston, Jamaica, a student wanted to know why program managers firmly believe that in any strategic framework — "logframe" for example — outputs (and their indicators) always contribute to outcomes, especially given the potential for misleading and unreliable results reported at the output level.

In my response, I agreed with the student while elaborating on the data collection and reporting vulnerabilities, especially in environments where very little appreciation is given to data that are subsequently converted to information. I explained the trade-offs that managers and other stakeholders are faced with. I described what it takes to address issues like these, including conducting a control study. I further shared an anecdote with the group; an experience I had conducting a program evaluation for a UN agency. In this case, the agency had spent 4.5 million dollars over a three-year period on national capacity strengthening. The participants, who were medical health workers, were trained both nationally and internationally. This was identified as one of the output indicators that contributed to a corresponding relevant indicator — improved quality of health services — at the outcome result level. During the evaluation assignment, I followed up (something that was never done after training), and as it turned out, most of those who benefitted from the training had moved on; some changed ministries, others had left the country, and some had even changed

professions! Obviously, any planning decisions made using that training report would undoubtedly be erroneous, misleading, and deceptive at best.

It is quite conceivable that the evolving, inclusive, and streamlining dynamic of PAAs will continue to have positive and unquestionable consequences on how programs are managed. The myriad implications are unfathomable with synergies that collectively yield both intended and unintended outcomes. If current thematic applications are any indications, introducing analytics in any intervention will continue to be a win-win initiative.

While different institutions condition their interventions towards their respective strategies, the ultimate outcome is improved productivity and optimization of resource (human, financial, and material) utilization. There is also the human (quality of life) dimension that can revolutionize, reverse, and mitigate certain nuances that affect our wellbeing. For example, academic institutions now apply some models for improving student performance. By using historical data these institutions are able to identify vulnerable students, counsel them on an individual basis, and enable them to set more achievable objectives based on their academic performance with respect to the average group standing. The ultimate outcomes demonstrate counterfactuals that are obvious. And the results have been quite impressive. Some students in some cases have even encouraged themselves to become their own agents of change.

There is also gradually and increasingly, an inclusive element of analytics that continues to encourage and involve members of different community populations: crowdsourcing. This strategy has mushroomed and generated an astounding dynamic amongst communities. It remains to be seen to what extent the strategy will contribute to improving people's quality of life.

In general, business institutions are ahead of the curve with marketing as one of the trailblazers. The competition remains extensive and brutal.
