We would like to keep you up to date with the latest news from Digitalisation World by sending you push notifications.
An increasing number of companies are embarking on their digital transformation (DX) journey, with global spending on DX predicted by IDC to reach $3.4 trillion by 2026. At the core of digital transformation is empowering data-driven decisions. However, as new architectures and processes are implemented, some companies are concerned that they will become data-dictated, that moving away from human-in-the-loop decision-making is both a logical and unavoidable consequence of implementing digital solutions. However, this fear is unfounded.
Data-informed vs data-dictated
While there are undeniable benefits to derive in moving from pure insight to decision-engineering, there are clear differences between the two concepts and questions exist over the role, or lack of, that humans play in each scenario. In the case of data-informed, humans remain in the loop to make decisions and take the appropriate actions based on data and analytics, whereas data-dictated refers to applications executing programmatic actions automatically in response to some stimulus or event.
Instead of just two options, to support decision-making, today’s enterprises typically implement four distinct categories of analytics: panoramic, predictive, prescriptive, and programmatic. Depending on the use case and the organisation, each of these types of analytics provide companies with immense value.
Data as a foundation for decision-making
Panoramic implies using analytics to provide a business with a real-time, accurate and expansive view of what is happening inside and outside the organisation. A 360-degree view of enterprise data delivers a range of benefits, key to which is improved decision support. In turn, this leads to better risk management, accelerated innovation, response to external events, and increased revenue among other clear benefits. A crucial component of enabling a 360-degree view is to incorporate more data from more and varied sources, providing a more complete and comprehensive view of the business
and more insightful analytics. The process of generating a panoramic view of an enterprise can assist in recognising and bridging departmental, application, or location-based silos. Through offering a transparent and unified view across a business, thereby generating situational awareness, it becomes easier to solve common challenges and orchestrate innovation and growth.
Predictive means using analytics to determine the probability of an event occurring, and therefore make timely decisions to maximise or minimise its potential business impact. For example, financial services firms can benefit enormously from using their own data, integrated with external sources, to generate signals focused on future outcomes such as changes in interest rates and how they will affect risk and profitability.
While this modelling is commonplace in the financial industry, predictive analytics have also found a natural home in the manufacturing industry. Developments in connectivity, sensor cost, size and technology, allied to advances in application processing speed and scope, mean that it is practical and cost effective to measure much more than before. In parallel with advances in remote connectivity and diagnostics, manufacturers are utilising Artificial intelligence (AI) and Machine Learning (ML) algorithms that leverage harmonised organisation-wide data to provide predictive maintenance, predict manufacturing quality and overall equipment effectiveness (OEE), and identify potential problems in advance so that they can be prevented.
Prescriptive involves the use of data and analytics to suggest the most appropriate actions for managers to perform, based on the probability of an event occurring, or on what is already happening. Prescriptive decision-making is a human-in-the-loop process, where the analysis of historic and current events, actions, and outcomes provides experienced individuals with a better foundation for making more accurate and optimal decisions. This mode of decision-making is not about replacing the human thought process, it’s about augmenting it with insights that, in combination with a person’s expertise, will deliver better results. This is currently an essential component within the supply chain industry. Given ongoing uncertainty facing supply chains globally, planners need to be able to rapidly understand how unfolding events will disrupt existing chains, and consequently use their domain expertise to act and react in the moment to minimise impact, mitigate risks, and reduce costs.
Programmatic is about providing the ability to execute in real time, based on predictive and prescriptive analytics when there is no time for human intervention. Within the financial sphere, programmatic decisions are employed for fraud prevention, pre-trade analytics, trading, and customer next-best action. Programmatic actions can also be deployed in use cases when there is simply no need for a human to be in the loop, which allows the organisation to streamline operations and
improve productivity. For example, in the logistics sector, programmatic decisions can be implemented by analysing unstructured video data of traffic congestion combined with weather predictions to re-route drivers to optimise on time and fuel costs.
Humans remain a key element in decision-making
Consequently, rather than moving away from a data-informed (human in the loop) to data-dictated (no human in the loop) state, firms are instead opting for the pragmatic application of most or all of these four Ps of analytics. This use of analytics gives firms the capabilities needed to improve business operations in a variety of ways. They can gain a current and comprehensive view of what is happening in the moment, and are better prepared for and swifter at managing events and disruptions, enabling the flexibility to take advantage of new opportunities as they present themselves.
Distributing the data to those who need it
Digital transformation is the enabler of informed decision-making, and firms are increasingly implementing a smart data fabric architecture to enable these capabilities. The smart data fabric is a next-generation data architecture that gives access to and orchestrates data from systems and silos both within and external to the organisation in real time, giving users a complete and transparent view of the organisation, its operations, customers, as well as a view on external data sources which have a potential impact on operations.
Embedded within the fabric is a set of services that allow users to explore, interrogate, and deeply understand the data, allowing them to drill down to greater resolution when required. A key element of this smart data fabric is that it does not rely on a continuous stream of requests and tickets to the IT team, thereby directly empowering employees to discover insights by themselves.
With the application of powerful analytics, enabled by the data fabric, the 4 Ps of decision-making form a basis for humans to take action, and for applications to optionally execute programmatic intelligent actions. Some decisions are supported by AI and ML, while other time-sensitive decisions are best made programmatically. Ultimately, modern data architectures and technologies offer businesses a palette of options, with human-only at one end, and computer-only at the other. Finding the right blend of decision-making and workflows depends upon the questions being asked, and determining this is reassuringly still a human decision!