The scope of data and its impact potential is growing; therefore, data is a strategic asset for organisations. Data volume is growing exponentially, mainly driven by the explosion in Internet Of Things usage and digitisation.
The vast increase in data volume and complexity of data requires computational power and infrastructure to analyse and access it. Both data and computational power enable machine learning algorithm execution.
Machine learning is a term that encompasses a range of algorithms from statistical methods like regressions to the neural network and deep learning. To harness the power of data and implement successful algorithms, the organisation should adopt an insight-driven value chain.
An insight-driven value chain consists of series of components. These components of the insight value chain are arranged horizontally and vertically in an organisation set up to create desired value. The Insight value chain components are classified as technical foundations and business foundations.
The strategy and vision-driven operating model encompassing the data operating model determines the success and failure of the data insight value chain. A systematic approach from identifying business needs through case scaling and rollout translates data insights into business value.
Business value is derived by identifying business needs through use case scaling and translating data insights into business value. Raw data rarely holds business value. Raw data needs to be cleaned and transformed to derive value.
Most of the time, people are focused on the execution and adoption of advanced technologies to deliver business value. However, a series of links in the insight value chain produce business value. The strength of a chain in a value chain is as good as its weakest component.
Therefore, capturing value from data requires excellence in all components (Refer: Figure 1) of the value chain. The components are made up of technical foundation and business foundation. The technical foundation is made of several components such as data, analytics, and IT infrastructure.
The links of the data component in the insight value chain are data sources, orchestration of data, unstructured data, privacy, and data security. The relations of analytics component are descriptive, predictive, machine learning, cognitive and optimisation. The links of IT components are cloud sourcing, horizontal scaling, analytical program language and data visualisation.
The business foundation component is made of components such as people and processes. The links of people component are cultural change, data-enabled decision making, roles and responsibility profiles and organisations. The relations of processes component are an adaption of new business process, automation of the business process, data and analytics governance, cross-functionality, and agile processes.
Deriving Business Value from Insight Value Chain
Deriving business value is a three-step process. These processes are generating and collecting relevant data and turning insights into action. The process of deriving business value from data is a three-step process.
First Step: Generate and collect data The process consists of data extraction, transformation, and loading. It also consists of appending external data sources and redefining data by using data mining. Predictive analytics is performed to support decision making and prescriptive analytics to drive value creation.
Second Step: Extract Insights from Data The steps involved in turning insights from data are process design and automated execution of machine learning models.
Third Step: Turning insights into Action: Once important insights have been extracted from the models, the next important step is to turn these insights into action and generate business impact. Man plus machines enable insight-driven action. Typical use cases for key based on business values are top-line use cases and bottom-line use cases. Top-line use cases typically include pricing, churn prevention, cross-selling, up-selling, and promotion open to driving growth.
Bottom-line use cases employ data-driven insights to optimise internal processes. Predictive maintenance, supply chain and fraud prevention are among the processes that can be improved with data benefit. The new business model is a data-enabled use case that moves beyond the process to improve data. A new business model requires data to be opened for external partnership and generate new sources of revenue.
There are three most critical structural challenges block maximising business impact from data. These challenges are:
The separation of data and business: Data science, data department, and business executions are separate in many organisations. This leads to a lack of understanding from the business side, possibly developing data science solutions that the business does not need.
The gap between insights and impact: Many times, decision-makers are ignorant about using the data. The data literacy programs and structured communication fills the gap.
No proper anchoring of data analytics at a corporate level: Business impact from data-derived insights only happens when data analytics is implemented deep within and consistently throughout the organisation. This requires commitment and direction from authority.
The challenges can be addressed by having proper tools, talent, and culture. Analysis should be used as a tool. Data translation is important, and change management is crucial, so organisation rules such as data product management, steward and governance are critical for the success of data programs.