Why would an organization refuse to buy into practices that would benefit them in the long run? In the modern digital age; more and more people are being connected to the internet and or other technologies that allow the generation of data. As a result of this growth in data production, analytics of these large datasets has become a critical success factor for organizations. After discussing the background of Big Data, the purpose of this essay is to discuss the why organizations are failing or struggling to adopt big data. This essay will also discuss the adoption of big data by considering the attributes of innovation theory. The stages of the diffusion of innovation model will be used to describe the adoption of big data and recommendations around the effective adoption of big data will be made.
What is Big Data
Big data refers to datasets that are large and unstructured as well as unsuitable for analysis through the use of standard spreadsheet and or relational database technologies. (Atkinson, 2014) What this essentially translates to is that big data is data that exists in unstructured and often unprocessed data that is so large and voluminous, that traditional data processing software is not enough to process it to render usable data for decision making in an organization. Following the growth of internet services such as Facebook, YouTube and others, organizations came to the understanding that users generated quiet a significant amount of data. This realization lead to the creation of various utilities and tools (such as Hadoop) specially designed to store and facilitate the analysis of big data datasets. (Oracle, 2017)
An expanded definition of Big Data is explained by explaining the 7 v’s of big data. The 7 v’s of big data are volume, velocity, variety, veracity, validity, volatility and value. In short; volume refers to the size of the data that is received from multiple sources, velocity refers to the rate at which the data grows and thus becomes more and more difficult to work with. Variety refers to the many types of data available, veracity refers to the accuracy and truthfulness of the data. The validity of data refers to “the correctness and accuracy of data with regard to the intended usage…” Volatility of data Data Volatility refers to the way in which Big Data is stored. The value of data is the cumulative result of the previously mentioned V’s as the purpose of Big Data analytics is to derive some sort of value from the analysis of the datasets. (Khan, et al., 2014)
Big Data Adoption Challenges
Although Big Data analytics has become a beneficial component of business operations, the fact remains that many organizations have still not successfully adopted it. The reasoning for this among many others is the fact that both large and small businesses encounter a series of challenges that inhibit successful adoption.
A major challenged posed by the prospective adoption of Big Data is interoperability. The reality in many organizations is that many organizations have already existing Business Intelligence tools and platforms that they have already invested in. The introduction of Big Data poses a challenge because it would require integration into existing systems and processes.
Big Data means that organizations need to develop methods to store a variety of large volumes of data from different sources. This creates an issue around security. There a many standards, policies and laws around how certain types of information should be stored and protected. Failure to adhere to these policies, laws and standards can lead to non-compliance troubles, data loss, and exposure of data to unauthorized users as well as the accumulation of data without the right quality. Storage of big data itself is also another challenge. Organizations that collect and use large sets of data are tasked with devising efficient ways to store the data accumulated.
Big Data analytics consists of resource intensive methods and practices. These practices require skilled individuals who not only know how to refine data search and processing queries, these individuals should also have mathematical and computing knowledge. This creates a complex environment where advanced skills that are not in high supply in the correct mix are required. Many organizations therefore lack the required skills to make a success out of Big Data analytics. The issue of resources or lack thereof is not limited to human resources. Complex technologies are required to efficiently process data and derive useful solutions. The lack of either the human or technical resources thus poses a challenge to the diffusion of the acceptance of big data and the ideas around it.
Explaining big data with the attributes of the diffusion of innovation model
As with many ideas, the spread of the concept of Big Data analysis is also by namely, 5 factors. These factors being Relative advantage, compatibility, complexity, trialability and observability.
The acceptance of a technology is more likely to be widespread if a technology has an advantage over existing technologies. With regards to Big Data, many companies need to see that Big Data analytics is better than the technologies currently in place. An advantage that current technologies continue to hold is that they require less resources, skills and investment than Big Data Analytics does. Big Data analytics however, wields some advantages on its own. These include the capability to analyze large datasets and make business decisions. According to New Gen Apps, business decisions derived from Big Data Analysis can lead to cost savings, new product development tailored to client needs, clearer understanding of client and market behaviors. In the end however, an organization is the only party that can decide how much marginal benefit will be reaped from implementing big data analytics. An organization will only adopt big data analytics if the adoption is more advantageous than current practices. (New Gen Apps, 2017)
Compatibility is another inhibitor or motivator for the diffusion of an innovation. A technology that is consistent with existing technologies, standards, practices and value is more likely to be adopted. Big Data analytics therefore, has not seen the widespread adoption by many organizations due to the intensiveness of the resources required. Big Data analysis is not consistent with many existing business practices and structures because it requires new technologies, methodologies and skills. It is thus not easy to integrate into existing structures and processes. The likelihood for the growth of Big Data analytics has however increased because vendors (such as vendors in Hadoop and NoSQL landscapes) are developing specific adapters to solve the issues compatibility of pre-existing systems.
If a technology or idea is to complex to understand, the likelihood of it being adopted decreases. This is something that can be observed in the diffusion of big data analytics. As previously discussed, the human resources and skills required for successful big data analysis are scarce. The methodologies and technologies involved are thus not easy to understand as they may require individuals who have knowledge in multiple areas such as mathematics, computing, business knowledge as well as querying knowledge. The complexity aspect could thus be used to explain the slow diffusion of big data analysis, particularly in countries like South Africa where the aforementioned skills are in extremely short supply.
An opportunity to try out data analytics technology would most likely increase the rate at which its being adopted. Trialability allows potential users to try and test a new innovation so that its potential benefit can be observed. The opportunity to experiment with Big Data analysis tools and technologies would establish an environment where more parties would be willing to adopt the practice as they have more certainty around its outcomes.
According to Nguyen ; Petersen (2017) observability refers to the extent to which the results of an innovation is visible to others. The benefits of Big Data analysis have been stated. Being able to see the results that big data analysis would render would increase the likelihood for adoption if more organizations believe that the results benefit them.
Stages of adoption
In accordance with the diffusion of innovation theory, there are 5 stages of an adoption of an innovation over time. There are 5 stages to the adoption of an innovation, namely knowledge, persuasion, decision, implementation and confirmation. These stages within organizations (the context for this paper) are however a little more complex and will be dealt with in this essay. These stages are agenda-setting, matching, redefining/restructuring, clarifying, and routinizing. These stages could be used to explain the current stance of Big Data adoption in organizations around the world.
Agenda setting stage serves as the first stage of the initiation phase. According to Rogers (1983) “… agenda-setting implies that one or more individuals in an organization identify an important problem and then seek an innovation as one means of coping with the problem.” With regards to big data, this means that the first step to the adoption of big data within an organization would be for a member of the organization to identify a need for its use. At this stage, business problems are defined and an innovation to respond to them is identified or developed. It is at this stage that product developers in an organization may raise questions like “what does the client want?” Questions like this allow for the seeking of innovations to respond to the need. The study of big data could lead to the answering of that question and thus big data should be looked into.
The next stage is matching which is defined as the comparing of the of the innovation to the problem space. This serves to establish whether or not the innovation will actually solve the problem identified by the organization. In the instance that the managers or other decision makers believe that the innovation will deliver desired results, then the innovation will be adopted. Should however, decision makers conclude that the innovation will not respond to the business problem in the desired way, it’ll be rejected. If the CIO or anybody put in charge of business intelligence establishes that Big data will not solve existing business problems, they’ll see no need to adopt it. It is thus worth mentioning that an organization will only adopt big data if it believes that big data solves issues that the organization identifies as business problems.
The third stage is referred to by rogers (1995) as redefining/restructuring. This stage is where an innovation selected to respond to an organizational need is finally brought into the organization. The innovation thus begins to loose its foreign character and adopts an organizational one. It is possible that it doesn’t respond to business needs in its original form and thus the organization could change/adapt it. This means that would be where big data technologies and processes are introduced in the base form and integrated into the organization. Should the shelf version of the big data processing/mining technology not respond adequately to business problems and questions, the organization would need to tailor it in order to integrate a technology that allows for the mining and analysis of data that responds to the needs of the user organization.
The fourth stage is clarifying and is defined as when an innovations use in an organization becomes is widespread or accepted, that the general organizational community understands its purpose and meaning more clearly. Misunderstandings are cleared up and the innovation is becoming ingrained into organizational culture. This would mean that everybody in an organization is aware of what purpose big data and its implementation serves in the organization. Individuals within the organization know what the end goal is of the adoption of big data is and are thus less resistant towards it. (Rogers, 1983)
The fifth and final stage is called routinizing. This stage is the stage at which an innovation is no longer regarded as a foreign construct but as a part of the organizational workflow and processing. This would be the stage in an organization at which big Data analytics are treated as an everyday and mandatory part of the everyday culture and operations. Big data analytics at this stage would no longer be treated as something foreign, but would maintain the same standing as any other organizational process and activity.
Recommendations for the successful adoption of big data
Successful integration of resource and skills intensive big data technologies requires a cross functional approach. This is because big data is not just a function of the IT department, but involves the entire organization and all its transactions, processes and information. As a consequence, the adoption of big data is a strategic move more than it is an operational move. This would thus require senior management such as the CEO, CIO and perhaps the CFO to act on the decision to adopt, with guidance from the board of directors. The entire organizational structure would need to change in order to integrate data sources and processors from all angles in the organization to one centralized location. Organizational structures need to be amended to integrate all business functions and departments so that there are no barriers to the flow of communication and information. This would establish a cross functional big data platform that strategic management can use to make business decisions. (Blissing, 2017)
In order for a tool to aid a business in decision making however, it should be effective. This means that any innovation adopted should respond to business problems while aiding the organization reach its goals. A tool in ineffective if it doesn’t satisfy stated outcomes. Organizations should invest heavily in the research and implementation of big data resources. Before a Chief information officer/ Chief executive officer approves a strategic way forward with regards to big data, the individual at hand should know exactly what the organization needs and if the proposed innovation resolves that need. Thus intense research is required. With regard to implementation however, an organization should invest in skills development to develop the skills required to operate successfully in the big data environment. An organization could invest in technology and other resources, but none of them are very beneficial if the required human resource is absent. This is because it renders the tools and technologies obsolete as nobody could use them effectively.
To function effectively however, a cross functional corporate structure needs a corporate culture that embraces change. Big data is not an old concept in itself and neither is the technology around it. New technologies are consistently being changed or developed for the purpose of data analysis. Legacy systems and processes will not be sufficient enough to allow for the effective use of big data. Strategic management needs to introduce a corporate culture that is dynamic and positively responsive to change. An environment that is resistant to change and new ideas is not one where big data can be successfully implemented.
According to Rogers (1983) “An immediate move to full implementation of an innovation may lead an organization to neglect important stages in the innovation process.” This simply means that an organization should not rush into the implementation of big data as rushing the full implementation of an innovation may lead to the bypass of mandatory steps or processes. A rush to fully implement big data without fully observing the implications around it could be problematic. The chief information officer should thus establish a clear strategic and implementation road map that would be rolled out to more operational levels. This plan should establish all areas that are mandatory so that they aren’t bypassed by lower or even senior management. The plan should be appropriately timed and budgeted.
Big data is a relatively new concept. Despite the fact that many organizations have not adopted it, it is growing in the digital age. As the technologies around it become more and more interoperable with existing infrastructure, it is becoming easier to integrate and to remove the barriers and challenges around big data. As time progresses and the innovation spreads however, more and more individuals will buy into it. With more and more vendors making it easier to integrate big data, more organizations will look to it as a way of answering business problems and deriving business solutions.