CHAPTER 4

Analytics and High-Velocity Decision Making

Managers must match the velocity of streaming data with high-velocity, data-based decision making and agile execution to successfully compete.¹ Digital transformation is ongoing and evolving. Managers engaged in ­finding ways to make greater use of digital data must understand the importance of analyzing and using digital data in decision making to ­actually improve the overall functioning and success of an organization. Digital transformation maturity refers to the progress of managers in ­implementing various actions to re-imagine or re-invent the ­business, typically leveraging technology. Managers implementing a digital ­transformation ­strategy should focus on adopting and using innovative technology that can enhance organizational decision-making capabilities.

In a collaborative research study conducted by MIT Sloan Management Review and Deloitte professional services network, Kane et al. (2015) found that among digitally maturing organizations, nearly 90 percent of strategies focus on improving decisions and innovation. The following six sections emphasize improving decision making and include: (1) the basics of analytics, (2) current trends in advanced analytics and applications, (3) a guide to identify and select data analytics tools, (4) analytics as an enabler for data-based decision making, (5) high-­velocity decision making, and (6) the ethical challenges encountered with data analytics and data-based decision making.

Basics of Analytics

Analytics is often described as the science of analysis and discovery. The term refers to quantitative analysis of data. People who conduct analyses and develop analytic applications are decision or data scientists. Analytics refers to a broad set of tools and capabilities that provide decision ­support. Analytic capabilities are important in data-driven and model-driven DSS and analysis with quantitative and statistical tools is the focus of special studies such as knowledge discovery or data mining.

Davenport and Harris (2007) define analytics as “extensive use of data, statistical and quantitative analysis, exploratory and predictive models, and fact-based management to drive decisions and actions. The analytics may be input for human decisions or drive fully automated ­decisions” (p. 7).

Analytics is a broad umbrella term that includes business analytics and data analytics. Business analytics (BA) use data and analytics to improve business operations and decision making. BA includes optimization techniques and Key Performance Indicators (KPIs). Data analytics applies quantitative and statistical methods to analyze large, complex organizational data sets. Managers need to understand the value of using data analytics and the decision support capabilities made possible by leveraging data.

Analytic applications have three main technology features: (1) data management and retrieval, (2) mathematical and statistical analysis and models, and (3) techniques for data visualization and display. Analytic applications are used to process large amounts of structured and unstructured data to find patterns and provide information. Analyzing data can be challenging and more data can increase the complexity of an analysis. More data does not mean better analytics. Like all computerized systems, for analytics to be useful the data must be accurate, complete, and representative of the real world.

Some sources consider analytics as a subset of business intelligence (BI), while some use the terms analytics and BI interchangeably, other commentators are more specific and consider only reporting analytics as another name for BI. In this discussion, data-driven DSS and BI are considered as reporting analytic applications. There are three major types of analytics: (1) reporting analytics, (2) prescriptive analytics, and (3) ­predictive analytics.

Information Systems vendors and analysts tend to use BI as a category of software tools that can be used to extract and analyze data from corporate databases. The most common BI software is query and reporting tools. This software extracts data from a database and creates formatted reports.

Prescriptive analytics manipulate large data sets to make recommendations. This type of decision support prescribes or recommends an action, rather than a forecast or a summary report. Prescriptive analysis relies on sophisticated analytics including graph analysis, simulations and machine learning. Through evaluating decision options, managers can use prescriptive analytics to take advantage of an opportunity in the future. Using What… questions such as “What should be done to achieve xxx in the future?” Predictive analytics is based on quantitative and statistical models and this category of analytics includes model-driven DSS.

Predictive analytics is a general term for using simple and complex models to support anticipatory decision making. Analysis of historical data is used to build a predictive model to support a specific decision task. The decision task may be determining who to target in a marketing campaign, what products to stock, possibility of fraud, or who the “best” customers are for a firm. Using historical data, predictor variables are identified for building quantitative or business rule models. The model makes a prediction for a decision task.

Managers in consumer-packaged goods, banking, gambling, energy, and health care industries are the most active users of predictive analytics. Predictive analytics is increasingly incorporated in day-to-day operations management tasks. New projects can be implemented faster because software has improved for analysis and development, but the number of IT professionals skilled in using the many varied analytical techniques is inadequate to meet the demand.

Developing analytics should involve both business and IT managers. This joint development process should help in understanding and in some cases automating business operations decisions. Creating a meaningful analytics development partnership can facilitate improved and enhanced routine decision making. Working together on analytics development can serve as a bridge between IT and business managers.

Development and use of analytics should be a core technology competency of many companies and managers should be reluctant to outsource or offshore the capability. Managers must realize the cost of each analytics project is an investment in building competency and it can also reduce operations costs and enhance operations. Certainly there is a learning curve associated with analytics but consultants can reduce the curve. Managers should not however assume that programmers outside their firm can easily understand the peculiarities and needs of their business. As organizations capture more and more data, it will be important to analyze and use the data to enhance business results. Implementing analytics is NOT just another routine Information Technology (IT) project.

Based on a number of sources including the IBM Analytics Quotient (AQ) quiz, the following questions and responses identify best practices for using IT to implement analytics.

Answering the questions presented in Table 4.1 can provide managers with a baseline guide to assess current development of analytics capabilities.

Table 4.1 Decision guide for implementing analytics

Ask questions	Guideline
1. What types of data sources should managers analyze?	In general, standard enterprise data sources across functions should be combined with data from external sources, point of sale, RFID, and social media
2. How important is the quality of data used in analyses?	Very important. An organization should have an enterprise data model. Common master data and metadata must exist and strong data governance practices must be in place
3. Should managers document outcomes of analytics initiatives?	Yes. Managers should initiate a documentation ­process to capture how the use of business analytics has changed business operations. Successful projects will lead to more projects
4. How important is using predictive models?	Very important. Integrated planning and predictive modeling can enable an organization to adjust policy and execution in response to shifting dynamics in the organization and business environment
5. How should managers assess and manage risk?	Risk metrics should be industry specific. Managers should share risk management assessment and mitigation processes across the organization, identify the most significant cross-departmental risks in an effort to reduce loss, and link risk reduction and specific risks to business objectives and improved performance
6. Should managers centralize resources for performing and developing analytics?	Yes, but analytics knowledge should be widespread throughout an organization. Using analytics should become part of the organizational culture. Managers should establish an analytics center of excellence and cross functional analytics team
7. What general analytics solutions should be implemented in organizations?	Solutions are in four categories: (1) Reporting and analytics, (2) Planning, budgeting and ­forecasting, (3) Predictive and advanced analytics, and (4) ­Governance, risk and compliance analytics
8. How should managers anticipate future events and results?	It is important to use both qualitative and quantitative methods, including: (1) experience and intuition, (2) predictive analytics for priority needs, (3) “what if” scenarios, and (4) integrated planning and ­predictive models

A joint MIT Sloan Management Review and IBM report identified three core competencies organizations must master to achieve a competitive advantage with analytics. The first is information management, which rely on standardized data practices. The second is analytics skills, which revolve on core discipline expertise, built on robust tools. Finally, there must exist a data-centric culture that sees analytics as a key asset to support evidence-based management.

Analytics is continuing and perhaps accelerating. Analytics are important decision support tools that lead to data-based decision making.

Current Trends in Advanced Analytics and Its Applications

Analytics and business intelligence (BI) technologies continue to be key enablers of most organizational data and decision support strategies. The opportunities for mobile BI are many, and the trends toward self-service analytics and visualization are both exciting and promising for data-­centric organizations.

BI and related technology initiatives attract a lot of attention from technology experts, managers, consultants, and vendors. A recent survey of BI professionals identified data discovery/visualization, self-service BI, and data quality/master data management as the most important trends. In her Datapine blog, Mona Lebied (2016) identified key areas for BI and analytics ranging from security and digitization to cloud analytics, embedded BI, and data storytelling.

AI and ML are areas attracting considerable attention. Managers in organizations large and small are trying to understand the prospect of AI in general and the possibilities for their business. Gartner has flagged advanced AI and ML as important trends for organizations by 2018. AI is a broad concept that has been around for many years. AI refers to the simulation of “smart” behaviors in computers. ML is a subset of AI that uses algorithms to learn and improve from experience. The opportunities for these technologies in business and health care are extensive. In a recent Forbes article, Marr (2017) highlights disease identification and diagnosis, crowd sourcing treatment options and drug response monitoring and disease surveillance, as areas where AI is having significant impacts in the health domain. Smart cities, smart manufacturing and smart cars are also capturing public attention.

Core to generating new value from decision support and analytics, managers should experiment with and explore new opportunities in AI/ML, visual data discovery, and data storytelling. Moving beyond traditional graphs and charts, this era of infographics pushes boundaries in terms of trying new ways to use data to tell a business story through the effective use of sophisticated approaches to data visualization.

Collaborative decision support is fostered through the increased availability and use of mobile technology and Web 2.0 technologies. BI is no longer solely for senior. Self-service BI provides opportunities to share data. Also, improvements in software facilitate new ways for embedding decision support features in existing software applications.

With increasing investment of resources in data and technology and heightened expectations for digital transformation, we anticipate many organizations will establish dedicated centers for analytics and self-service BI (cf. Lebeid 2016).

Data governance and security are high priorities for most organizations. Security is a ­consideration for managers in all businesses looking at cloud analytics and cloud storage options. While self-service BI in the cloud affords many benefits to individual users, realizing this level of flexibility and agility is challenging in terms of security and data governance (Potter 2015).

Analyzing data is often challenging and more data can increase the complexity of an analysis. More data does not mean better analysis. ­Hiring data scientists, buying more hardware or software or hiring consultants does not guarantee success. It may be necessary to implement capabilities like in-memory processing or software like Hadoop. Additional training is probably needed and possibly new staff might also be needed.

McKinsey director David Court (2012) argued multiple success ­factors must be present to use new technologies: creative use of ­internal and external data, developing workable models, and transforming the company to take advantage of data. Finally, Court notes “you’ve got to make a decision support tool the frontline user understands and has confidence in.”

To exploit these trends six critical success factors must be present:
(1) managers must want to use analytics and decision support, (2) knowledgeable and innovative data and decision support analysts, (3) high quality data, (4) accurate models for forecasting and prediction, (5) appropriate technology, and (6) effective data-centric culture of the process.

Select Analytics Tools and Technologies

Selecting the most appropriate analytics approach and tools for a specific task or project is important. The approach that is the best fit depends upon many factors including: (1) the need and objective, (2) data availability, (3) training and background of current data analysts, (4) vendor support, and (5) the industry/type of organization. Choosing the wrong approach and tools often results in a difficult and incorrect analysis. ­Analytics is about asking specific questions and finding the best answers. As discussed in a previous chapter, a question asking technique should also be used for choosing analytics approaches and tools.

Descriptive analytics primarily uses data aggregation and statistical tools like averages and differences. Predictive analytics use more complex statistical models like regression and correlation and forecasting techniques like moving averages. Diagnostic analytics uses tools like drill down, interactive data visualization and data mining. Finally, prescriptive analytics uses tools like optimization, simulation, scenario analysis, and case-based reasoning.

In a Forbes article, Bernard Marr (2016) sets out six approaches to analytics. In considering these approaches it is important to understand when specific tools should be used. We have briefly assessed these approaches in terms of the nature of the data needed and the business goal(s) that can be met. The following is our assessment.

1. Business Experiments

Testing ideas is often a goal of data analytics. A business ­experiment approach is used to test the validity of an idea. This may be a ­strategic hypothesis, a new product package or a marketing campaign. ­Davenport (2009) advocates a “test and learn” approach to conducting business experiments noting that experiment design is key to ­generating a sound evidence base. We recommend using a business experiment approach when seeking to test ideas systematically.

2. Causal Analysis

Finding causes helps understand a situation so changes or ­prediction is possible. Regression is a primary causal analysis tool that is ­useful when understanding and/or prediction is required and adequate data are available on plausible predictor variables. Regression is a ­statistical tool for investigating the relationship between variables. For ­example, managers might use regression analysis to understand the causal ­relationship between price and product demand. Use causal ­analysis when a complex situation is data rich and managers want better understanding.

3. Correlation Analysis

This is a statistical technique that allows managers to determine whether there is a relationship between two separate variables. It also helps to determine the strength of the relationship between the variables. We might use correlation analysis to understand if there is a relationship between ­positive customer experience and customers sensitivity to changes in the price of a product or service. Use correlation analysis to explore relationships.

4. Forecasting Analysis

This approach uses a time series of data values to forecast or predict other ­values. For example, managers may use sales data from the past to predict future sales values. Perform a forecasting analysis when the primary goal is estimation of a variable(s) of interest at some specified future date.

5. Scenario Analysis

Managers can consider “what-if” questions by analyzing a variety of ­possible future events or scenarios considering possible alternate outcomes (Power and Heavin 2017). Use scenario analysis where there are numerous possible course of action and a high degree of uncertainty about the potential outcome.

6. Visual Analytics

Data can be analyzed in many ways and the simplest way is to create a visual or graph as a means of identifying patterns or trends. This is an interdisciplinary approach integrating data analysis with data visualization and human interaction. For example, a sales manager could use an interactive map to better understand customer purchasing behaviors by region. Use data visualization when managers are interested in directly deriving insights from large volumes of data.

As mentioned previously, business analytics and data analytics are terms used interchangeably to describe a systematic process of purposefully examining and using data sets with statistical and quantitative models, and leveraging the capabilities of sophisticated algorithms and technologies. Often the goal is to draw conclusions about the underlying meaning or implications of the data. Business analytics emphasizes business uses of analytics, while data analytics has a broader focus across organizations and settings. There is no single best approach for meeting every analysis goal. Once the most appropriate analytics approach for a specific situation is selected, a manager or analyst must select analytics tools and technologies to conduct the analysis. There are many analytics and decision support software tools available and many of them are open source and widely available. Managers should consider and evaluate the technologies currently available and the technical capabilities of staff available in-house before going to the market for new software tools or to hire experts.

Managers and analysts continue to investigate new opportunities in visual data discovery and data storytelling (Heavin and Power 2017). Moving beyond traditional graphs and charts, sophisticated data visualization technologies promote new ways of telling a business story from data insights to data visualization. Analytics approaches are evolving.

Analytics Should Inform Data-Based Decision Making

Analytics is an integral component of a successful organizational digital transformation strategy. Use of analytics can offer individuals, organizations, governments, and our global society “data-based” perspectives on existing challenges and possible solutions. Analytics can provide facts to improve data-based decision making and help find meaning in an increasingly complex environment.

In her Teradata blog, Yasmeen Ahmad (2016) identified advantages of analytics, including: (1) increased proactivity and the ability to anticipate needs, (2) delivering the right products and services at the right time, (3) improved personalization and service, and (4) optimizing and improving the customer experience. Analytics provide these benefits when incorporated in data-based decision processes. Analytics has been lauded by some as the “silver bullet” solution to finding an organizational competitive advantage. Sadly, it is NOT a “silver bullet,” but rather analytics is more like the raw unformed silver! Gaining Ahmad’s advantages does seem reasonable, but the task is an ongoing challenge.

Data albeit “good data” on its own will not result in good decision making (cf. Shah et al. 2012). In their study of data-savvy practitioners, Shah and colleagues identify five challenges to data-based decision ­making in organizations, these include:

1. Few employees have analytical skills;
2. IT departments need to invest more resources in providing information and less in the technology aspect of IT;
3. There is broad acknowledgment good quality data exist, however it is often a challenge to locate important data sources;
4. Managing data and providing information is widely perceived as the sole responsibility of the IT function. Traditionally, business ­managers did not engage in data and information management. The typical manager often neglects understanding information that is received.
5. There is an urgent need to develop more informed sceptics. “Employees need to recognize that not all numbers are created equal—some are more reliable than others.”

Notably, the Shah et al. study highlights the widespread misperception that analytics is the responsibility of an organization’s IT function. Analytics is the everyone’s responsibility, it should “put information in the hands of business analysts and business users and offer significant potential to create business value and competitive advantage” (Jones 2016).

Benefits of analytics are always constrained or limited by the ­manager or managers who use the results to make decisions. In order to achieve meaningful integration between analytics tools and technologies, ­analytics must support and reinforce data-based decision making and ­management. Key performance data should be included in a ­data-based management process that provides routine evaluations to improve ­organizational ­outcomes, including quality, and financial metrics. This integrated more holistic and balanced view of providing analytics is ­illustrated in Figure 4.1.

Figure 4.1 Finding a balance between “Data analytics” and “Data-based management”

Negotiating the balance between an organization’s approach to data analytics and a data-based management strategy can be difficult. To achieve this dual focus, it is essential that managers participate in the process of analyzing data in a way that leverages new data insights and integrates them into the organization’s management and decision-­making processes. Shah et al. (2012) suggest that managers need to be better trained to use new analytics tools, paying particular attention to building analytics into managerial processes. If a balance is not achieved, managers run the risk of investing in expensive analytics technologies that are not used by managers. It is important to implement the “right” amount of analytics to support a data-centric culture and that should lead to better data-based decision making.

High-Velocity Decision Making

Consequential decisions can and often should be made using a high-­velocity, high-quality decision-making process. High-velocity decision making is a new label for a familiar idea: make reversible decisions using streamlined, rapid data-based decision processes focused on issues, yet make sure the processes are thoughtful and goal-oriented. Jeff Bezos, Founder and CEO of Amazon.com, Inc., in his 2016 and 2017 shareholder letters discusses decisions and decision making. His ideas are very relevant to managers interested in using analytics and decision support to improve decision making outcomes. In this era of digital transformation, Bezos advocates for using high-velocity, high-quality decision making for consequential and reversible decisions. The following paragraphs explain the rules of high-velocity decision making:

Rule 1: Know what kind of decision you are trying to make. Is it a Type 1 consequential and irreversible decision? or a Type 2 changeable and reversible decision?

Bezos notes Type 1 decisions are

consequential and irreversible or nearly irreversible—one-way doors—and these decisions must be made methodically, carefully, slowly, with great deliberation and consultation. If you walk through and don’t like what you see on the other side, you can’t get back to where you were before.

Type 2 decisions are different.

Most decisions are changeable, reversible—they’re two-way doors. If you’ve made a suboptimal Type 2 decision, you don’t have to live with the consequences for that long. You can reopen the door and go back through. Type 2 decisions can and should be made quickly by high judgment individuals or small groups.²

Bezos argues

As organizations get larger, there seems to be a tendency to use the heavy-weight Type 1 decision making process on most decisions, including many Type 2 decisions. The end result of this is slowness, unthoughtful risk aversion, failure to experiment sufficiently, and consequently diminished invention.³

Rule 2: Strive to be a Day 1 company.

Bezos explains

I’ve been reminding people that it’s Day 1 for a couple of decades. I work in an Amazon building named Day 1, and when I moved buildings, I took the name with me. I spend time thinking about this topic. ... Day 2 is stasis. Followed by irrelevance. Followed by excruciating, painful decline. Followed by death. And that is why it is always Day 1. .....

Day 2 companies make high-quality decisions, but they make high-quality decisions slowly. To keep the energy and dynamism of Day 1, you have to somehow make high-quality, high-velocity decisions. Easy for start-ups and very challenging for large organizations. The senior team at Amazon is determined to keep our decision making velocity high. Speed matters in business—plus a high-velocity decision making environment is more fun too.

Rule 3: Strive to make high-quality, high-velocity decisions.

Never use a one-size-fits-all decision making process. Many decisions are reversible, two-way doors. Those decisions can use a light-weight process. For those, so what if you’re wrong?

Rule 4: Consider the trade-off between seeking more information and a slower decision.

Most decisions should probably be made with somewhere around 70 percent of the information you wish you had. If you wait for 90 percent, in most cases, you’re probably being slow. Plus, either way, you need to be good at quickly recognizing and correcting bad decisions. If you’re good at course correcting, being wrong may be less costly than you think, whereas being slow is going to be expensive for sure.

Rule 5: Disagree and commit when appropriate for Type 2 decisions. Trust other managers and know when to respectfully disagree and go along with the group.

The phrase “disagree and commit” will save a lot of time.

If you have conviction on a particular direction even though there’s no consensus, it’s helpful to say, “Look, I know we disagree on this but will you gamble with me on it? Disagree and commit?” By the time you’re at this point, no one can know the answer for sure, and you’ll probably get a quick yes.

Rule 6: Go for quick escalation. Know when senior management should make the decision.

Recognize true misalignment issues early and escalate them immediately. Sometimes teams have different objectives and fundamentally different views. They are not aligned. No amount of discussion, no number of meetings will resolve that deep misalignment. Without escalation, the default dispute resolution ­mechanism for this scenario is exhaustion. Whoever has more stamina carries the decision.

Know when a decision needs to be escalated to the senior team. As Bezos concludes “‘You’ve worn me down’ is an awful decision making process. It’s slow and de-energizing. Go for quick escalation instead—it’s better.”

Rule 7: Strive to have the spirit and the heart of a small company.

Bezos asks

Have you settled only for decision quality, or are you mindful of decision velocity too? Are the world’s trends tailwinds for you? Are you falling prey to proxies, or do they serve you? And most important of all, are you delighting customers? We can have the scope and capabilities of a large company and the spirit and heart of a small one. But we have to choose it.

Case studies and news stories at DSSResources.com suggest that high-velocity environments are increasingly common and that managers must learn to make high-velocity decisions if they are to remain relevant and part of the decision process. Managers must develop fast, incremental processes for Type 2 decisions in high-velocity decision environments. In some situations, those decisions should be made by software algorithms. In rapidly changing environments, people will be eliminated from some decision processes because they are too slow. Highly-structured Type 2 decisions will be automated. Also, analytics can be used to analyze and understand business data that can be used to make high-velocity decisions.

Rule 8: Strategic, consequential, irreversible Type 1 decisions should be made using a high-quality decision-making process that is methodical, careful, and thoughtful, and made with deliberation and consultation.

High-velocity decision making is not the same as high-speed decision making. High-speed decision making describes only how to make fast decisions, while high-velocity decision making means decisions are made quickly and, in a goal-oriented direction. High-velocity decision making means a decision maker is aware of and considers the goals that are being pursued. If the decision is a reversible, nonprogrammed decision about new, novel situations requiring innovation can be made using a high-­velocity, high-quality decision-making process.

Bezos’ views are similar to but do differ from results of Bourgeois and Eisenhardt’s (1988) decision making research. In their classic study, ­Bourgeois and Eisenhardt investigated how executives make strategic decisions in industries where the rate of technological and competitive change is so extreme that market information is often unavailable or obsolete, where strategic windows are opening and shutting quickly, and where the cost of error is involuntary exit. They noted

Our results consist of a set of paradoxes which the successful firms resolve and the unsuccessful firms do not. We found an imperative to make major decisions carefully, but to decide quickly; to have a powerful, decisive CEO and a simultaneously powerful top management team; to seek risk and innovation, but to execute a safe, incremental implementation. Despite the apparent paradox, effective firms do all of these simultaneously.

In a related study published in 1989, Eisenhardt reported that her “results link fast decisions to several factors, including the use of real-time information, multiple alternatives, counselors, consensus with qualification, and decision integration” (p. 573). She also noted “the emergent perspective highlights emotion as integral to high stakes decision making … emotion is critical for understanding strategic decision making” (p. 573).

Strategic decisions are important, usually with long-term consequences, and with large resource commitments. By definition strategic decisions are consequential and some are not reversible. Reversibility means senior managers are able to change, roll-back, and reverse a decision and that the actions to implement the decision can be undone. Rather than two states, reversible and irreversible, there seems to be a vague continuum of decision reversibility ranging from completely irreversible to completely reversible. For example, Bezos notes some decisions are “nearly irreversible.”

High-velocity decision making at Amazon under Jeff Bezos is apparently effective. Making successful strategic decisions in an environment of rapid change does create paradoxes we only partially understand. One aspect of the decision process paradox is captured in the question—How do we know when and if a decision is reversible?

Ethical Challenges for Decision Making

Modern managers are becoming more strategic about the capture, storage, and value extraction from data. This strategic approach typically means leveraging large data sets to extract insightful information that was previously unknown. These insights are used to maintain or create competitive advantage for the organization. Analytics may provide results to identify new markets, new products and services, opportunities to grow revenue, and opportunities to drive down costs with the end goal of boosting organizational performance. These new uses of data are creating ethical challenges.

For example, the general data protection regulation (GDPR) in Europe and significant differences in data protection legislation across jurisdictions has stimulated much discussion about data privacy, security and protection by government agencies, regulatory bodies and managers in both public and private organizations. This ongoing discussion is widely underpinned by ethical scenarios in business related to IT. Ethics refers to “moral rules, codes, or principles which provide guidelines for right and truthful behavior in specific situations” (Lewis 1985, p. 382). The areas of analytics, BI, decision support, and big data are relatively new, we continue to uncover new and increasingly complex ethical questions on a regular basis.

Some people think that building and using a computerized decision support and analytics capability is ethically neutral. That view is poorly informed and incorrect. People are faced with ethical choices when dealing with computerized decision support that we are only beginning to recognize, consider and evaluate. Using a stakeholder perspective illustrated in Figure 4.2, Asadi et al. (2016) consider three main stakeholders when evaluating the ethical implications of analytics. These stakeholders include (1) individuals who use social media, (2) organizations who capture, store and manage data, and (3) society. The societal impacts are significant as governments and industry associations struggle to regulate and create ­policies for emergent and rapidly changing markets (Asadi et al. 2016).

Figure 4.2 Analytics Stakeholders

Given the multi-stakeholder perspective presented in Figure 4.2, the ethical considerations associated with analytics and decision support are complex. While technology is neutral, the decisions that people make about how technology is used to capture, store, analyze, and share data are not ethically neutral. IBM Engineer Mandy Chessell (2014) ­proposes nine categories of questions that should be considered by individuals and organizations when tackling ethical issues, related to data and technology including:

1. Context: Why was the data originally collected? How is the data now being used? How far removed from the original context is its new use? Is this a fair and appropriate use of this data?
2. Consent and Choice: What are the choices given to all stakeholders involved? Do they know they are making a choice? Do they really understand what they are agreeing to? Do they really have an opportunity to decline? What alternatives are offered?
3. Reasonable: Is the data used and the relationships derived appropriate and reasonable given the purpose it was collected for?
4. Substantiated: Are the sources of data used suitable, respected, ­complete, and timely for the application?
5. Ownership: Who owns the new insights generated as a result of data analysis? What are the owners’ responsibilities?
6. Fair: How fair are the results of the application to all stakeholders affected? Is everyone properly compensated?
7. Considered consequences: What are the potential consequences of the data collection and analysis?
8. Access: What access to data is given to the data subject?
9. Accountable: How are mistakes and unintended consequences detected and repaired? Can the interested parties check the results that affect them?

Exploring plausible critical ethical incidents that may be faced by managers, data scientists, and other users can help understand the complexity of ethical decision making. So what situations might occur? Using Chessell’s questions, please contemplate the scenarios in Table 4.2:

Table 4.2 Scenarios for data-based decisions with ethical implications

Scenario 1: A builder of a BI solution chooses not to include a key metric because the data are hard to capture and display. Eventually that missing metric, for example the weight of a prototype airplane, becomes a critical flaw that leads to major cost overruns.

Scenario 2: A sponsor proposes combining individual sales affinity card and credit score data and a software developer becomes concerned that the privacy rights of customers will be in jeopardy. The sponsor is a powerful figure in the company who does not like dissent.

Scenario 3: A software developer realizes the quality of data for a proposed data analytics solution is flawed and inaccurate and still proceeds to build the system. The system is never really used because of complaints of poor data quality.

Scenario 4: A software development team fails to validate a forecast model in an automated inventory replenishment system and managers report large inventory problems. The company takes a major write-down on obsolete inventory.

Scenario 5: A manager/user of a data-driven DSS notices a sales problem in a store and drills down into the underlying data and sees a large transaction by her husband. The manager confronts her spouse with the information he found using the system.

Scenario 6: A manager fails to use an investment management and control system in a timely manner and a subordinate makes a large, unauthorized trade. The trade is ill-advised and significant losses result.

Scenario 7: The knowledge base for a knowledge-driven DSS derived with AI and ML seems out of date and no one acts to fix the problem. The recommendations of the system become increasingly error prone and erratic. Managers start ignoring the results.

As a thought exercise ask yourself: What would you do in each situation? Why would you take that action? Who is the responsible party? Is the situation avoidable?

In many of these situations we encounter an ethical dilemma. Initially the situation seems clear cut, but other scenarios are gray and they require additional information, investigation, and even some expert advice.

Principles and values play an important role in making many significant organizational decisions. When analytics and decision support solutions are constructed, software developers make assumptions that can have ethical impacts on user choices. Also, some decisions are considered so value-laden that many people would be uncomfortable with developing decision support to assist a decision maker. One cannot specify all of the ethical issues that might be relevant to a specific decision support proposal, but once a proposal reaches the feasibility stage, the project sponsor needs to specifically address the ethical issues associated with the project. Also, during development developers need to be sensitive to how representations like charts and tables that are designed to present information impact a decision maker.

Privacy concerns are also easy to ignore during the evaluation of a analytics proposal. In many societies, people expect that certain personal and behavioral information about them will be kept private. This information belongs to the person and does not belong to a company, the public, or the government. Managers must ensure that data used in the organization does not infringe on the privacy rights if individuals. The exact extent of privacy rights for employees, customers, and other data providers is not always clearly defined. In general, unless there is a clearly compelling reason to risk violating an individual’s privacy, the “fence” to protect privacy of data should be higher and larger than any minimum requirements.

The following potential analytics and decision support ethical issues require more thought: (1) data quality assurance, (2) hidden data capture, (3) propagating data errors, (4) ongoing use of obsolete decision support, (5) data mashups, data linking, and data fracking, (6) combining internal and external data sources, (7) inappropriate use of customer profiles/data, (8) legal liability issues from failing to use or from misuse of a decision support capability, (9) data/key metrics exclusion, (10) analytics/decision support model validation, (11) unauthorized data transfers, (12) lack of policies or poor policy enforcement, and (13) invasion of personal privacy. Organization policies and National and Local laws should guide the behavior of managers and developers on these and related topics.

We want to encourage and promote open discussion and proactive behavior to insure ethical use and construction of analytics and computerized decision support. To do so we need to explore the subtleties of a wide variety of ethical situations that managers, developers, and system users might encounter. When in doubt about the ethical use of a decision support or analytics tool or the need to use decision support or the consequences of poor design decisions on the behavior of decision makers, do not ignore the question, rather ask others, consult, and discuss. Ignoring ethical issues associated with building and using computerized decision support is not an option.

Summary

The path to digital transformation is not an easy or even a straight one. More wide-spread adoption of analytics is a key element of digital transformation for an organization. This chapter considered analytics and decision support paying particular attention to the increasing need for high-velocity decision making. Many tools including BI, analytics, and decision support and other tools and technologies can be used to support the development of an ethical, data-based approach to organizational decision making.

Amazon’s Jeff Bezos has pursued the development of a data-driven, high-velocity decision making approach using analytics and decision support to improve decision-making outcomes. He warns of the key challenge for decision makers one-size-fits-all’ decision making.” Bezos’ approach advocates rapid decision making with less than complete information, where the decision makers use their judgment to react quickly in situations where a decision is reversible.

While this approach has seemingly worked well for Bezos, it is ­important to consider the role of ethics in making decisions in ­high-­velocity environments. Perhaps there is an opportunity to build questions, such as those proposed by Chessell (2014) into the decision process. Ethical ­decision making is important for data-driven, data-based, and data-informed decision making. Further consideration must be given to the challenges of ethical decision making in this data-intensive era of digital upheaval and transformation.

Type 1, consequential and irreversible decisions, requires ethical, data-informed decision making, including use of analytics and ­decision support. Ethical considerations are also very important in making ­high-­velocity, reversible Type 2 decisions. Considering ethical issues is relevant when managers use data-based decision making or implement an algorithmic data-driven decision automation solution. Even though some decisions can be “reversed,” some harms cannot be undone and consequences associated with an unethical decision cannot be reversed.

Chapter 5 provides additional actionable advice to managers who are implementing a digital transformation vision and strategy.

¹ Dykes, B. 2017. “Big Data: Forget Volume and Variety, Focus On Velocity.” Forbes, June 28. https://forbes.com/sites/brentdykes/2017/06/28/big-data-forget-volume-and-variety-focus-on-velocity/#5de2336f7d67

² Note: We assume a “high judgment” individual is a person with experience and good judgment.

³ Bezos“The opposite situation is less interesting and there is undoubtedly some survivorship bias. Any companies that habitually use the light-weight Type 2 decision-making process to make Type 1 decisions go extinct before they get large.”