Companies that are data-driven demonstrate improved business performance. McKinsey says data and analytics can provide EBITDA (earnings before interest, taxes, depreciation, and amortization) increases of up to 25% to the business [1]. According to MIT, digitally mature firms are 26% more profitable than their peers [2]. Forrester’s research found that organizations using data and analytics are three times more likely to achieve double-digit growth [3]. Fundamentally, data and analytics have the potential to improve the company’s revenue, reduce expenses, and mitigate business risk. Despite this, many companies do not have the necessary data captured for managing their operations, compliance, and performance. McKinsey Research found that the US economy is only operating at 18% of its digital potential – an estimated $2 trillion hit to the economy. The study also found that digitization is happening unevenly, and organizations with advanced digital capabilities are winning the battle for market share and profit growth [4].
But even when companies have digitized their business processes and captured data, the data quality levels are low in most organizations. While operational and compliance processes are largely deterministic and often yield high-quality data, the inherently stochastic nature of analytics presents unique challenges for data quality. Unlike operations and compliance processes, the effectiveness of analytics relies heavily on the questions posed and the data available. For instance, a question about historical performance—such as “What were the sales in the third quarter of last year?”—can be answered relatively easily and accurately. However, forecasting sales for the fourth quarter of the upcoming year, particularly on days with snowfall, is far more complex. Basically, ensuring robust data quality in analytics, therefore, requires not only reliable data sources but also a strategic approach to question framing and data acquisition.
However, most enterprises struggle with data quality. According to a Harvard Business Review (HBR) study, only 3% of data within enterprises meet quality standards. Experian Data Quality reports that poor data quality can impact up to 12% of revenue, while Gartner estimates that 27% of data within top global companies is flawed. Research by IBM and Carnegie Mellon University reveals that 90% of enterprise data is unused for business purposes [5]. This raises a crucial question: are firms better off with bad data, or with no data at all?
But what exactly is “bad data” and “no data”? Bad data refers to inaccurate, incomplete, outdated, biased, or corrupted data that leads to poor operations, compliance violations, and errors in decision-making. NO DATA is the absence of data, meaning that business operations, compliance validations, and decisions are made without empirical evidence using assumptions, intuition, or experience. Now the key question is – are firms are better off with bad data than no data or vice-versa. The answer depends on the degree of impact or consequence of bad data and no data on operations, compliance, and analytics.
Bad data poses significant problems and risks for the organization:
Below is the impact or consequence of bad data on operations, compliance, and analytics.
- Operations: Bad data leads to inefficiencies, increased costs, communication issues, delays in managing business processes, and more. All these issues ultimately hinder operational performance and profitability. For example, incorrect stock levels can result in overstocking or stockouts, delayed shipments, and billing errors leading to longer cycle time and dissatisfied customers. In addition, storing and processing bad data consumes valuable storage and computing resources, unnecessarily increasing costs, complexity, and carbon footprint of the firm without providing much business value [6].
- Compliance: Bad data increases the likelihood of regulatory violations, data breaches, and reputational damage, and this can have severe financial and legal consequences for the organization. Firms in the Banking, Insurance, and Healthcare sectors are subject to strict regulations such as GDPR, HIPAA, and SOX. BAD DATA can lead to non-compliance with these regulations, resulting in audits, investigations, sanctions, hefty fines, legal penalties, and more. In 2014 Home Depot paid out over $135 million to credit card companies and banks because of a data breach. In 2021, regulators in Luxembourg fined $877 million to Amazon for GDPR breaches [6].
- Analytics: If decisions are based on bad data, stakeholders may lose trust in the system or process. Bad data undermines the accuracy and reliability of analytical insights, reducing the effectiveness of data-driven strategies and initiatives. Key performance indicators (KPIs) derived from bad data can misrepresent the true performance of the business, leading to misguided insights and decisions. Predictive analytics performed on bad data can skew forecasts, leading to incorrect business strategies and missed opportunities.
No data too poses significant challenges for the organization:
Below is the impact or consequence of no data on operations, compliance, and analytics.
- Operations: Without data, companies cannot identify bottlenecks, redundancies, or areas for improvement. This leads to inefficiencies, increased costs, suboptimal resource allocation, and missed opportunities in business operations. Without data, businesses have no way of identifying which departments, products, or projects are most profitable or require additional resources. Automation and workflow optimization require data inputs to function properly. Without data, processes are prone to human error and will be slow, preventing businesses from scaling efficiently. Also, without data, tracking and verifying financial or operational transactions becomes extremely challenging, increasing the likelihood of fraud, mismanagement, or unauthorized activities.
- Compliance: Many industries are required to provide regular, detailed reports to regulatory authorities related to financial, environmental, health, or safety data. Without data, businesses cannot meet these requirements, leading to fines, penalties, or even operational shutdowns. For example, when Nexen, an oil company based in Canada spilled over 31,500 barrels of crude oil in July 2015, the Alberta Energy Regulator (AER) ordered the immediate suspension of 15 pipeline licenses issued to Nexen due to a lack of maintenance data records [5]. Also, the absence of proper records or data can result in stricter regulatory scrutiny, reputational damage, and potential legal consequences.
- Analytics: Without data, organizations cannot perform any analytics, making it impossible to derive insights to measure and improve business performance. Today customers expect personalized products, services, and experiences. Without data, businesses cannot create personalized offerings, leading to generic marketing strategies and lower customer engagement. Overall, without data, businesses fall behind competitors who are leveraging analytics to innovate and grow,
So, what is the conclusion? While both no data and bad data are undesirable for the business, in most cases bad data might be better than no data provided the level, or the state of data quality is known. Bad data can still offer a better starting point and provide a general sense of direction than no data at all. This is especially true if data is clearly understood and supplemented by realistic assumptions and expert insights. Data profiling using EDA (Exploratory Data Analytics) techniques can help in assessing the level of data quality. But in high-stakes situations like medical diagnoses and financial investments, no data is often better than bad data, as faulty data could be misleading resulting in disastrous outcomes for the business. Overall, the key is to understand or assess the level of data quality on a regular basis and devise strategies for data cleansing, data governance, data literacy, and more to validate and improve data quality in the entire data lifecycle (DLC).
Read more interesting articles:
Future of APIs and Data Integration: Essential Trends to Watch
Unmasking Algorithmic Bias: Hidden Injustices within AI
References
- Bokman, Alec; Fiedler,Lars; Perrey,Jesko; Pickersgill, Andrew, “Five facts: How customer analytics boosts corporate performance”, https://mck.co/2Ju0xYo, Jul 2014.
- MIT, “Digitally Mature Firms are 26% More Profitable Than Their Peers”, https://bit.ly/2xBTPNe, Aug 2013
- Evelson, Boris, “Insights Investments Produce Tangible Benefits — Yes, They Do”, https://www.forrester.com/blogs/data-analytics-and-insights-investments-
- https://www.mckinsey.com/~/media/mckinsey/industries/technology%20media%20and%20telecommunications/high%20tech/our%20insights/digital%20america%20a%20tale%20of%20the%20haves%20and%20have%20mores/mgi%20digital%20america_executive%20summary_december%202015.pdf
- Southekal, Prashanth, “Data Quality: Empowering Businesses with Analytics and AI”, John Wiley, Feb 2023
- https://www.forbes.com/councils/forbestechcouncil/2021/04/06/can-data-be-a-liability-for-the-business/
I’m a consultant, author, and educator with over 80 client collaborations, including P&G, GE, Shell, Apple, and SAP. I’ve authored three books—Data for Business Performance, Analytics Best Practices, and Data Quality. My second book, Analytics Best Practices, was recognized by BookAuthority in May 2022 as the #1 analytics book of all time. Alongside consulting and advisory work, I’ve also had the privilege of training over 4,500 professionals worldwide in data and analytics.
