All others bring data…

We are a data driven company!

Ever heard this before?

In the digital age, data is king. A data-driven strategy places data at the core of business decision-making, offering unparalleled insights into customer behavior, market trends, and operational efficiencies. 

This approach is meant to not only enhance decision-making but also to fortify competitive advantage through personalized customer interactions and optimized operations. 

Businesses are amassing vast amounts of information, to then leverage data for predictive analysis, better customer engagement and strategic agility, ensuring they stay ahead in today's data-centric world.

In God we trust. All others… 

A quote by Dr. W. Edwards Deming goes along these lines:

In God we trust, all others bring data.

Although the one above is most probably the most famous quote / joke about the importance of data, I personally prefer the one that reads: "Without data, you're just another person with an opinion".

Those who bring data must feel accountable for the findings and insights that should be extracted from them. 

If you are good at storytelling, your data can tell any story.

We often hear that "if you are good at storytelling, your data can tell any story". Which is often true. Data analysis, presentation and visualization must align to your business objectives. Data do not really tell any story unless you present them within the appropriate context. 

But what are the real, concrete advantages of a data-driven approach to decision making?

Advantages of a data driven approach.

Taking data-driven decisions offers several significant advantages for companies. In essence, data-driven decision-making harnesses the power of data to guide strategic decisions, optimize operations, and enhance customer experiences, which are all crucial for sustaining and growing in today's competitive business environment.

Some examples:

• Improved Decision Making. Because they rely on analytics, facts, and empirical evidence rather than intuition or personal experience, data-driven decisions tend to be more informed and objective. This leads to better strategic decisions.
• Increased Efficiency. By analyzing data, companies can identify opportunities to optimize processes, reduce costs and improve operational efficiency. This can lead to streamlined workflows and reduced waste.
• Enhanced Customer Insights. Data analysis provides deeper insights into customer behavior, preferences, and trends. This allows companies to tailor their products, services, and marketing strategies to better meet customer needs and boost satisfaction.
• Risk Mitigation. Good data helps in predicting potential risks and uncertainties, enabling companies to develop strategies to mitigate these risks ahead of time. This includes a better ability to respond to external changes and challenges effectively.
• Scalability. Data-driven strategies support scalable business growth. By understanding market demands and operational constraints through data, companies can make informed decisions on where and how to grow.

But what if data are incorrect?

Yes, you heard that right. Although the “data-driven” argument sounds extremely appealing, having relevant, accurate and consistent data about the company itself, its core business metrics, market trends and so forth is often a challenge. 

So, what are the risks of running a business based on data that are not totally relevant, consistent or accurate?

Incorrect data pose significant risks that can affect various aspects of operations and strategic decision-making. For instance:

• Misguided Decisions. Easy one. Taking strategic decisions based on inaccurate data can lead to strategic missteps, operational inefficiencies, and financial losses. It can misguide businesses on critical issues like market trends, customer preferences, and competitive strategies. With the big risk being not just where a bad decision will lead the company but, more importantly, when will the leadership realize that things are going south… and why?
• Wasted Resources. Again, pretty obvious. Allocating resources based on flawed data analysis can result in wasted efforts, time, and capital. This could involve investing in the wrong projects, targeting incorrect market segments, or pursuing ineffective marketing strategies. Or just reorganizing your business in the wrong way (and maybe wait for a U-turn opportunity down the line?). 
• Damaged Reputation. Especially for companies that have some market presence and visibility (this is typically the case of fast growing public companies, but not only those), making public statements or decisions based on incorrect data can damage a company's credibility and trust with customers, investors, and partners. Once lost, reputation can be hard to rebuild.
• Financial Misreporting. Somewhat related to the one above, inaccurate financial data can lead to erroneous financial reporting, affecting investor relations, stock prices, and the ability to secure financing. 
• Decreased Employee Morale, General Dissatisfaction. A culture that continuously makes decisions based on faulty data can lead to frustration and decreased morale among employees, especially if they are aware of the data inaccuracies but are powerless to change the decision-making process. This can lead to a general dissatisfaction, loss of loyalty, and negative word-of-mouth among multiple stakeholders, including customers and partners, impacting long-term revenue and growth.

How do we prevent or fix data inaccuracy?

There’s an interesting TED talk called “Leadership in the age of AI”. In this interview, Paul Hudson , CEO at Sanofi, speaks about how big data has changed the way decisions are taken in a large, highly innovative organization such as Sanofi.

Paul, don't look, the data is not 100 percent correct…

After explaining how his organization thrives by making a smart use of big data, he addresses a concern that all of us faced one way or another: what if the data that we look at, that we carefully analyze, that we beautifully present in visual, multicolor dashboards… are not accurate? What if they are just wrong?

Well, then make it correct! Because the data is live!

In his interview, Paul Hudson explains: "If you really jump in and make it correct, it'll better reflect what you're doing. But if we wait for perfection it's simply not going to happen."

Boom. There you go. Do you really get better insights from bad data than no data at all? Well, for what it's worth, in my personal experience (and in situations of high uncertainty) this approach has often helped to make good enough decisions (as opposed to no decision at all, which is often detrimental to the business).

More specifically, finding data correlations and linking them to business goals and outcomes has proven to be more effective than trying to build mathematically correct causal relationships among data and between data and business outcomes. And yes, you normally don't need perfect data to identify correlations.

A “good enough” approach to data is often... good enough. 

Now. Let’s take this one step further. Can you rely on partially accurate data if you ask an AI to figure them out?

AI and Data Analysis

AI excels in areas where decisions are based on large volumes of data, patterns can be discerned, outcomes can be predicted with high accuracy. This makes AI particularly suited for routine and tactical decisions, such as:

• Operational Efficiency. Automating processes like inventory management, logistics, and scheduling, where decisions are repetitive and can be optimized based on historical data.
• Data Analysis and Reporting. Handling vast amounts of data to identify trends, generate reports, and provide actionable insights without human bias.
• Customer Interactions. Managing routine customer service inquiries through chatbots, personalizing user experiences on websites, or recommending products based on previous behaviors.

These areas benefit from AI's ability to process and analyze data far more quickly and accurately than humans, leading to increased efficiency and effectiveness in decision-making.

But how does AI deal with incorrect data? How important is the quality of data when you adopt AI for insights and predictions? Well.. it is important. Even very important at times. 

AI, particularly machine learning and data analytics, excels in processing and analyzing large volumes of data far beyond human capabilities. It can identify patterns, trends, and correlations that might not be immediately apparent to humans. However, its efficacy is contingent on the quality of the data it processes. 

• Data Quality Dependence. AI systems base their analyses and recommendations on the data they are trained on. If the data is biased, incomplete, or inaccurate, the AI's conclusions will be flawed, leading to potentially misguided decisions.
• Lack of Intuition. AI lacks human intuition and the ability to contextualize data in the broader spectrum of human experience and societal nuances. It cannot question the data in the way humans can, especially if the data contradicts known patterns or expectations.
• Objective Analysis. One of AI's strengths is its ability to perform objective analysis without the cognitive biases that humans may have. This can lead to insights that humans might overlook due to their biases or preconceptions.

I am not a robot

Human intuition is a product of experience, cognitive patterns, and subconscious information processing. It enables individuals to make quick judgments and decisions without the need for detailed data analysis. This "gut feeling" can be especially valuable in specific contexts.

• Complex Decision-Making. When faced with complex situations that involve ambiguous or incomplete information, human intuition can fill in the gaps, drawing on past experiences and cognitive shortcuts to make informed decisions.
• Creativity and Innovation. Intuition plays a crucial role in creative processes and innovation, where connecting disparate ideas and envisioning novel solutions are required.
• Emotional Intelligence. Of course. Human decision-making is not just about logical analysis; it also involves understanding emotions, both of oneself and others. Among other business areas, emotional intelligence is crucial in leadership, negotiation, and customer interactions.

Human decision-makers are extremely well suited for strategic and business-critical decisions. Indeed, such decisions require:

• Contextual Understanding. Humans can consider broader economic, social, and political factors that might not be captured in data but are crucial for long-term strategic planning.
• Ethical Considerations and Empathy. Decisions that have significant ethical implications or require understanding of human emotions are better handled by humans. AI lacks the ability to make value judgments or empathize with human conditions.
• Complex Problem Solving. Humans are better at dealing with ambiguity and complex problems where there may not be clear data or historical precedents to guide decisions. This includes navigating new markets, developing innovative products, or responding to unprecedented crisis.

Can we get the best out of both worlds?

Is there a winner out there? Should we trust human intuition and creativity over AI’s ability to process enormous amounts of data? Well, this is how humanity survived and even thrived for a few thousand years. But we have to acknowledge that things are changing very rapidly and, especially when decisions must be taken quickly and off of dozens, hundreds of data points… Humans can use some help from machines. 

Can we possibly avoid relying exclusively on either human intuition or AI and instead combine the strengths of both approaches? We surely can. For instance:

• Human Oversight. Humans can provide oversight, context, and ethical considerations in the interpretation of AI-generated insights. They can question AI recommendations when they conflict with known facts or ethical standards.
• AI as a Tool. AI can augment human decision-making capabilities, providing detailed analyses that can inform and enrich human judgment.
• Continuous Learning. Incorporating feedback loops where human decisions can inform and refine AI models can help in mitigating biases and improving the accuracy of AI analyses over time.
• Balanced Decision-Making. Human intuition has the unique ability to balance data-driven insights with smart and well pondered decisions, especially in areas where human values, ethics, and emotional intelligence play a critical role.

In Conclusion…

Where AI can process data at scales and speeds unattainable by humans, it lacks the ability to question data quality and contextualize information within the broader human experience. Human intuition, complemented by AI-driven insights, can lead to more nuanced, ethical, and effective decision-making.

For many businesses, a hybrid approach maximizes the strengths of both AI and human judgment.

• AI-Assisted Decision Making. Humans can use AI-generated insights for data-driven decision support in strategic planning, while applying their understanding of context and ethical considerations to guide the final decision.
• Human-Moderated AI Decisions. In tactical areas, AI can take the lead in decision-making but operate within parameters set by humans, who can intervene when situations fall outside of normal patterns or when unexpected events occur.

While there is a big debate on the advantages and risks of using AI in high impact decision making processes, some best practices in leveraging AI and human decision-making in business are arising, which can hardly be challenged. While AI is ideal for enhancing efficiency and accuracy in routine and tactical decisions, human judgment remains essential for strategic decisions that require a deep understanding of context, ethics, and complex problem-solving. A synergistic approach, where AI supports and enhances human decision-making processes, often results in the most effective outcomes.