Smarter Business Decisions: Using Data Wisely
Smarter Decisions Start with Data
Since we were kids we’ve learned what’s right and what’s wrong by making mistakes. For example, our parents say “Don’t touch the hot surface.” We don’t understand, we touch it, and our hands get burned. We learn not to do it again.
You might wonder how this relates to data. Sometimes entrepreneurs overlook data, rush decisions, and focus only on quick results. Similar to the example above, think of data as a guide. When data tells us not to do something and founders ignore it, they make mistakes that set the business back or even cause it to fail. Because we have emotions, we use our feelings when we make decisions. However, data has no emotions; it reports patterns objectively. Either way, we must use data in ways that yield accurate, unbiased results. So, the main challenge today is how to use data well. I’ll discuss this later in the article.
Why Businesses Use Data
Data plays a significant role in business. Businesses use data to develop, formulate strategies, predict the future, manage risks, or plan.
Smarter Decisions
Companies use data in several ways. First, data helps companies make better decisions with well-backed evidence. They avoid relying on gut feeling; instead, they analyze data to understand what’s happening and avoid investing in ineffective strategies or projects. As humans, we make decisions based on our knowledge, experiences, and emotions. However, this doesn’t always work. Data widens our view and reveals patterns we’d miss in anecdotes.
Understanding Customers
Companies understand customers by analysing digital footprints to see who they are, what they need, and what they do. This longitudinal data is better than a brief ten-minute conversation where people might not reveal everything openly. I recall the dilemma of honesty: it’s when telling the full truth may hurt someone or cause problems, so people are unsurewhether to be completely honest (Zolkefli, 2018). So we should focus on real data rather than dealing with these uncertainties.
Targeted Marketing
Furthermore, businesses use data to prepare targeted marketing campaigns. This is where data moves from opinion to action. Data makes marketing far more effective. Instead of one-size-fits-all strategies, companies analyze data to segment customers and personalize campaigns.
I’ve always believed that if there is no demand, a product will not sell under normal circumstances. For example, if we meet a medical student who has no need for a finance textbook, selling the product directly would be difficult. However, if we understand their interests and challenges, we can create a connection between the product and their needs. For instance, if I know they struggle with managing a budget, I can spark their interest by presenting the book as a practical guide to personal finance, budgeting, and saving money. This shows that using data to understand customers is vital and businesses actively rely on such information to align products with customer needs. Businesses analyse customers’ data and develop personalised ads to attract more customers.
Operational Effectiveness and Risk Management
Data is also used to enhance operational effectiveness and to manage risks. For instance, companies create performance metrics by analyzing internal operations to identify inefficiencies and patterns that can cause problems in the future.
Innovation
Data isn’t only used to optimize the present but also to foster innovation. By using data, we can find unmet customer needs and ways to fulfill them with creative ideas.
This is just a small part of what data can do. There are many other ways to use data.
The Hidden Power Behind Everyday Choices
We can see successful examples of data tracking in daily life. A basic example is when we search on Google and type one or two words in the search bar; Google shows suggestions that start with those words. These come from a large dataset of other people’s searches.
Also, when we want to watch something on Netflix, we see suggested movies. Those recommendations come from algorithms that use our activity data (collected through tracking and analyzed with analytics).
About 80% of viewing on Netflix starts from homepage recommendations rather than the search bar (Wired, 2018). That illustrates how effective data-driven recommendations are in practice.
Most of us check restaurant rankings when deciding where to eat. For instance, Yelp aggregates reviews and ratings from many people to help guide our choices. When we see a restaurant with a high ranking, we tend to trust that information and expect its food or service to be better than those of lower-ranked restaurants. In this case, we rely on crowdsourced data.
These are just a few examples of data tracking. Even when we don’t notice it, it works behind the scenes to improve our choices. Whether it’s live traffic data that helps us cut commute time or music apps that build playlists from our interests, data makes life easier and more convenient. So, what we experience as users becomes the raw material for how companies decide.
The Challenge: Using Data Accurately
The main point is to use data accurately, with sound strategy and method. This is one of the most challenging parts of the process. If we have a huge dataset but can’t use it, it’s pointless to have it. In some cases, using data in the most effective way is harder than collecting it. I won’t go deeper into the data analysis part in this article.
In simple terms, data analysis steps include defining the question, collecting the right data, cleaning it, analyzing it, and sharing the findings. Doing each step well is harder than it seems. Here’s a brief personal example from my own work:
During my bachelor’s degree, I created YourTravelBudget, a website that provides travelers with travel tips from locals. I searched the internet for ideas on how to structure the website. Because I assumed that travelers preferred long, blog-style articles, I ignored any criticism or examples that suggested short, structured articles. I launched the site with lengthy posts, but the results were disappointing. Users spent little time on the website, and many left without finding what they needed. To get a clear answer, I conducted a survey of randomly selected people and analyzed the data objectively. The results revealed that most people preferred concise, structured information such as price cards, checklists, and brief guides over lengthy paragraphs. After that, I realized I had fallen into a basic trap. This trap has a name.
The Traps and Biases
Confirmation Bias
It is called confirmation bias. Confirmation bias is the tendency to seek, interpret, and remember information in a way that confirms one’s existing beliefs (Simply Psychology, 2023). This is a common bias people fall into, often unintentionally. It stems from human psychology. Sometimes, to protect our self-esteem, we seek approval for our opinions.
As Casad notes, to make themselves feel confident, people tend to look for information that supports their existing beliefs (Casad, 2019).
This bias often leads to unexpected outcomes in life. I’ll focus on the business side of it rather than daily-life examples.
Let’s assume that we have a business and want to create a marketing campaign on different platforms such as Facebook, Instagram, and Google. If we believe Instagram ads work better, we’ll pay more attention to the months when Instagram ads did well but ignore other months when other platforms did better. If we’re running a business or want to start one, we need to avoid this trap.
Another example: I want to import products from China and sell them in my home country. I’m really interested in X product, but I ignore the overall market demand for this product only because I love it. That interest can push me to look only for positive information and feedback about the product because I want to buy it. In the end, I see that most people are not interested in buying that product. The fact that you like something doesn’t mean others do as well. In this scenario, that bias misled me, and my entrepreneurial attempt failed.
Of course, you can avoid this bias! For instance, set rules before you act. Let’s say your aim is to increase sales by 10 percent.
- Define clear rules that you will follow.
- After defining these rules, collect all relevant data (not just what supports you).
- Give equal weight to all evidence you collect
- Example: In this case, even if you believe that aggressive marketing strategy is good to increase sales, always remember there are other factors too.
Another solution is collaborating with other people. Even if you’re confident about your opinion, ask for others’ opinions. Collect all this information and analyze it objectively. Someone might tell us something completely different, and it can make us think in a way we’ve never considered. Even if we avoid confirmation bias, another trap appears.
Survivorship Bias
Another common trap is survivorship bias. Survivorship bias arises when a successful subgroup is mistaken for the entire group because the failure subgroup is invisible (Survivorship Bias — The Decision Lab, n.d.). People often look at successful subgroups and ignore others with similar characteristics.
For example, when we read the success story of Bill Gates, we see that he dropped out of school, started a business, and became successful. People might think, “Okay, I should drop out of college and start a business,” but they miss one point. What happened to people who dropped out of school to start a business and failed?
Another example is an entrepreneur who wants to copy the habits of successful businesspeople to be successful like them. What’s missing is the fate of entrepreneurs with the same habits who failed.
We shouldn’t focus only on success stories while ignoring alternative cases. A well-known example explains this bias. Historically, the U.S. military wanted to armor planes that had the most bullet holes when they returned from World War II. According to statistician Abraham Wald, they should strengthen the areas without bullet holes because planes hit there never returned (Wallis, 1980). The key insight was to reinforce the areas without bullet holes.
From the business perspective, if we want to do targeted marketing campaigns with the data we have, we tend to focus on previous successful campaigns and try to reuse or improve the same patterns. However, we forget to analyze the unsuccessful ones. Maybe we’ll find better outcomes or a way to attract more customers when we analyze them.
Let’s assume we’re business leaders and our company manufactures high-quality laptops. We want to offer new models with high RAM, GPU, and processor performance. First, we find similar laptops that did well in the market. We see that they used high-budget marketing campaigns and people preferred these laptops because of their high-quality hardware. We try to implement similar strategies to be successful.
What about other companies that produced laptops with the same characteristics and similar marketing campaigns but did not succeed?
We must analyze why they failed, too. If we take failure stories into account as well, our analysis and decisions will be better and more effective.
Correlation vs. Causation
Next, we need to make the correlation–causation distinction clear. A correlation occurs when two variables change together. Causation is when changes in one variable directly produce changes in another (Madhavan, 2025). That means just because two variables show a relationship, it doesn’t prove that one causes the other (Wikipedia contributors, 2025b).
For example, we observe that ice-cream sales and drowning incidents increase at the same time. At first glance, we think there’s a strong correlation. However, we forget about other factors: hot weather. In summer, people buy more ice cream and go swimming, so both numbers rise. Eating ice cream doesn’t raise the number of drowning incidents, but hot weather does.
We must take other factors into account while analyzing data. Otherwise, we can get wrong results, and it can ruin our entire analysis. This hidden factor is a confounding variable. A confounding variable is a hidden factor that influences both the putative cause and the effect, making it appear there’s a direct link when there may not be one (Schroeder et al., 2016).
Personal example:
I opened a customized clothing store and did marketing campaigns every week with different strategies to boost sales. One day, I analyzed all the results of my campaigns and compared them. I saw that last week my sales doubled compared to other weeks. The first thing I thought was that the strategy I used last week worked better and boosted my sales, so I should keep using it going forward. What did I forget here?
Yes, a confounding variable. Even if I saw a direct link between my ad campaign and sales, there was a hidden factor. Last week was a national holiday, and people bought more clothes than in other weeks. I didn’t count that factor, so it misled me. Always ask: “What else could explain this pattern?”
To avoid this problem, businesses should run small, controlled trials that change only one element at a time while holding everything else constant. For instance, to compare findings equally, businesses can run advertisements in one location but not in another over the same time period if they want to determine whether Instagram advertising boosts sales. Also, companies can separate true causes from coincidence by varying who sees the ads, tracking variables like holidays or sales, and adjusting for them with statistical techniques like regression.
Conclusion
However, power comes with responsibility. We face several challenges when working with data, and if we’re not cautious, it might lead to misdirection. The examples I covered in this article were limited to confirmation bias, survivorship bias, and the correlation–causation distinction. These are typical pitfalls, but we face many more difficulties in gathering, analyzing, and interpreting data on a daily basis. So what should we take from all of this?
In conclusion, data helps only when we use it with care. In other words, data adds value when it’s applied precisely, objectively, and carefully. Having the largest dataset is not the point. It includes asking pertinent questions, conducting multi-angle analyses, learning from both achievements and setbacks, and becoming conscious of our own analytical and cognitive biases. Data becomes more than just numbers on a screen if we use it wisely and understand its limitations. It turns into a guide for smarter decisions, more robust companies, and more intelligent long-term plans.
References:
Casad, B. (2019, October 09). Confirmation bias. https://www.britannica.com/science/confirmation-bias
Madhavan, A. (2025, September 4). Correlation vs Causation: Learn the Difference. Amplitude. https://amplitude.com/blog/causation-correlation
Schroeder, R. D., Carey, T. A., & Wooten, L. P. (2016). Applied multivariate statistics for the social sciences (6th ed.). Routledge.
Simply Psychology. (2023, June 22). Confirmation bias in psychology: Definition & examples. https://www.simplypsychology.org/confirmation-bias.html
Survivorship bias — The Decision Lab. (n.d.). The Decision Lab. https://thedecisionlab.com/biases/survivorship-bias
Wallis, W. A. (1980). The Statistical Research Group, 1942–1945: Rejoinder. Journal of the American Statistical Association, 75(370), 334–335. https://doi.org/10.2307/2287454
Wikipedia contributors. (2025b, May 30). Correlation does not imply causation. Wikipedia. https://en.wikipedia.org/wiki/Correlation_does_not_imply_causation
Wired. (2018, January 18). How Netflix’s recommendations system works. Wired. https://www.wired.com/story/how-do-netflixs-algorithms-work-machine-learning-helps-to-predict-what-viewers-will-like/
Zolkefli, Y. (2018). The ethics of truth-telling in health-care settings. Malaysian Journal of Medical Sciences, 25(3), 135–139. https://doi.org/10.21315/mjms2018.25.3.14