How to Become a Business Data Analyst in 2022
The days when a business data analyst only needed to be a spreadsheet ninja are long gone. Modern-day business analysis requires robust data analysis skills and knowledge in data science methodologies like predictive analytics or causal inference. The familiarity enables you to support non-technical teams and bridge the gap with IT-based departments. In other words, you become an analytics translator.
Nevertheless, you need to have more than the ability to correlate data to identify problems for the business. Your advanced hard skills simply get your foot in the door. It’s the soft skills that keep it open. Highly proficient abilities in communication, stakeholder management, and business acumen are keys to becoming a critical resolution key for senior management.
Here are the hard and soft skills needed when starting as a business data analyst.
Improve your hard data analyst skills
In my current profession, all interns and junior colleagues are incredibly skilled. However, since they do not have years of business analytics experience, we look at what else they bring to the team. Either it’s self-taught or learned at school through data and computer science classes. These are the same principles considered during my hiring process.
To put it another way, focus your post-education energies on improving your skills as a data analyst. By doing so, you help stakeholders commit to positive business decisions.
Last Updated January 2023
Learn Causal Inference & Statistical Modeling to solve finance and marketing business problems in Python and R | By Diogo Alves de ResendeExplore Course
Have a solid statistical foundation
Statistics is the root of business data analytics. This science empowers you to comprehend a company’s data better. Additionally, it gives you a chance to see the most common biases in business. For instance, you comprehend the definitions behind omitted variables, mutual casualties, or selection bias.
You need to completely understand concepts like the mean, correlation, or t-test and how they’re applied to an organization’s data. Not only does a solid statistical foundation enable you to be more secure in your data analysis, but it also helps you move to the next level in your journey.
The main branches of business data analytics
Econometrics, segmentation, and predictive analytics are the main branches of data analytics. While it’s both tremendously underrated and unpopular, this application of statistical methods used in econometrics is a powerful science. The algorithms within this science focus on the decision-making process.
What is also insightful is segmentation and the associated techniques. Here, you separate markets into groups of prospects and customers with similar characteristics. This includes what they purchase.
Finally, I recommended predictive analytics as the third priority to study. This form of statistical science is currently popular among businesses. Its goal is to predict possible future outcomes from a series of variables.
To illustrate these areas of study, here are some examples from my professional career. The first uses the matching algorithm for econometrics. The second applies the k-means algorithm for segmentation.
Matching is an econometrics algorithm to find causal incrementality with non-comparable groups.
I work for the German fashion company Zalando. There are two language possibilities for their website: English and German. The English version was introduced in 2018. However, a few months after launch, the question of whether it was worth it was put on the table.
The English website is dedicated to immigrants and ex-pats. This profile is entirely different from the average German. Expats are younger, live almost exclusively in city centers, and have a higher income, just for the people’s characteristics. Thus their buying behavior is different.
As the English and German consumers aren’t comparable, matching was necessary to determine whether the separate website was worth the investment. Thus, I tried to determine every characteristic that makes a Zalando customer. I opted to go with 15 of them.
For the robustness check, I used the repeated experiment approach and created a 1,000-fold repetition. After the correlation of the data, I showcased the methodology and approach to senior management and colleagues.
Top courses in Data Analysis
From value-based segmentation to a customer behavior
As for segmentation, my first experience was in a closed contest for a conglomerate. The company had a value segmentation in place, which helped them prioritize its customers. They knew the worth each customer brought to the organization. In turn, this data enabled them to concentrate their focus.
However, despite the robustness of their value segmentation, there was no behavioral separation. Hence, the company requested one for the contest. To obtain the correct information, the organization supplied us with a massive amount of anonymized data.
The total number of variables was more than 600, which goes beyond a concept to keep problems to a minimum. You don’t want to encounter issues like multicollinearity, the curse of dimensionality, or even interpretation. Thus, the first thing was to trim the data set.
Our first step was hypothesis-driven: What elements could define the customer’s behavior and be insightful simultaneously? The answer got us down from 600 variables to 30. Next, I ran a correlation matrix to see where we had variables explaining the same thing, and it resulted in the removal of 15 more characteristics.
We then used the k-means to cluster the data further. It produced seven segments, and we presented our findings to the contest. Although we lost, the experience was amazing.
The reason to understand the three main branches is to learn what can be done better. For instance, for the contest, seven segments were on the upper end of complexity. Perhaps simplicity would have been wiser.
Additionally, new customers weren’t part of the solution because no information was available. We needed to create a process for them, like a temporary eighth segment. Then, we would have allocated it once the company had updated information.
The idea’s presentation and communication also required improvements. It was too technical rather than insightful. This leads to a discussion on soft skills.
Communication is the next step
Proper communication becomes increasingly relevant as you expand your role as a data analyst. It doesn’t matter if you have a degree in a business field, information technology (IT), or English. Proper communication isn’t something to ignore.
This doesn’t mean you’re prone to instantaneous inspirational speeches. I don’t do this. The fact of the matter is, I read from a script for my videos. Yet, I still have to be prepared to present a stripped-down version of the results you compiled.
Let me share some of the principles I follow for improved communication.
My presentations are uncompelling and confusing If I don’t know everything about my topic or data. Each time I have a presentation, no matter how small, I make sure I understand the facts.
I attempt to create an outline of the story in my mind. This enables me to speak fluidly and with confidence.
I once got feedback that my body language seemed closed and conflicted. I didn’t realize this until I gave a presentation in a room with mirrors. I observed myself and saw what the person meant.
Now, I try to show more openness. I keep my back straight and body loose. I keep a small smile on my face throughout the presentation.
Be open and inquisitive
I’m confident in what I say because I put effort into knowing the facts. At the beginning of my career, if someone presented a concept that contradicted my own. I’d immediately dismiss it and then counteract the idea with my own. This came off as dry and abrasive.
Today, instead of an instant dismissal, I ask the speaker questions. If they’re prepared and present a logical argument, then I consider the idea’s feasibility.
Granted, being open to questions and criticism isn’t easy. However, you should adopt honest curiosity to create a pro-active environment.
Your role as a business data analyst is considered the second most proficient stakeholder in an organization. Thus, the hard and soft skills you learn and increase your competence in the eyes of management.
It’s hard work to achieve this position. You need to establish solid analytical skills and its three main branches. At the same time, consider how to properly communicate your experience and the information you compile.
Through the combination of soft and hard skills, you gain an advantage in landing a position many organizations demand.