Problem-Solving — What Every Data Analyst Must Have
Wait. Wait. Wait. Don’t go beyond this paragraph yet. Imagine problem-solving as a data analyst. Imagine the scenarios. The benefits. The approach. Now write down what’s in your head. Ensure you write oh because we’ll be using it at the end of this article.
Now continue reading🤗
Maybe you’ve seen different takes on how Data analysts shouldn’t just be building dashboards or reports but solving problems with those dashboards. Maybe you don’t like those takes. Provided the dashboards are beautiful and people use them, then that’s okay. Your responsibility as a data analyst is done.
Well, as much as I would love to agree with you as my reader, you’re wrong if that has been your thoughts all along.
But don’t worry if you’ve been finding it difficult to solve problems as a data analyst. I believe all you need is a shift in your mindset. The moment you understand what problem-solving is, you’ll know where and how to apply the skill.
Because I want you to relax and read this article while sipping cold orange juice (keep my share by the way), I won’t be talking in technical terms. Nope. I won’t be discussing all those problem-solving frameworks. They are important oh and I’ll link to articles that go into detail about them at the end of this article. But I’ll be doing something different. Something exciting! Something I always do anyways. I’ll be informing you with stories. And much more. Okay, no spoilers. Just read this and see how impactful it will be.
Let’s start with why you need to solve problems as a data analyst…
Why Problem-Solving for Data Analysts
Problem-Solving attracts business value.
Have you been guilty of not sticking to a friend because you feel like you’re not gaining anything from him or her? It must not be financial gain but at least you know you’re gaining something that’s why you’re with this person.
That’s how companies see their employees too. Imagine being a data analyst in a company but you’re not providing value because you’re not even solving any problem. You just build dashboards and don’t make the attempt to get feedback on who is using it and if their problems are being solved.
That company won’t see you as a person of value and not long after, they’ll certainly drop you off where they picked you from. Yes, it’s as harsh as it sounds and it happens. But you can change that.
Problem-Solving is required to milk your data dry for insights and data-driven recommendations
Think for a second. How do you approach a dataset you’ve worked on? Did you just start making visuals without having any questions to answer? If yes, then you’re not solving any problem.
Efficient problem-solving skill is driven by an inquisitive and curious mind. It’s only when you’re in this state that you can ask your data lots of questions and see if it gives you answers or not. If it gives you answers, you still need to go further to see if those answers align with the problem the organization is experiencing before bringing it out to stakeholders.
Yes, it’s a whole lot of work no doubt. But you signed up for this, right? So you can definitely do it💪.
The truth is if I go on and on why you should solve problems as a data analyst. All of the items in the list will still be a subset of the main goal: providing business value. Hold that in your mind or write it somewhere.
The next question you might be asking or might have asked is, “What does it mean for a data analyst to solve problems?” I’ll be showing you different scenarios of what is and what is not problem-solving.
What Problem Solving Is not
Making a presentation to stakeholders on several factors that affected company sales and eventually getting them overwhelmed.
What Problem Solving Is
Making a presentation to stakeholders showing one or two factors that greatly affected company sales while giving them recommendations on how to tackle the cause so that sales can improve.
What Problem Solving Is not
Having several unrelated metrics/KPIs on a dashboard.
What Problem Solving Is
Having several charts on a dashboard that answer sub-questions that all come together to answer one big question.
What Problem Solving Is not
Delving into analyzing data given to you without asking stakeholders questions on their needs and challenges they are experiencing
What Problem Solving Is
Asking stakeholders questions to know what their challenges are before even collecting the data. Identifying other unstated needs during your analysis to know if their needs are actually the true problems or if there are more hidden problems.
What Problem Solving Is not
Making one dashboard or presentation to accommodate different kinds of stakeholders like Managers, C-executives, department heads, and team members.
What Problem Solving Is
Tailoring a dashboard or presentation to suit one kind of stakeholder being aware of the fact that metrics/KPIs that C-executives are interested in are different from the metrics/KPIs that managers are interested in.
(Exhales). Giving several scenarios like this can take a whole day or even more. I tried to make it generic because I don’t know the industry you’re currently in. One thing you should know is, these scenarios can be applied anywhere. If you need a more specific scenario, I told a story in this LinkedIn post.
And I’ll be telling another story here😊. This story will be showing you how a data analyst solved a problem using the problem-solving framework recommended by McKinsey as summarized in this article.
Sarah just joined a fictional e-commerce company called Nick-Mart as a data analyst. The company complained about high shopping cart abandonment rates and she was asked to investigate and figure out a solution to the problem. Sarah got on a meeting with the relevant stakeholders to ask when they noticed this problem. She asked them about several events that might have happened during the period they noticed the problem. They mentioned that they added more products, removed the low-performing products, and changed their website user interface during this period.
With their answers, Sarah was able to generate some hypotheses which include
- The addition of new products had an effect on the shopping cart abandonment rate.
- Removal of certain products had an effect on the shopping cart abandonment rate.
- Changes in the website user interface had an effect on the shopping cart abandonment rate.
When Sarah carried out her research, she realized that the addition of new products and the removal of old products could not affect the shopping cart abandonment rate because customers already pick their products before abandoning them. If they didn’t like the product, they wouldn’t pick it in the first place so she decided to begin with her third hypothesis: Changes in the website user interface had an effect on the shopping cart abandonment rate. She did this while keeping her mind open to other factors that could come up at some point.
Sarah didn’t just limit the hypothesis to this. She was open-minded to other factors that could contribute to this problem.
She collected the necessary data and dived into her analysis where she figured out that there was a high correlation between the change in the website interface and the shopping cart abandonment rate. To confirm this, she carried out an A/B testing alongside some other persons on the team where the users were split into two and directed to two different website interfaces while the shopping cart abandonment rate was measured on these two interfaces.
As expected, there was a significant difference. It turned out the old interface had a lower shopping cart abandonment rate than the new interface.
Sarah had probably figured out the reason for the high shopping cart abandonment rate so she made a presentation starting with her conclusion that they should go back to the old interface. In her presentation, she continued further to show why she made this conclusion by briefing them about the A/B test she carried out and showing them the results. At the end of the presentation, they saw reasons why they should go back to the old interface and obliged.
But she didn’t stop there…
She joined the team that would implement this to measure the shopping cart abandonment rate and other metrics that measured the performance of the e-commerce website as a whole. She built a dashboard to track these metrics so that the relevant stakeholders could also see them and so she could monitor them and identify any problems that could be encountered in the future.
Now this is what problem-solving is! Take note of how she already developed a hypothesis at the very beginning before even collecting the data. She was able to get this hypothesis because she asked stakeholders relevant questions. Being a good problem solver as a data analyst requires you to listen!
This takes us straight into the opinion of several experts in the data field. I got to communicate with them and was able to get their perspectives on problem-solving. Yes, I went that far because I wanted to serve you hot hot!
The key is to actually identify the problem first. So many times we set out to solve the problem that our stakeholders tell us is a problem, just to spend hours trying to solve it. But 3 hours in, we realize that the problem presented to us really isn’t the problem at all. So the key is to listen, understand, and determine what the real problem is regardless of what the stakeholder tells us.
Kevin Flerlage, Data Visualization Consultant at Moxy Analytics
In addition to technical skills, effective problem-solving also requires strong communication and collaboration skills. As a data analyst, it is essential to be able to communicate complex information in a clear and concise manner to stakeholders and team members.
Komolafe Seyi, Business Analyst at RightClick Solutions
I’ve seen a lot of people talk about analytical thinking, communication, critical thinking, attention to detail, etc but I’ll say that you should be creative. Being creative makes you stand out in every form whether in communication, or thinking( critical or analytical), and that will reflect so much on how you tell the stories you’ve seen in your data.
Ayoade Adegbite, Freelance Data Analyst and Data Analytics Tutor at CareerFoundry
My biggest suggestion is to learn how to become a good listener. That’s more than just hearing the specific words they are saying but listening for what they are not saying. Get to the root of what they need rather than what they say they need. And it’s also important to know how to ask good questions…questions that help you to truly get to the root of what they need.
Ken Flerlage, Data Engineering Consultant at Moxy Analytics
Problem-solving is the data analyst’s secret weapon for creating order in the midst of intense chaos. This means that if you’re poor at putting your thoughts together to solve problems, you’re very likely going to be poor at being a data analyst. Being a good problem solver sets you apart. As a must-have skill for every data analyst, problem-solving can be learned. My all-time recommendation is to always have a requirement document and a problem-solving plan. The requirement document tells you what your audience expects, while the problem-solving plan guides you on how to deliver this expectation.
Linson Abah, Analytics Engineer at SafeBroda
Problem-solving gets better with domain knowledge or experience which is why insights communication can still be permissible for Junior roles as this is a repetitive process in most industries. For Senior roles, ownership and leadership come into the bucket as well. Seeing beyond the insights and leading the solution. Being able to communicate as a skill actually drives proper investigation.
Bukunmi Adebanjo, Data Analyst at Nutrien
Perhaps one of the biggest problems in the problem-solution approach lies in technological packaging. When everything becomes a tool, even reasoning. This effect, which I call “self-induced collective dumbing down”, has affected the way data analytics professionals solve their clients’ problems. In summary, the practice of constantly using ready-made solutions, and adapting them to the problem, is distancing analysts from the legitimate solution. This is the bad news. The good news is: Go back to Cartesianism. Descartes, influenced by Aristotelian logic, has already taught us how to solve problems.
Luciano Gavinho, Data Science Specialist at Newoway
When you want to solve a problem, check the most efficient way of solving the problem, the fastest approach to solving the problem, optimum method. Focus on optimization. Businesses are more interested in people that can deliver value. It’s not just about the tools.
The root cause of an analysis is a problem. If there is no problem there wouldn’t be any analysis to begin with. Also to measure the effectiveness of any analysis, check if the result solves the stakeholders' problems or reveals insights. Above all problem solving is a must-have skill for a Data Analyst.
Abioye Hamid, Freelance Data Analyst
Effective problem solvers in the data field have a skill set that allows them to look beyond the surface-level data and identify the root causes of a problem. It’s important for data analysts to have a deep understanding of the business problems they are trying to solve, as well as the context in which those problems exist. By taking the time to understand the business context and objectives, and by developing a skill set that goes beyond basic statistical techniques and visualizations, data analysts can become true problem solvers and provide valuable insights that drive business value.
Jasmin Jusufbegovic, Data Analyst at Volvo Cars
I see problem-solving as a skill that involves three steps: identifying problems, analyzing to find the root cause, and finding solutions to address them. The technique I recommend for data analysts to use is the 5W H framework. The 5W H technique simplifies problem-solving by providing a clear and straightforward framework for analysis. Instead of getting lost in a sea of data and details, the technique helps to focus your attention on the most critical aspects of the problem. By narrowing down the scope of the problem and identifying its underlying causes, you can develop effective solutions that address the root cause of the issue. With the domain knowledge and understanding acquired about the problem and the root cause of the problem identified, providing techniques and steps to solving the problem is the next step to problem-solving.
Pheww! That was a long journey for us. I’m hoping you now understand what it means for you to solve problems as a data analyst. Even though you didn’t take anything from my perspective and my stories (which I know is not possible), I’m sure you learnt one or two from the perspectives of the data experts I mentioned.
Before we part ways (for a short while of course), I want you to know that it’s not just about knowing but practice. You might read about one million problem-solving frameworks but if you don’t put them into practice, you’re not a problem-solver. So yes, start from somewhere and start solving those problems even in the littlest projects. Know that I’m rooting for you.
Oh yes! One more thing. Go back to the first paragraph of this article. Yes, that thing I told you to write. Does it still stand now that you’ve read this article? Or has it changed? Tell me in the comments.
Till we meet again!🥰
