How to turn business questions into analytics
At its core, the business intelligence selling point is simple — business intelligence software allows organizations to use data to answer their business questions. BI tools do this by providing useful data analysis that helps managers and executives make decisions.
However, the business intelligence experience is rarely that simple. Businesses can’t just plug in their BI tool and start getting their questions answered right away. These sorts of tools require work and knowledge to work properly. Organizations that invest in a BI tool are often disappointed that their tool isn’t as effective as they’d hoped.
Generally, this isn’t because the tool itself is ineffective (though there are cases where that might be true). Usually, if a business isn’t getting much out of their BI tool, it’s because they’re not using it to its full potential.
Since BI tools are so complex, it can be very easy for a business without much technical staff to underuse their tools. Data novices who just click around semi-randomly in a tool can’t usually produce useful data analytics or create useful visualizations. They’re very limited by what they know how to do and the data they can access.
Thankfully, many modern BI vendors are attempting to make their software easier for the average worker to use. These tools are designed to be self-service, meaning that businesses using them don’t need any technical staff to help them with creating visualizations or doing data analysis.
Even though self-service tools can help the average person use their BI tool without any technical support, they don’t necessarily make it so that anyone can design useful visualizations or do important data analysis without any training or expertise. Users of BI tools still need to know the fundamentals of data analysis to get anything useful out of their information.
This leads to a fundamental question that lies at the heart of business intelligence– how do users turn their business questions into useful analytics? How can users design analytics and visualizations that actually answer the questions that they have about their data?
The ability to turn an important business question into a visualization that can simply and effectively answer that question is one of the most important skill-sets in the entire data industry. Learning to do it effectively is essential for anyone who wants to work with data in a serious way.
Asking the right questions
To effectively turn business questions into analytics, BI users first need to ask the right questions. Usually, the actual content of the question isn’t the problem; it’s how it’s being framed and what the answers to the question should look like.
In general, BI tools aren’t very good at making value judgements or providing qualitative results. They’re much better at producing metrics, KPIs, and other numbers that can be measured and counted. That means users need to ask questions that can be answered with a metric, and not with a value assessment.
For instance, think of what’s probably the most basic business question — ‘Is my business doing good or bad?’ A human analyst like a financial advisor or consultant might be able to answer that by looking at relevant data points and synthesizing them into their own value judgment. A BI tool, though, can’t really provide that sort of analysis.
When using a BI tool to answer a question like this, users need to reframe the question in a way that the BI tool can actually understand. Generally, BI analysts do this by measuring metrics and KPIs associated with the question.
A ‘KPI’ is a key performance indicator. The idea behind using a KPI is that some metrics are indicators of more general, qualitative trends. For example, business health might be measured by a few different metrics, like revenue, click through rate, conversions, and so on. By tracking these metrics, we can track performance.
Instead of asking a BI tool a question like ‘Is my business doing good’, it’s much more useful to look for the KPI or KPIs that are important in that situation and measure those. That’ll change from business to business and is one reason it’s important to have good business sense when building out analytics and visualizations.
Some businesses may look at pure revenue when determining business health, while others may use a blend of multiple different KPIs to decide how their business is doing.
By using KPIs, users can turn general business questions like ‘How is my business doing’ into answerable, actionable data streams like ‘How is my revenue trending’ or ‘What’s my average conversion rate?’
This approach does rely on the user having at least some sense of what these metrics mean for their business. It’s not always clear what a KPI’s movement means; some metrics, like revenue, have clear connotations, but other metrics might not be as simple. Users have to make a value judgment on what a metric means and, more importantly, what to do about it.
How data analysis answers business questions
After a user has decided what question they need answered, it’s time to figure out how to answer that question using data. In some cases, this can be very simple, but in others, the data analysis required can be much more complex.
The first step in this process is selecting the data that will power the analysis. Usually, this isn’t terribly complicated, but users still need to consider what metric they want to measure, where they can get the data they need, and how to connect to their data source.
If a business wants to analyze its financial data, for example, they need to connect their BI tool to their accounting software or manually upload that data in some way. A BI tool can’t analyze metrics it can’t access. From there, users need to make sure they’re using the right data; it’s very easy to accidentally use data from the wrong section of a business or from the wrong date.
During this process, users need to think about how to transform their data to get the best result. Data transformation is a complex topic, and we’ve got all sorts of articles to help you learn about it. Right now, it’s enough to just say that data transformation is an essential part of the data experience.
After the data has been collected and transformed, it’s time to analyze it. In a lot of situations, data analysis is very simple. For example, if a user wants to see how earnings are trending over time, it’s very easy to graph the relationship between earnings and time.
In other situations, though, more complex analysis might be necessary. It’s important to know the question, know the metric or metrics, and know how the metric needs to be expressed to best answer the question. That’s not a skill that comes naturally to most, and users generally need some data training to figure out these more complicated use cases.
How to visualize business questions
Data visualization is the last step in the data analytics process, and it’s often the most challenging for users. Data analysts have to build graphs and dashboards that best answer their business questions while providing a good user experience.
Choosing the best visualization for a given analytic is a complex process, and users often need at least some training to do it effectively. We have a handful of articles that can help new business users and data analysts better understand which visualizations are the best for a given situation.
There are a few key guidelines that beginner visualization designers can keep in mind.
First, it’s best to prioritize readability over style. Often, especially in self-service BI tools, it’s tempting to use the flashiest visualizations possible to express analyses. Novice designers should resist this urge and make sure their visualizations are legible and understandable.
Second, the simplest visualization isn’t always the best choice. More casual designers often stick with the visualizations they’re familiar with, like line charts or bar graphs. While these visualizations are generically useful, they can lack a lot of the nuance that a more specialized visualization has.
Lastly, it’s important to step back, look at the visualization holistically, and ask whether or not it’s actually answering the business question. As a worker moves an analysis through the data process, the reality of the visualization can get disconnected from the vision.
Since the data process can distort the goals behind a visualization, it’s important to do a reality check occasionally. Does the visualization look generally correct? Do the numbers look right? Does its conclusion make sense? Can a non-expert use it to draw conclusions about a business question?
Beginner visualization designers can often make simple mistakes that distort their analytics, especially if they need to use calculated fields or SQL queries. While it may seem unnecessary, it’s good practice to make sure everything’s pointing in the right direction.
Turning business questions into analytics
When businesses invest in a BI system, they often expect to see the benefits immediately. However, good data analytics doesn’t spring up overnight. Users have to create the analytics themselves, and that can be a major challenge.
With self-service BI tools, more and more people without any data analysis experience are building their own dashboards and visualizations to answer their business questions. While this change is generally good since it allows more people to interact with data, it does come with some downsides.
Users don’t always understand how to actually turn their business questions into useful visualizations. Building an effective, actionable visualization is a hard job, and data beginners often struggle to make their data dreams a reality.
There are some strategies that data analysts can use to translate their business questions into data analytics. First, ask questions that can be answered with KPIs and metrics rather than open-ended, qualitative questions. Next, make sure to collect, transform, and analyze the data correctly. Lastly, have a good understanding of data visualizations and use the right one for your question.