Diagnostic analytics enables your business to take a dive deep into why something happened—whether it’s a decrease in monthly sales or sudden increase in membership subscriptions. This critical information leads to more informed, data-driven decision-making across the enterprise.

Analytics play a critical role in decision-making for most companies today. Understanding what the data says about your customers, your brand, your products, and the overall business is essential to remaining innovative and competitive.

Most businesses rely on descriptive analytics to tell them what happened and then jump to predictive analytics to get a grasp on what will happen, and sometimes go as far as exploring prescriptive analytics to choose a course of action. Within the four pillars of modern analytics, there is one step that is too often skipped during data analysis—diagnostic analytics, which is also known as ‘root cause analysis’.

In this blog, we’ll discuss what diagnostic analytics is and how it’s different than other types of analytics, provide use case examples, explain tool options, and take a deep dive into one of the tools we recommend using for diagnostic analytics—Sisu.

What Is Diagnostic Analytics and What Makes It Different from Other Types of Analytics?

To understand how you can use diagnostic analytics, you need to first put it into context with how other types of analytics work. The four pillars of modern analytics include:

  • Descriptive analytics uses historical data to tell you what has already happened. Most organizations with some level of maturity on their analytics journey are already doing some degree of descriptive analytics.
  • Diagnostic analytics puts context behind the “what” by helping organizations understand the “why”. Taking this step first may lay the groundwork for predictive analytics, or in some cases, might even negate the need for predictive analytics because you can solve the issue by understanding the root cause of the problem. For example, if you are investigating why there is an increase in customer churn and diagnostic analytics reveals it’s because customer service is poor, you can take immediate steps to improve customer service to improve retention.
  • Predictive analytics is the next step. It also uses historical data but does so to predict what will happen in the future. In the example above, if you had an increase in customer churn and diagnostic analytics revealed that it is because certain promotional deals weren’t incentivizing customers to renew, predictive analytics can be used to help predict what kind of promotional deals will result in more renewals.

Predicting Customer Behavior Using Data Science Starts Here

  • Prescriptive analytics recommends automated actions you can take to affect those outcomes. It is about empowering the people who are making decisions to make the right decisions based on predictions. Understanding the “what” and “why” will lead to more confidence behind the decision-making.

What Are the Different Use Cases for Diagnostic Analytics?

Use cases for diagnostic analytics are very similar to those of predictive analytics. Ultimately, if you are collecting data and analyzing “what has happened”, the next logical step is to understand why it has happened—what specific events and circumstances led to that outcome. Some common use cases include:

  • Marketing and Sales Analysis: Diagnostic analytics can help you look at specific aspects of a marketing campaign in relation to sales data—to help you understand what about it performed well versus what didn’t. It can answer questions such as:
    • Why did sales drop in the first quarter of this year?
    • Why did this specific segment of customers not renew their subscriptions last fall?
    • Why did the weekly promo deal lead to a 42% increase in purchase size this week when it’s traditionally led to no more than an increase of 20%?
  • Anomaly Detection: Whether it is financial fraud or hidden performance opportunities, understanding why something is occurring can help to effectively address an issue. Diagnostic analytics can look at specific root causes in correlation with less known variables to help answer questions such as:
    • Why did mineral markets have a sudden sell-off?
    • Why does system downtime spike at the end of each month?
    • Why did customer adoption skyrocket?
  • Part Failure and Optimization: For supply chains, operating at full capacity and preventing downtime is critical. Analysis into why breakdowns occur can help with maintenance, as well as improving processes. Diagnostic analytics help answer questions such as:
    • Why does part failure in our rail fleet increase in the month of October?
    • Why did part breakdown increase when we switched to a new vendor?
    • Why did the price of parts increase over the last six months?

What Tools Can be Used for Diagnostic Analytics?

Most modern analytics tools contain a variety of search-based, or lightweight artificial intelligence capabilities. These features allow for detailed insights a layer deeper (for example: the Key Drivers visualization in Power BI, or Qlik’s search-based insight functionality).

However, while these tools provide an effective lightweight means to address diagnostic analytics use cases, they are not a means to a full-scale diagnostic analytics implementation. The feature functionality is often limited to a select set of dimensions and measures to apply against, and its detail is kept relatively surface level. If you’re looking to go a layer deeper in your analysis, it won’t be possible here.

Other tools that can facilitate diagnostic analytics are often primarily geared toward data science and machine learning. Lityx and DataRobot both have feature functionality in which users can discern influencing factors or variables that caused lift within their data. However, those insights are not the primary focus, as both vendors are geared toward a robust focus on machine learning and predictive modeling. Vendors like Sisu, however, have built their core business around addressing diagnostic analytics use cases.

What is Sisu and How Does Their Diagnostic Analytics Tool Work?

Sisu is a decision intelligence engine used by data teams and business leaders to diagnose changes in metrics to enable better, faster, and more data-driven decisions. The company’s proprietary engine uses scalable machine learning to rapidly test every possible combination of factors in your data to uncover the subpopulations having the biggest impact on your metrics. Sisu’s machine learning algorithms can perform analysis across millions of dimensions of data in seconds, helping comprehensively answer diagnostic questions at a rate far faster than is possible using traditional business intelligence tools.

Because Sisu’s diagnostic analytics platform is designed to make the most pertinent data easier to find, data teams can use it to augment their manual hypothesis testing by revealing where data specialists should focus their time and resources to maximize outcomes.

A diagnostic analytics workflow within Sisu reflects how businesses prefer to answer questions. It is iterative, with many questions asked in rapid succession and with each question building on the one prior. Key steps in the workflow include:

  • Explore a Metric Change: Sisu’s metrics-first interface improves the way you can investigate changing KPIs and empower your entire team to do the same. You can quickly discover data trends and metric changes on a chart or dashboard. Explore the metric and underlying data using the “drag and drop” interface, as well as pivot, filter, and dig into the data.
  • Reveal Why Change Occurred: To reveal hidden drivers behind the change, Sisu’s machine learning algorithms automatically analyze complex cloud-scale data in seconds to identify key drivers of metric performance. This augments manual analyst-directed analysis.
  • Take Action, Quickly: The analytics platform delivers clear and understandable analysis of key drivers using tools such as interactive waterfall charts. Businesses can use these insights to take immediate action and drive business impact. Decisions based on comprehensive diagnostic analysis are faster, better, and more trustworthy.
  • Share Insights and Outcomes: Insights need to be communicated to have maximum impact. Sisu allows teams to collaborate and iterate in real time, sharing metrics, diagnostic outcomes, dashboards, and presentation-ready visualizations.
  • Iterate and Keep Asking: Sisu’s rapid machine learning-driven analysis helps teams iterate in real time, answering follow-up questions as they arise and with minimal delays. This leads to greater use of data in decision making and better decision outcomes.

It’s never been easier for a business to understand the why behind what happened when analyzing data. If you need something high level, native features in your existing data viz solution may work. If you are looking for heavy duty advanced analytics using tools such as Lityx and DataRobot is the way to go. If you are looking for augmented analytics, however, Sisu is a tool our clients have had success with.

Kevin Lobo Kevin is our VP of Consulting and is based out of our Chicago office. He leads the entirety of our consulting organization, including 100+ consultants in the U.S. and Europe. Outside of work, Kevin enjoys spending time with his wife and two daughters, going to concerts, and running the occasional half-marathon.
Subscribe to

The Insider

Sign up to receive our monthly newsletter, and get the latest insights, tips, and advice.

Thank You!