SAP Analytics Cloud: Analytics Designer Hackathon – KPI Analysis

This blog post is a submission for the Analytics Designer Hackathon

More information about the Hackathon

Introduction

Any company evaluates its success based on Key Performance Indicators. Therefore, it is important to have a convenient tool for KPI analysis.

The developed KPI model is based on the structure of our client’s company. This is one of the largest energy companies in Ukraine. Its top-level managers need to track KPIs performance based on actual and plan data. Our objective is to help a company achieve strategic and tactical goals by quick response to problems.

The selected data structure represents KPIs in detailization that is sufficient for top-level analysis. All data are relevant, but scrambled and anonymized. My data model provides for new KPIs import without the necessity of changing the data structure or the application. Such approach makes this solution sustainable for long-term usage.

The purpose of the application is to provide actual information about the KPIs performance of plans and identify problem areas (bottlenecks). For each KPI we can track the dynamics over time and conduct a flexible structural analysis. One of the key features is the ability of modeling a forecast.

Application description

On the main page we can see an overview of all KPIs. At a glance, we can identify unperformed indicators that are highlighted in red. KPI is considered as unperformed if its plan variance exceeds its permissible variance.

While opening the application, a filter by date is automatically set. However, the user may change the period. There are additional features like Search to Insight and Help Page in the upper right corner.

Also I created user-friendly ad-hoc attractive navigation panels for strategies and departments. Such panels make it possible to see the whole picture in any direction and filter KPIs as we wish. Using JavaScript I was able to implement the flexible cross filtering logic. It was one of the most difficult components. It is worth noting that the KPI chart is always the right size. The point is that its height varies depending on the number of the selected KPIs. Bless JS!

We can open a popup with detailed analysis by clicking on any KPI.

Then user can switch from detailed analysis to structural analysis by clicking on the table icon. A suitable dimension is automatically added to the rows of the table depending on recommendations. User can easily change the table structure.

By the way, I’m a huge fan of Value Driver Tree. Therefore, I created it in Analytics Designer!

Based on the current year actual data, an automatic forecast for the remaining months is built. If the high-level KPI (EBITDA, EBIT or Net Profit) is not performed, then simply adjust the forecast value with the sliders. Such a “what-if” simulation determines what to do in order to perform the plan.

Conclusion

The usage of JavaScript in Analytics Designer allowed me to set up flexible navigation through the application and implement other specific customer requirements. I tried to develop the intuitive user interface and make the visualization easy to interpret. As a result, I’ve got the application that supports quick decision making.

This application is ready to be used by the client. It can be implemented for any company with minimal modifications. I see the further application extending with planning capabilities and advanced modeling.