This post is also available in:
Deutsch
In today’s data-driven world, businesses must leverage analytics to make informed decisions and achieve their goals. At our consulting firm, we specialize in IT projects, with a focus on data, business intelligence, and IT infrastructure. Recently, we had the opportunity to work with an electric mobility company in Europe, helping them understand whether their goals for the year were reasonable through the power of analytics.
The Challenge: Too Many KPIs
The electric mobility company faced a significant challenge: they had an overwhelming number of Key Performance Indicators (KPIs) that varied across teams. Different teams had different definitions and understandings of these KPIs, making it difficult to align their efforts and measure success accurately. Our task was to gather and understand all KPIs from both a business and data perspective, and streamline them to create a cohesive and effective measurement system.
Streamlining KPIs for Better Insights
We began by collecting and analyzing all the existing KPIs, identifying their sources and definitions. This process involved combining data from various sources and establishing a proper core data source to ensure accurate data definitions and business understanding. Once we had a clear picture, we categorized the KPIs into primary and secondary KPIs. Primary KPIs were the most critical metrics, while secondary KPIs could be derived from the primary ones. By focusing on improving the primary KPIs, we could ensure a more significant impact on the company’s overall performance.
Through this process, we reduced the number of KPIs from around 30 to just 5 primary KPIs. Among these, we identified one KPI that truly mapped the business’s performance: the “number of usage sessions.” This KPI indicated the volume of usage of the electric mobility product and provided insights into the revenue vs. costs ratio.
Understanding the “Number of Usage Sessions”
The “number of usage sessions” refers to the total count of times consumers use the electric mobility product, such as charging sessions for electric vehicles. This metric is crucial as it reflects the product’s adoption and usage frequency, directly impacting the company’s revenue and operational costs. By monitoring this KPI, the company can gauge its market penetration and customer engagement levels.
Factors Influencing the KPI
The “number of usage sessions” was influenced by both intrinsic factors, such as customer preferences, and extrinsic factors, such as the price of electricity, the cost of electric vehicles, and external events like the COVID-19 pandemic. Other extrinsic factors that could affect the sessions include government policies and incentives for electric vehicle adoption, the availability of charging infrastructure, and seasonal variations in usage patterns. It was crucial to consider these factors when determining the future of the KPI. To forecast the KPI’s development over time, we built a forecasting model using historical data.
Building the Forecasting Model
We employed two models for our analysis: the ARIMA model and the LSTM model. The ARIMA model is well-suited for time-series data with clear trends and seasonality. Before running the ARIMA analysis, we ensured the data was stationary, meaning its statistical properties did not change over time. We also examined the mean and variance to confirm the data’s suitability for the model.
For the LSTM model, which is designed for more complex patterns and long-term dependencies, we prepared the data by one-hot encoding categorical variables and normalizing the data. This model allowed us to predict the KPI’s value based on multiple factors, providing a more comprehensive analysis.
Results and Insights
Using these methods, we analyzed the patterns of development over time for the “number of usage sessions” KPI. Our initial analysis focused on predicting the KPI for the next seven days, but we also extended the analysis to several months in advance, updating the model monthly to account for changing factors. This approach helped us determine if the company’s goals for the KPIs were achievable and identify the optimal targets.
Our analysis revealed that while increasing the “number of usage sessions” indefinitely was not feasible, understanding the factors influencing the KPI allowed us to set realistic and achievable goals. For instance, the number of people purchasing electric vehicles versus traditional vehicles, the frequency of charging sessions, and the introduction of new electric vehicle models to the market were critical considerations.
What Did We Do for This Business
- Business Unification: We unified the business by aligning all teams with a common understanding and definition of KPIs, ensuring everyone was working towards the same goals.
- Data Unification: We combined data from various sources to create a single, accurate core data source, providing a clear and consistent view of the business metrics.
- Roadmap for the Company: We developed a roadmap for the company to follow, outlining the steps needed to achieve their goals and maintain focus on the primary KPIs.
- Methods to Avoid Straying onto Other KPIs: We established methods to ensure the company remained focused on the primary KPIs, preventing distractions from less critical metrics.
- Understanding the Impact of Various KPIs: We provided insights into how different KPIs impacted the business, helping the company prioritize their efforts and resources effectively.
- Dashboard Creation: We created a dashboard that was regularly updated, allowing the business to monitor their KPIs and make informed decisions based on real-time data.
Skills and Expertise
Developing such models requires a combination of skills, including data analysis, statistical modeling, and machine learning. If your company faces similar challenges and needs assistance in leveraging analytics to drive success, our team of experts is here to help. Contact us to learn more about how we can support your business in achieving its goals.
This post is also available in:
Deutsch
