Noreja Process Intelligence

Connect, transform, and analyze your process data with Noreja. Our containerized and micro-service based approach provides a flexible and scalable solution for cutting-edge process mining and analysis  

Different Noreja tool components
Different Noreja tool components

Causal Process Template

Source Layer How is data integration and process definition handled?

On the one hand, business processes cover a lot of domain-knowledge which is hidden inside the minds of process owners, analysts, and business experts. On the other hand, IT-systems generate huge amounts of data alongside the execution of day-to-day operations.

Consequently, we believe that the basis of effective process mining is a combination of process data and domain-knowledge. Our approach merges both knowledge sources by allowing the user to define and sort the processing order of certain business objects (e.g. order, picking slip, invoice, etc.) based on the underlying database structures. 

In the next steps, this defined order of causal relationships serves as the basis for importing the data from the source system into the graph database.

1. Causal Process Template

Source Layer How is data integration and process definition handled?

On the one hand, business processes cover a lot of domain-knowledge which is hidden inside the minds of process owners, analysts, and business experts. On the other hand, IT-systems generate huge amounts of data alongside the execution of day-to-day operations.

Consequently, we believe that the basis of effective process mining is a combination of process data and domain-knowledge. Our approach merges both knowledge sources by allowing the user to define and sort the processing order of certain business objects (e.g. order, picking slip, invoice, etc.) based on the underlying database structures. 

In the next steps, this defined order of causal relationships serves as the basis for importing the data from the source system into the graph database.

Causal Event Graph

Database Layer What is the underlying data structure?

Based on the data-aware process modeling, the underlying source data is transferred into a graph structure in order to model temporal interconnections between business objects. This structure provides an ideal foundation for causal process mining but also other advanced data science use cases. Thereby, graph databases …

  • Represent process-like data structures: Graph data structures resemble processes naturally because they do not just store data objects themselves but also their relationships – and a business process consists, more or less, out of activities (nodes) and relationships (edges).
  • Contextualize data: The network-like data structure between business objects (e.g. order, invoice, product items) allows uncovering interesting relationships more intuitively. This allows for more explorative data analysis.
  • Allow pushdown: Computational power is pushed down from the process mining application layer to the graph database engine.
  • Accelerate complex queries: Especially complex queries with many interrelations can be handled faster and more efficiently by graph database engines since relational database engines need to perform cost-intensive joins first.

Below we depicted one simple process instance represented as a graph structure:

2. Causal Event Graph

Database Layer What is the underlying data structure?

Based on the data-aware process modeling, the underlying source data is transferred into a graph structure in order to model temporal interconnections between business objects. This structure provides an ideal foundation for causal process mining but also other advanced data science use cases. Thereby, graph databases …

  • Represent process-like data structures: Graph data structures resemble processes naturally because they do not just store data objects themselves but also their relationships – and a business process consists, more or less, out of activities (nodes) and relationships (edges).
  • Contextualize data: The network-like data structure between business objects (e.g. order, invoice, product items) allows uncovering interesting relationships more intuitively. This allows for more explorative data analysis.
  • Allow pushdown: Computational power is pushed down from the process mining application layer to the graph database engine.
  • Accelerate complex queries: Especially complex queries with many interrelations can be handled faster and more efficiently by graph database engines since relational database engines need to perform cost-intensive joins first.

Below we depicted one simple process instance represented as a graph structure:

Causal Process Mining

Application Layer How is causal process mining working and what kind of analyses and visualizations can be used? 

Based on the graph database structure, it gets possible to plot, filter, and sort process instances and select areas of deviating cases for further investigation. The workflow-like tool will guide the user from general to specific visualizations. In the end, the user gets target-oriented implications for process optimization.

Below we depicted a selection of visualization types: 

3. Causal Process Mining

Application Layer How is causal process mining working and what kind of analyses and visualizations can be used? 

Based on the graph database structure, it gets possible to plot, filter, and sort process instances and select areas of deviating cases for further investigation. The workflow-like tool will guide the user from general to specific visualizations. In the end, the user gets target-oriented implications for process optimization.

Below we depicted a selection of visualization types: 

Process Performance Indicators (PPIs)

Application Layer – Which key figures can I use to continuously measure my process? 

Our tool allows calculating all kinds of PPIs that are built to support higher management or a process owner in order to monitor the process execution over time. PPIs give implication on whether a certain process optimization initiative has led to positive business outcomes. Below we highlighted a selection of important PPIs from the Order-to-Cash process*.

Avg. Throughput Time

Calculation of the average time it takes for a certain process from its first to its last activity. By using filter functions, it is possible to limit the calculation to just a customer, an order type, a single process instance, a region, etc.

Rework Rate

Uncover how many rework actions take place along a business process. This allows not only uncovering the additional costs necessary for manual work but also the intensity a customer or supplier changes its requests. Of course, there is the possibility to filter by customer, order type, resource, region, etc.

Activity Batching Factor

The batching factor calculates how intensive certain process activities bundle multiple business objects along the process. For instance, a delivery activity can encapsulate multiple orders for the same customer. From a cost perspective, it can be helpful to combine the delivery of multiple orders for the same customer in order to save money for transportation.

Number of Process Variants

The number of process variants is an indication of process complexity. We count one process variant as one individual process path taken. In reality, we often experience hundreds or thousands of individual process paths. Usually, the goal is to reduce the number of process variants in order to get closer to the standard process paths.

Number of Process Violations

A process violation occurs if one process activity is executed in a way that contradicts its initial definition. For instance, an process activity was executed in the wrong order, a four-eye-check was skipped or certain thresholds were exceeded.

Days Sales Outstanding (DSO)

Days sales outstanding indicates the average amount of time it takes between the invoice date and the actual payment of a customer. The number is highly important for assessing the liquidity and cash flow based on outstanding account receivables. The number helps optimize cash discount rates in order to nudge the customer towards an earlier payment. By using Noreja, it is possible to filter by certain business objects like customers, orders, order types, etc.

Cash Discount Received

The cash discount received is a measure of how intensive a customer uses a cash discount in exchange for early payments. A high number indicates that the cash discount is attractive for the customer and that the customer has good payment morale. A low cash discount received leaves room for maneuver during upcoming payment terms negotiations.

Hidden Credit Time

The hidden credit time is the average number of days between delivery and invoicing. It indicates how much time a seller grants credit for its customers, hence, is not paid for its services or products. By reducing the hidden credit line, the cash flow is optimized.

* Depending on the source database and the availability of relevant data

Process Performance Indicators (PPIs)

Application Layer – Which key figures can I use to continuously measure my process? 

Our tool allows calculating all kinds of PPIs that are built to support higher management or a process owner in order to monitor the process execution over time. PPIs give implication on whether a certain process optimization initiative has led to positive business outcomes. Below we highlighted a selection of important PPIs from the Order-to-Cash process*.

Avg. Throughput Time

Calculation of the average time it takes for a certain process from its first to its last activity. By using filter functions, it is possible to limit the calculation to just a customer, an order type, a single process instance, a region, etc.

Rework Rate

Uncover how many rework actions take place along a business process. This allows not only uncovering the additional costs necessary for manual work but also the intensity a customer or supplier changes its requests. Of course, there is the possibility to filter by customer, order type, resource, region, etc.

Activity Batching Factor

The batching factor calculates how intensive certain process activities bundle multiple business objects along the process. For instance, a delivery activity can encapsulate multiple orders for the same customer. From a cost perspective, it can be helpful to combine the delivery of multiple orders for the same customer in order to save money for transportation.

Number of Process Variants

The number of process variants is an indication of process complexity. We count one process variant as one individual process path taken. In reality, we often experience hundreds or thousands of individual process paths. Usually, the goal is to reduce the number of process variants in order to get closer to the standard process paths.

Number of Process Violations

A process violation occurs if one process activity is executed in a way that contradicts its initial definition. For instance, an process activity was executed in the wrong order, a four-eye-check was skipped or certain thresholds were exceeded.

Days Sales Outstanding (DSO)

Days sales outstanding indicates the average amount of time it takes between the invoice date and the actual payment of a customer. The number is highly important for assessing the liquidity and cash flow based on outstanding account receivables. The number helps optimize cash discount rates in order to nudge the customer towards an earlier payment. By using Noreja, it is possible to filter by certain business objects like customers, orders, order types, etc.

Cash Discount Received

The cash discount received is a measure of how intensive a customer uses a cash discount in exchange for early payments. A high number indicates that the cash discount is attractive for the customer and that the customer has good payment morale. A low cash discount received leaves room for maneuver during upcoming payment terms negotiations.

Hidden Credit Time

The hidden credit time is the average number of days between delivery and invoicing. It indicates how much time a seller grants credit for its customers, hence, is not paid for its services or products. By reducing the hidden credit line, the cash flow is optimized.

* Depending on the source database and the availability of relevant data

“A company’s ability to flexibly change its organizational processes indicates its readiness to undergo other radical reconfigurations.”

– Kim, Shin, Kim, and Lee (2011: 488)

Advanced Data Science

Application Layer – Which analytics capabilities are possible in addition to causal process mining? 

Our modular approach allows us to plug in additional micro-services for advanced data science to the graph data structures. This offers users different types of analysis, including correlation, clustering, or discriminant analysis on process data. Moreover, graph database structures are capable to handle machine learning or big data use cases far more efficiently than relational databases. Currently, we are working on developing the first advanced data science service in the area of clustering.

Stay tuned!

4. Advanced Data Science

Application Layer – Which analytics capabilities are possible in addition to causal process mining? 

Our modular approach allows us to plug in additional micro-services for advanced data science to the graph data structures. This offers users different types of analysis, including correlation, clustering, or discriminant analysis on process data. Moreover, graph database structures are capable to handle machine learning or big data use cases far more efficiently than relational databases. Currently, we are working on developing the first advanced data science service in the area of clustering.

Stay tuned!