Glossary

Here you can find all terminology with a short explanation. In the Navigation on the left you can jump to the specific letter you are searching for.

A

A-Activities

The number of distinct activities that are included in all the A-Variants.

A-Variant

A-variants are the most common process variants that make up 80 percent of all cases. In the process analyzer, use the + symbol to increase the number of variants until you reach 80% of all cases and that number will match this PPI.

All the activities that make up all the A-Variants in our process are called A-Activities.

Activity

An activity is a specific step in an overall process.

Example: An IT ticket process could have the activities "Insert ticket", "Assign seriousness" , "Take in charge ticket", "Wait", "Resolve Ticket" and "Closed".

💡 Note: Activities are general steps in a process and have no timestamp.

For a purchase order, activities could be registering the order, checking stock, handling payment, shipping the order, or cancelling the order. When considering a specific case as it goes through the process, it will then get timestamps for each step and they are then called events.

Activity Type (ID)

• The Activity Type field holds the name of an activity.
• It is used in the app as a dimension.
• The ActivityTypeID is a number to represent this activity.

ActivityUserName

This fields holds the user who is responsible for completing an activity. It can also be a resource, such as a computer, if the activity is automated.

Aggregation

In data analytics, aggregation refers to the process of combining multiple data points to produce a condensed data set for analysis or processing. The values of multiple rows are processed together to form a single summary value.

The key goal of aggregation is to provide a more manageable and interpretable representation of the data, which is useful for various analytical tasks.

Use different types of aggregation to group items, e.g. Sum, AVG, Max, Min, Count, Median.

AL_ActivityTypes Table

This table includes all activities from the process log but it is separated so that the filters don’t apply on this table.

It is used as dropdown/list that should always show all activities (no filtering).

Automation Rate

The automation rate indicates the percent of processes that are fully automated.

💡 Note: The higher the automation rate, the more efficient your process runs.

AutoML

Qlik AutoML is a powerful tool that makes it easy for analytics teams to generate models, make predictions and test business scenarios without having to write code.

B

B-Activities

The number of distinct activities that show up in the B-variants.

B-Variants

While A-variants make up the first 80% of all the cases, the B-variants are those variants that cover the next 15% of cases (or from 80-95% of all cases in the process analyzer).

Benchmark

When you benchmark something, you compare it to a standard.

In process mining, benchmarking can help identify gaps in performance and opportunities for improvement. You can compare your existing process to various options, e.g.:

• time: last year vs this year,
• region: North America vs Europe,
• Supplier A vs Supplier B, etc.

By identifying which areas do better or worse you can identify best practices or areas of improvement.

Bookmarks

You can use bookmarks to save certain selection statuses. You then have the option of checking them later or sharing them with other users.

To learn more on bookmarks have a look at "How to create and use bookmarks".

Bottleneck

A bottleneck is a specific point in the process where an increase in delay before an activity occurs.

A bottleneck can be identified by tracking process events and their timestamps. Bottleneck analysis is used to identify constraints and inefficiencies in a process that limit its capacity and productivity. It involves identifying and prioritizing the steps that take the most time, resources, or effort and analyzing them to identify the root causes of the inefficiencies.

BPMN

Business Process Model and Notation, or BPMN, provides an easily understandable visual for all users.

BPMN Models are made from a few basic elements, including flow objects, connecting objects, pools, swimlanes, and artifacts.

BPMN can be used to design initial sketches of a process, as well as executable processes that can be used as a template for implementing software, where it would take an interface function between the design of a process and its implementation.

BPMN Import Wizard

It converts the information in a BPMN Model into a process in the mpmX app.

C

C-Activities

The remaining distinct activities that only show up in the C-Variants.

C-Variants

C-Variants cover the last 5% of all cases and thus are the least common process variants.
(A-variants cover the first 80% and B-variants cover the next 15% of variants.)

Case

A Case is a specific instance of going through a process, following a single process path.

In an IT Ticket process, each new ticket that is processed is a case. Each case is a row in a data table. A case can have specific attributes, e.g. the ticket type, the customer, the product, the workgroup, and so on.

Case (Context) Information

Case information is specific to a case, can be identified by the CaseID, and does not change throughout a specific case.

It can be the Purchase Order Number in a sales process, or a Ticket Number in an IT HelpDesk process.

Case Counter

The CaseCounter field efficiently counts the total number of cases. It is used in the app KPI for showing the total cases.

Case ID

A Case ID is a unique marker that will identify the specific process instance that you are tracking. It can be used to track the different events that occur in a specific case.

concatenate

In data analytics, to concatenate is the joining of two or more elements. It can be applied in expressions and the data load script to load tables, or in expressions by using the ampersand string operator to give meaning to a value in labels, titles, etc.

Conformance Checking

Conformance checking monitors how closely the real world cases (and event paths) match the recommended process model. It compares existing process models with event logs of these processes by calculating deviation metrics.

Calculating how well the reality of the process conforms with a set process model provides valuable insights into process deviations.

Typical problems include

• skipping a step,
• sadding a step,
• repeating a step, or
• performing steps in the wrong order.

A core element of Conformance Checking is the concept of happy paths. In combination with specific Conformance Metrics these happy paths provide a comparison method that is easy to understand.

Context Information

We differentiate between two kinds of information:

• Case Context Information: Information that is related to the CaseID. It is unique and doesn’t change throughout a process instance.
• Event Context Information: Information that is related to an Event. It is possible that it will change within a case.

Convergence

When one event is replicated across different cases, possibly leading to unintentional duplication. This replication of events can cause misleading diagnostics.

D

Data Extraction

The process of retrieving raw data from various sources to capture all the activities within a business process.

• Data can be extracted from databases, ERP systems (like SAP, Oracle), CRM systems, logs, or other business applications.

Data Mining

Data mining is an interdisciplinary approach that uses methods from computer science and statistics. The aim is discovering new cross-connections and trends.

Data Model

The mpmX Data Model always consists of the mpmX core tables with the enhanced process log, a process variant table and some help tables. Additionally it can contain conformance checking tables, case/event context with extra information and/or the root cause analysis table.

For more information on this see our Help Documentation Page.

Data Set

Is a sample or table which consists of instances (individuals, entities, cases, objects, or records)

Deviation

A variant that does not follow a desirable path.

Dimension

A Dimesion is a data field whose values contain descriptive information for each record.

• It can be either a word or a number.
• Usually it repeats in a table and so can be used to easily group items.

Numeric examples: Year, Product ID, Price, Weight & Size dimensions, etc.
Alphabetic example: Department, Document name, Color, Status, etc.

Dimensions can be used in visualizations to sort and display specific data.

Divergence

When there may be multiple instances of the same activity within a single case. Events on a lower entity level may become indistinguishable and seemingly related.

Example: The pick item and pack item events happen only once per item and in a fixed sequence (e.g., you pick the item, then pack it), but within a flattened sales order event log these events may appear to happen many times and in a seemingly random order for an individual case.

E

Edge

An edge is the arrow or line connecting two activity nodes in the process analyzer. It indicates the flow of work through the steps in the process (or activity transition).

• The thickness of the line indicates the volume of cases that follow that path.
• It is also possible to indicate the average waiting time between two activities, if configured this way.

Edge Details

The number showing on the arrow (or edge) connecting two activities (in a process) or events (in a case) will show how often the first activity was followed by the second.

Clicking on an edge will open a popup box with more edge information and allow you to filter by select cases that only contain this edge or that do not have this edge.

EL_EdgesTypes Table

This table will show all the different edges that connect two distinct activities in a process.

Each edge has an EL_EdgeID, made of two ActivityID numbers joined by a hyphen (e.g. 103-109) and an EL_EdgeType, made of the names of the two Activities (e.g. Resolve Ticket - Closed).

ETL - Extract, Tranform, Load

There are two different ways to use the mpmX Template App. It is possible to read in a self-made event log or to process the process model created by the mpmX EventLogGeneration Apps.

Data Load Without ETL

• The log has previously been created.
• The log contains at least one timestamp, a case ID and an activity description.

Data Load with ETL

• The log has been created with the mpmX ETL process by the mpmX EventLogGeneration Apps.
• The ETL output tables pa_activity_log, pa_process_variants, CaseTimes, EventTimes, ProcessPathContinuation, and so on are accessible.

Event

An event is a process activity related to a specific case and with a distinct timestamp.

A process may have an activity (e.g. “Resolve Ticket”) that occurs just once, but each time a new case goes through the process, that activity will occur and each time have a different timestamp, so it forms three distinct events.

An individual event can have attributes that describe the context of its occurrence, e.g. start and end timestamps, a specific event ID, a resource (user or machine that executed the activity), the event costs, etc.

Event Context Information

Event Information is specific to an individual event in a case and could change at each process step.

• A timestamp is a good example of event information.
• Also User often changes as different people carry out different steps in the process.

Event Log

An event log is a table that contains event information in all cases that have gone through a specific process.

Event logs must contain at least three fields that are necessary to run the algorithms:

• CaseID as a numeric identifier,
• Activity name to specify what activity took place, and a
• Timestamp to give a precise time.

Additionally, the event log can contain context information, such as the company, areas, machines, users. In a business environment, event logs are made of the data collected in various system databases, for example CRM, ERP, MES, PLM, or SRM.

Expression

In data analytics, it is a calculation which produces a value or dataset, and can include components such as numeric values, text values, data field names, operators + - * / , and functions like Sum () and Avg ().

Extensions

As mpmX is a Qlik-based product, thus many of the capabilities are based on Qlik software. The Extensions are the coded features that are especially developed for the mpmX product.

G

Geofield

A geofield is a data field that contains geographic information. This can be country, region, city, co-ordinates, addresses, or other spatial data.

• Geofields can be used to show information on a map of the globe or to display locations of machines in a warehouse.
• They can be used to visualize distribution, to optimize production layouts or delivery routes, or to understand how location impacts process performance, such as identifying regions with higher sales, delays, or costs.

H

Happy Path

Each process has one or more acceptable and recommended paths that are expected to be followed. The path is usually the most efficient one. Ideally it should also be the most common.

I

Idle Time

Idle time is a period of time in which an asset (machine or an employee) is ready and available, but is not doing anything productive.

💡 Note: Idle time is the opposite of Lead Time which is how long a case has to wait after one event before being able to start the next process activity.

Import Wizard

The mpmX Import Wizard is a tool that makes it easy for users to create a new app (with already prepared data logs) by walking you through the necessary steps and configurations.

You can reach it by clicking here!

K

KPI

Key Performance Indicators are metrics used to measure a company's success against a set of targets or by observing their change over time. A Key Performance Indicator is a calculation based on one or more aggregations. For example, the sum of sales is a single aggregation, while the sum of sales divided by the number of customers is a measure based on two aggregations.

Examples for KPIs are:
TIME | response time, processing time, cycle time, control time, idle time, setup time, transfer time, value added time

COST | error costs, total costs, resource consumption, value added costs

QUALITY | output quality, process variants

CAPACITY | bottlenecks, resource capacity, employee capacity, output, resource capacity

INTEGRATION & COMPLEXITY | Automatization, information flow, timeliness (of information), degree of standardization

L

Lead Time

The amount of time it takes to get from one activity (or node) to another is called Lead Time.

💡 Note: A long lead time indicates a inefficient process where not enough resources are assigned to complete the tasks as they arrive.

Lead Time Spread

Shows the volatility of the lead time among cases by calculating the “inter-quartal range” of the lead time, also known as the midspread or middle 50%. It covers the data between the 25th and 75th percentiles. In other words, it shows the time difference between the slower and faster cases while excluding the extreme cases. A low spread score is better, when you want to rely on assumptions of lead time.

M

MasterItem ID

When you create and build your visualizations, you can save assets to reuse them in other visualizations and on other sheets.

You can save visualizations, dimensions and measures, as master items in the assets panel. When your app gets published these master items will be available to others as ready-to-use visualizations, dimensions and measures. The master items are connected between the front end and back end via the _MasterItemID.

Maverick Buying

Maverick buying refers to the practice of employees or departments making purchases outside of the established procurement policies and procedures of an organization. This typically involves bypassing approved vendors, neglecting formal approval processes, not involving the Finance Department, or not adhering to negotiated contracts.

Maverick buying poses significant risks to organizations, including financial inefficiencies, increased costs, and compliance issues. One way it is usually detected in a Purchase To Pay process is when an invoice is posted before a purchase order was created.

Measure

A numeric data field whose values are used in mathematical calculations – so Salary, Items Sold, or Time (Duration) – might be good examples, because they can be calculated.

In contrast, an ID Number or telephone number could not be used in calculations, so they would not be Measures.

Machine Learning and AI

With a machine learning model that is integrated into mpmX, predictions for open processes can be calculated.

The mpmX platform offers numerous options for making predictions:

• Connection to Python

• Connection to existing ML platforms

• OpenAI Connector

N

No Touch Case

It identifies whether the case has been completed entirely by automated users. (No real user has worked on or touched the case).

Node

A node is a circle on our process analyser that represents a specific activity in a process, or an event if you are looking at a specific case.

• The node icon is an inner circle and outer ring containing various KPI information.
• The inner circle changes colour from light to dark as its measure increases.
• The outer ring displays a colour grading scale that fills in as the measure increases.
• Clicking on a node opens the context menu popup.

O

OCPM

Object Centric Process Mining (OCPM) (in contrast to traditional, Case Centric Process Mining) is an approach to process mining that is relatively new. It allows for the analysis of processes that involve multiple interacting objects. This approach provides a more holistic and accurate representation of complex processes, enabling better insights into interactions, dependencies, and overall process behavior.

Open Case

An open case is an instance going through a process which has not yet come to its natural end.

Optimization Threshold

An optimization threshold is a predefined limit or criterion used to determine when a process has reached a level of efficiency that is considered acceptable or optimal.

It sets the benchmark for achieving desired outcomes, beyond which further improvements may yield diminishing returns or require additional resources. It is specifically needed for the Root Cause Analysis to optimize too fast or too slow cases or the automation rate.

P

pa_activity_log Table

The “pa” here stands for process analyzer. The table contains five types of information:

1. event information,
2. edge information,
3. process variant information,
4. lead time information, and
5. information about the events that occur before and after each instance.

PPI (Process Performance Indicator)

PPIs provide insight into how well a process is performing and whether it is achieving its intended objectives. They are used to measure various aspects of a process, such as speed, quality, cost, and compliance.

Predictive Analytics

Predictive analytics involves using statistical techniques, machine learning algorithms, and historical data to forecast future events or trends. By analyzing patterns and relationships within the data, predictive analytics can provide actionable insights and help organizations make informed decisions by anticipating potential outcomes and risks.

Process Activity Filter

With the ProcessActivityFilter, you can choose activities that should or should not be part of the process variants that you would like to see.

Process Analyzer

It is a tool for visualizing your process and identifying places to start your additional analysis.

You will see your process laid out from Start to End with all the steps or activities showing as nodes. The volume of cases that follow the various pathways between the nodes is indicated by the thickness of the edge lines.

Various KPIs can be seen related to each node and edge. You can filter cases based on what you select in the process analyzer.

Process Discovery

Creating an end-to-end visualization from the event log by following every step any case has taken.

• Cases: How many cases are going through this particular process.
• The more cases that can be processed, the more effective the process is.
• The timeline underneaths shows the development over time.

Process Homogeneity

This is the number of process variants compared to the number of cases in the variant. The more cases covered by a single variant, the higher the homogeneity percentage.

Process Model

A process model is a visual representation of a business process, illustrating the sequence of activities, decision points, and interactions involved. It helps in understanding, analyzing, and improving processes by mapping out how work flows through an organization.

Process Modeler

With the ProcessModeler, target processes can be defined so that deviations from them can be investigated with the help of the ProcessAnalyzer.

Process Pattern Search

It is a tool to build any sequences of events, using specific activity names and wildcard placeholders, which allows you to filter and narrow down your analysis to specific process patterns.

Process Variant Inspector

A tool to look at the patterns in your process. You can highlight activities and compare different process variants for different angles of analysis.

Process Variants

A process variant is a unique sequence of activities from start to finish in a process.

While a process may have its ideal model, it may also have many other variants each with a unique combination of activities.
Variants may be unique because they:

• have extra steps,
• skip steps,
• repeat steps, or
• follow steps in a different order.

Q

Qlik Sense

Qlik Sense is a data analytics and business intelligence platform that enables users to visualize and explore data, generate insights, and make data-driven decisions through powerful data integration capabilities. It is designed for self-service data visualization and exploration. The intuitive interface makes it suitable for a broad range of users.

Qlik View

QlikView is a business intelligence tool that allows users to develop interactive dashboards and perform guided analytics by creating custom data visualizations and reports. It is a more traditional, developer-focused tool for creating guided analytics applications and dashboards, offering greater control over layout and scripting.

R

Resource Log

In the resource log, the process steps are not made up of activities, but rather the resource that completed the activity. You are able to view the process from the user or machine that completed the steps, based on resource or department in a company.

Rework

Rework is an indication that the process did not go 100% correctly the first time, so an inefficiency. Rework events are usually related to changes (change order, change price, etc) or cancellation.

Robotic Process Automation (RPA)

RPA (Robotic Process Automation) automates repetitive, rule-based tasks using software robots. These robots interact with software through the graphical user interface (GUI), performing tasks without needing process changes or specialized interfaces. They effectively take over user roles and interact with other systems.

Root Cause Analysis

The Root Cause Analysis is a method used for identifying underlying reasons for issues within a process. By adressing the root causes rather than just treating the symptoms, effective and lasting solutions can be found.

The RootCause table shows two probabilities for each process optimization potential:

• how likely each different attribute is to have caused the optimization potential
• or conversely, the chance of an attribute to occurring when an optimization potential has been detected.

S

Set Expression

In data analytics, it is an expression which uses an aggregation with a defined set of field values.

Qlik Sense uses the identifiers

• 1 to use all data and ignore all selections, e.g. Sum({1}Sales),
• $ to use all selections and ignore the rest of data, e.g. Sum({$}Sales),
• and modifiers are nested to specify selections, such as Sum({1<Year={‘2023’}>}Sales).

We can configure set expressions with multiple identifiers and modifiers embedded with additional fields, using operators to determine how the data sets are combined.

Spaghetti Diagram

Refers to a zoomed out visualization of all variants from start to end which looks like a plate of spaghetti. 🍝

T

Template App

The Template App contains over a dozen different Dashboards and Toolsheets with valuable information. By default, the following sheets are available in the mpmX Template App (if configured):

• Activity Details for "x": Focus on a single activity or resource and understand it’s  impact on your process.

• Ad Hoc Lead Time Analysis: Use this sheet to easily define custom subprocesses and reveal their lead times.

• Automation: See the automation rate of your process and reveal its automation potential in order to improve it.

• Benchmarking: Benchmark two different processes for a direct side-by-side comparison.

• BPMN Modeler: On this sheet, you can create, edit, and export a BPMN model from the ProcessAnalyzer. Also, an imported BPMN can be shown here.

• Conformance: Analyze the process conformance by seeing how it aligns with your defined process model.

• Edge details for "x": Analyze in detail a specific edge of your process and gain new insights such as ways to avoid unwanted edges.

• MasterItem Import: Import MasterItems defined in the backend and by the mpmX algorithms.

• Monitoring: Monitor your unfinished cases and take corresponding actions.

• Optimization Dashboard: Quickly detect the most relevant optimization potentials and navigate to the corresponding sheet to analyze the topic further.

• Process Discovery: Dive into your process by filtering and narrowing down your analysis to discover additional insights.

• Rework: Analyze rework events in your process such as unwanted or repeated process steps, and ways to improve your approach.

• Root Cause Analysis: Find the root causes of optimization potentials in order to optimize the process.

• StartToEnd Lead Time: Have a look at your processes lead time and identify cases that negatively impact the lead times. You can also analyze lead times of specified sub-processes.

• User Analysis: See the flow between users that work on one process instance to identify wasted or overworked resources.

• Variants: Observe the unique variants your process takes. Understand what drives your processes complexity and find out how to improve it.

Timestamp

A timestamp is the time that is assigned to an event in computer records as the event occurs. It is saved in the event log.

Timestamps are used for various synchronization purposes, such as to assign a sequence order for events in a process, to determine how long each step takes, or to record time in relation to a starting point. Data management relies on timestamps to ensure the integrity and quality of data.

Timestamp formats differ in regions, programming languages, etc. and so the format being used needs to be set using y (year), d (day), h (hour), m (minute), and s (second). One example of a timestamp format is DD.MM.YYYY hh:mm:ss.

Traditional Process Mining

A particular path that a case has taken from start to end. Because there are often many ways a case can go from start to end, there are usually many variants of a process.

V

Variable

• Also often referred to as attribute, feature, or data element.
• In data analytics, a variable is a container storing a static value or calculation. It needs a name and a definition (e.g. field names & values, aggregations, functions, expression, color codes, numeric values, or calculation symbols).
• In an event table, a variable is the data found in the columns.
• A variable can be set up once and applied multiple times to visualizations in that app, like in a dimension, measure, when configuring reference lines, setting limitations, and establishing conditional thresholds.
• As a naming convention, variables begin with a “v”. For Example Mehrwerk’s variables begin with “mv”.

Examples: $(vField) $(vProfit) $(vColor)

Variant

A particular path that a case has taken from start to end. Because there are often many ways a case can go from start to end, there are usually many variants of a process.

Visualization

Data can be presented in a visual way to make it easier to understand. mpmX does this for example in the different sheets where you can see the data presented in different charts, or the process analyzer which shows the different steps in a process and their order.

The different charts are explained in our Dashbord&Sheets Section.

W

Wildcard

A wildcard represents a placeholder for one or more steps or elements within a process. It allows for flexible matching of different process variations by accommodating variations in the sequence or presence of activities.

There are two different types of Wildcards in the Process Modelers Activity List:

• Wildcard [*]: It can be none or any random activities.
• Wildcard [???]: It can be just one random activity. 1-2 of the Questionmarks can be filled in to search for specific tasks e.g. [??2]

1-9

5-Layer Model

The mpmX TaskApps (EventLogGeneration Mode) have a 5-Layer model. The layers are:
• 01qvdbuilder
• 02transform
• 03eventsource
• 04processlog
• 05datamodel