High-value performance management solutions are highly sought after by organizations in highly competitive global markets. Despite the potential of large returns, many organizations experience challenges to establish even basic functions required to drive complex solutions. To better understand advanced processes and sophisticated models, we have provided a general glossary with basic terminology.
AI – Artificial Intelligence: The use of computer systems to simulate, augment & advance tasks that can be performed with human intelligence – computerizing this processing enables tasks to be completed faster, with increased complexity, and producing actionable insights humans may overlook. Relevant EPM & xP&A use cases: Descriptive, Diagnostic, Predictive & Prescriptive Analytics, ML (Machine Learning), RPA (Robotic Process Automation).
Algorithm: Algorithms are used to analyze data trends that help identify the main influencers and relationships leading to a specific outcome. Examples of algorithm categories are classification, regression, time series, decision trees, clustering and segmentation, forecasting, outliers, and association.
BI – Business Intelligence: is a technology-driven process that consists of business analytics, data mining, data visualization, data tools, data infrastructure to help organizations make data-driven decisions. Working with BI requires familiarity with SAP’s programming and database.
Bottleneck Utilization & Constraints: Generally used as an indicator that measures how well resources are being used. An example is how available capacity is unable to meet pending production demands that could impede the flow of operations. Compare to a bottleneck in a physical bottle where there is only so much fluid that can be pushed through the neck at any time – causing a bottleneck (backup).
BPC: BPC is a single platform for financial and operational process planning, forecasting, and reporting. By eliminating the need for separate modules or proprietary applications, we can simplify the technology, data, and meta-data management, the cost of ownership, and the user experience. With Microsoft Office and Internet Explorer as the database interface, users have an easy-to-use way to access a single version of the truth—decreasing training time and improving adoption. BPC is compatible with on-premise and private cloud deployments.
Budgeting: Budgeting is the process of allocating financial or human resources to support business activities with the intent to deliver optimized business returns. Investing in streamlining and adding efficiencies to budgeting, planning, and forecasting processes can pay significant dividends. Budgets, typically done annually at a low level of detail for a fixed time horizon of future activity (for example a budget to plan for next year’s activity is completed at the end of this year), have been criticized as too rigid and prone to obsoletion because of this timeline and long lead time for development. Rolling forecasts done more frequently in response to changing market conditions are put forward as an alternative. Some process experts – Beyond Budgeting Round Table for example - promote the complete abandonment of budgets.
Business Plans: All successful projects producing useful outcomes from the business point of view require a business plan. Success must be measured objectively, requirements and assumptions are known, risks and contingencies are highlighted (i.e. usable data or available personnel skill sets), resources are aligned and assigned, constraints are known and handled, deployment requirements spelled out, and most of all costs and benefits are clear. From a data science perspective, the plan to achieve project goals meets intended outputs that meet business objectives.
Classification: Classifications, or groups, are used to answer the who and when-type data analysis questions. For example, with fraud-based cases, we need to understand what transactions are fraudulent and what types of customers may be susceptible.
Cloud: While there are different types of cloud-based solutions (private, public, hybrid), a cloud solution is typically defined as being deployed on hardware located outside the organization. As of current trends, nearly all EPM deployments are cloud-based. In some less precise references, “Cloud EPM Solutions” refer to vendor solutions that can only be deployed in a hosted public cloud mode, offer subscription pricing (SaaS), have Managed Services built-in, and are multi-tenant.
Cognitive Analytics: This is an AI model equivalent to the way the human brain processes information such as learning from interactions and the resulting data from those experiences. Cognitive is the simulation of a large number of interconnected data processing units arranged in input-output layers that resemble an abstract version of neurons in the brain. The values are propagated from each node to other nodes between layers.
Consolidation – Financial Close: Financial consolidation is the process of combining financial data from several departments or business entities within an organization, usually for reporting purposes. The process includes importing data, mapping general ledgers to a single chart of accounts, normalizing the consolidated data, and producing reports called consolidated financial statements. How often the finance department performs financial consolidation varies from company to company, requirements are driven by regulatory and management reporting requirements. Depending on the complexity of the process and the company's preference, finance managers can use Microsoft Excel spreadsheets, ERP (enterprise resource planning) software suites, or, specialty software modules in EPM solutions for financial consolidation.
CPM: Corporate Performance Management, see also EPM - Enterprise Performance Management. Gartner was one of the last holdouts clinging to the term CPM, but has since replaced this market label with xP&A.
Darwin: Pre-built and pre-packaged multi-model application software that supports a wide range of functions now known as xP&A. These functions include Legal Consolidations and Statutory Reporting, Management Reporting, Revenue Planning, Sales Planning, Project Planning, Asset Planning, Cost Center Planning, Employee Planning, and Headcount Planning. When applying the standardized Darwin platform, greater potential solution value is possible to enable high performance and advanced analytics functionality with less time and cost compared to alternatives.
Data Analyst: A role within an FP&A, Operational analysis, or data science team, the data analysts’ goal is to uncover important factors during the data discovery phase that can influence the outcome of the project. The key to achieving success within this role is knowing important business questions that need to be answered, how to source and proportionally weight available data indicators, and understanding how the results will be displayed.
Data Discovery: The conventional definition defines it as the process of exploring data to reveal relationships and hidden patterns, which can be accomplished through machine learning algorithms. SAP defines data discovery as ‘the business user-driven and iterative process of discovering patterns and outliers in data. Contrast discovery from the process of report production where reports with known metrics can be produced in a highly automated fashion on a monthly (or recurring) basis. Discovery is more free form, exploratory and unstructured.
Data Mining: Data mining can either be performed manually by a data mining engineer or as an automated process to get information from raw data to find correlations used to pinpoint key parameters during data exploratory and discovery. Data mining applies modeling and discovery techniques to assess and evaluate data/models quality and accuracy in accordance with a predetermined assessment. The goal of data mining is to define intended outputs and success criteria to achieve business objectives within business terms.
Data Modeling: Considering the variety of modeling techniques available, most have specific requirements for the structure, type, or form of data in use. For example, the size of the data may or may not constrain use for some modeling techniques. Examples are diagnostic modeling, predictive modeling, prescriptive modeling, and descriptive modeling.
Data Scientists: Data scientists are highly trained data professionals with technical expertise specifically designed to solve complex data problems. They are also responsible for building out data models to support algorithms used in delivering data-related solutions and expertly perform tasks known as ‘data wrangling’ to clean up messy data. New York Times reports data scientists to spend 50%-80% of their time mired in the mundane task of collecting and preparing unruly data before it can be explored for anything useful. Data scientists work to remove bias from models to expose true patterns of activity that support improved decision-making and business actions.
DI – Data Intelligence: applies mostly to data scientists and IT specialists. It consists of a cluster of tools of artificial intelligence (AI) and machine learning (ML) algorithms to transform and analyze data. DI encapsulates the concept of uncovering information such as hidden patterns in data, correlations, and trends.
DNA – Dynamic Networked Analytics: is a successor performance analytics business process to Serial Analytics, both of which were first articulated by Column5. Serial analytics focused on a one-way process to extract data from sources, transform the data to a standard, and then display the data in prescribed formats. The time expectation for Serial Analytics is calibrated by the due date of the output – a report or completed plan or budget. Dynamic Networked Analytics can be a vast acceleration of business process cycle time by focusing more on optimizing data refresh rates, enabling reports to be completed on-demand in real-time. DNA requires the use of Artificial Intelligence technology to automate and enhance analysis tasks formerly done by human analysts. For these reasons, DNA encapsulates the most mature business process built on xP&A and AI platforms.
Deep Learning: Deep learning is a machine learning technique that trains computers to learn by example as humans do. It can be likened to mimicking the workings of the human neuropathways in the brain to process data for use in detecting objects, recognizing speech, translating languages, and making decisions. A technique for learning without human supervision, the process is structured to draw correlations from data that is both unstructured and unlabeled. Examples using this technique are driverless cars, voice control consumer devices such as cell phones, tablets, TVs, and hands-free speakers.
Descriptive Analytics: What happened? Is the type of analytics examination of data or content to answer questions using reporting tools, business intelligence (BI), and visualizations tools such as pie charts, bar charts, line graphs, or tables. For example, a report showing dollars spent per customer and year over year changes in sales. This capability represents minimal sophistication from analytics solutions. Incorporating planning into analytics allows estimates as to what might happen in the future as well as identification as to where results deviate from expectations so management can be informed as a precursor to taking action.
Diagnostic Analytics: Why something happened, most often in the past. Diagnostic analytics is a form of advanced analytics that examines data or content to answer the question, “Why did it happen?” It is characterized by techniques such as drill-down, data discovery, data mining, and correlations. Without AI, this function is performed by asking management to explain relationships between market drivers and business outcomes. Because of varying data, experience levels, and timeliness, it should not be surprising that the response quality in explaining why outcomes are the way they are is typically lacking. Incorporating AI into this process enables much larger data sets to be evaluated, and otherwise difficult to detect relationships to be exposed. Maturing an organization’s capability to assess more complex data more quickly drives business value by enabling more nimble responses to change market conditions. This capability ensures optimized business outcomes are more likely.
Drilling Down: Analysis of focusing on a certain slice of the data or particular widget for more insight. SAC, for example, provides users with no data science skills the ability to analyze data to better understand relationships to make confident decisions.
Driver-Based Planning: Driver-based planning is a standardized and mostly automated method of planning that focuses on key drivers to ensure key business targets and objectives are achieved. Normally identified by the business entity, key drivers, or key inputs that drive operational and financial results, are built into the financial model and used to drive relationships between activities, processes, data, and outputs. Examples of key drivers may be the number of cost centers, salaries, products or services sold, prices of actual services rendered, efficiency rates of production, rent, exchange rates, and headcount. Not to be confused with KPI’s, key drivers are ones that influence high-level aspects of the business, whereas KPIs (key performance indicators) are quantifiable measurements, such as targeted percentage increase in market share or revenue at any specific time, that contribute to and influence planning outcomes.
Ensemble Modeling: Ensemble methods in AI/Data Science use multiple learning algorithms to optimize predictive performance, making them more robust because target variables of the dataset are averaged across multiple algorithms. Ensemble modeling strategies are used by data scientists and data analysts to reduce the limitations of any one model and provide better information to business decision-makers. It averages out biases, reduces variance, improves predictive power, and is less likely to over-fit. Examples of ensemble methods for modeling are bootstrapping, bagging, boosting, and stacking multiple machine learning models. The same technique can be applied to performance management analytics processes where AI-generated scenarios and multiple scenarios generated by humans are compared to establish consensus, expose risk, and highlight calls to action.
EPM – Enterprise Performance Management: is the process of maintaining and monitoring performance across the enterprise with the goal of improving business performance. What makes this tool popular among Finance and Accounting is that it is Excel-based, making it easily customizable for an organization with diverse reporting needs. It integrates a variety of performance management processes and tools, such as automation of Excel-based processes through API’s and Serial Analytics. The most potent AI-enabled performance management solutions support visionary Dynamic Networked Analytics (DNA) requirements. See also What is EPM?
Financial Planning: The process of financial planning in business is designed to forecast future financial results and determine how best to use the company’s financial resources in pursuit of the organization’s short- and long-range objectives. Because planning involves looking well into the future, it is a highly creative thinking process as well as an analytical one.
Flash Reporting: A flash report is a summary of the key operational outcomes and financial results of a business. Typically provided by an accounting department to the management team, the report is used to identify issues and potential issues the management team can act on. The information is time-delimited, meaning it will change daily, weekly, periodically, or annually. Some topics will be settled and no longer require attention, while new areas will crop up over time. A flash report can track different kinds of metrics, for example, bottleneck utilization, the status of overdue receivables, customer order fulfillment rate, or the volume of storage currently available in a warehouse. The report is circulated internally since it contains confidential information.
Forecasting: Forecasting is the process of predicting future developments in business based on analysis of trends in past and present data. Business forecasting refers to the tools and techniques used to predict developments in business, such as sales, expenditures, and profits. The purpose of business forecasting is to develop better strategies based on these informed predictions. historical data is collected and analyzed via quantitative or qualitative models so that patterns can be identified and can direct demand planning, financial operations, future production, and marketing operations. For example, forecasting revenue per month based on trends, seasonality, and other external factors.
Hosting: A host is a relationship between a computer connected to other computers for which it provides data or services over a network. A client, or computer hardware device or software that accesses a service and environment available through a hosting server (or provider). Hosting partners will lease environments featuring a specified amount of computing power and deployment options for a subscribed time period. These services can be bundled with managed services and are often accessed via some sort of cloud deployment.
IBP – Integrated Business Planning: is a solution that supports aspects of operational planning and analysis. SAP offers a software product called IBP, which integrates with S/4, SAC, and HANA as a key component in a cohesive analytics stack. As an example, an effective IBP strategy supports multiple departments within an organization in making informed and strategic decisions for product marketing such as new products or types of capital investments to make, projection of customer demand, production capacity to meet this demand, supply chain, and inventory round out the operational activity planning and reporting. SAP IBP, BPC, Column5’s Darwin, and SAP’s SAC (Analytics Cloud) are examples of technology that can support integrated business planning solutions.
IFP – Integrated Financial Planning: This term refers to applications that support the planning of financial activities and outcomes. Financial planning outputs statements and key financial ratios to equip management to evaluate performance momentum and make adjustments to shape financial performance to better meet expectations. Typical components include support for financial statements (income statement, balance sheet, statement of cash flows), and can also include supporting models like labor planning, operating expenses, capital expense, sales & margin planning (from financial purposes as distinct from volume focused planning for operational activity forecasting).
In-memory database: database technology where records are loaded into memory for the fastest possible access and processing. RAM-based data is volatile, meaning losing power will lose data, but it is backed by a slower-moving, more persistent storage of data. Non-volatile memory is the memory that is slower to access than RAM or volatile. HANA is an example of an in-memory appliance, which means specially built hardware to maximize the amount of RAM available to support large-scale database operations in memory. Many database software programs also support in-memory via the allocation of physical RAM on a server. In-memory databases and hardware can be on-premise, procured via PaaS, or bundled with Hosted & SaaS solutions using public or private cloud solutions.
Last Mile of Finance: A term used in the EPM space, is the timeline of the financial business process after close through the filing of reports to stakeholders. It is comprised of all steps finance executives must perform after the monthly, quarterly, or annual close to prepare for financial reporting and disclosure to public and regulatory authorities. Deloitte refers to the last mile of finance as such it "covers the processes and activities in between the trial balance and a company's 10-K." Notably, it consists of four main components: account reconciliation, internal management & financial reporting, financial controls & workflow, and disclosure management & regulatory reporting.
Managed services: this refers to an agreement to have third-party care for hardware or software on behalf of a client organization. For a public cloud, multi-tenant solution, these services are built-in. But those services can be contracted for any solution deployment mode.
ML – Machine Learning: Machine learning focuses on applications that learn from experience and improve their decision-making or predictive accuracy over time without being programmed to do so. Machine-learning algorithms use statistics to find patterns in massive amounts of data. And data, here, encompasses a lot of things—numbers, words, images, clicks, what have you. If it can be digitally stored, it can be fed into a machine-learning algorithm.
Multi-tenant: this term involves multiple other concepts in one offering. Like an apartment building, there is a shared front door, and an amenities package is granted to each tenant, but specific private areas also exist for each resident. Residents do not have access to shared infrastructure components like the heating system, the common roof, or the security system - this is administered by the landlord on behalf of each tenant. Also like a leased apartment, limited authority is granted to each tenant...they cannot remodel or change too much, they have to keep within certain constraints so as not to disrupt the shared infrastructure or other tenants’ experience. A Multi-tenant technology solution
Outliers: During the data analysis phase of a project we study the structure, assess quality, look for errors, and uncover missing observations. Take for example the use of a scatterplot that easily shows anomalies (aka deviations and exceptions), which are referred to as outliers. Outliers may indicate gaps in the model or in data and can be a flag to call attention to areas where the action is required.
PaaS – Platform as a Service: refers to a complete hardware and software stack ready to support one or more applications. PaaS is always hosted in the cloud. SAP’s Cloud Platform (SCP) & Microsoft’s Azure are examples of PaaS. See also: Hosting, Cloud
PBF – Planning, Budgeting, Forecasting: See also: Planning, Budgeting, Forecasting. The process of planning, budgeting, and forecasting should sit right at the heart of every organization PBF is a key component of how information is generated and processed, how decisions are made, and how responses are formulated to steer the organization and impact future performance. The processes involved are tightly linked to many others. They involve and connect many people and functions across the entire enterprise. Integrated and effective PBF processes have a fundamental role to play in identifying and exploiting areas where new growth exists and in modeling and managing risk and uncertainty in plans and forecasts. In this way, PBF becomes the critical business process that it ought to be rather than the financially skewed exercise that is ingrained within many organizational cultures
Predictive Analytics: What could happen? Predictive analytics uses statistical models, machine learning, and forecasting to understand the future. Using historical data it analyses patterns and develops models that predict the probability of future outcomes. The objective is to go beyond knowing what has happened and assess what will happen in the future. This form of analytics has become popular to analyze large volumes of data collected to make decisions, understand customers, recommend products, improve business performance, fraud and risk, sales and marketing, and finance and HR. Care must be taken to not over-fit or under-fit predictive models. Without AI, the function of predictive analytics relies on bottoms-up manual processes – in other words, the process is to ask managers in business areas what they believe will happen in terms of business outcomes. A key cause of variances can be traced to incomplete, inaccurate, and skewed perspectives from those management resources that may rely on their individual ability to evaluate available data. Augmenting & validating human-provided scenarios with computer-generated views should accelerate & improve the quality of the output.
Prescriptive Analytics: What should we do? Prescriptive analytics predicts not only what will happen, but also why it will happen to provide recommendations regarding the action that will take advantage of predictions to make short and long-term business decisions. It uses a combination of techniques and tools such as business rules, algorithms, optimization, machine learning (past performance), current performance, and mathematical modeling and processing. The opposite of prescriptive analytics is descriptive analytics, which examines decisions and outcomes after the fact. Without AI, the function of prescriptive analytics relies on bottoms-up manual processes – in other words, the process is to ask managers in business areas what they believe is the right combination of targets and asset allocations to achieve business outcomes. A key cause of variances can be traced to incomplete, inaccurate, and skewed perspectives from those management resources that may rely on their individual ability to evaluate available data. Augmenting & validating human-provided scenarios with computer-generated views should accelerate & improve the quality of the output.
RPA - Robotic Process Automation: the application of computers to mimic activities performed by humans. In the context of EPM & xP&A activities, this may be the opening of spreadsheets or EPM interfaces, the evaluation of certain data, the calculation of adjusting journal entries, the submission of journal entries, the approval of journal entries, the generation and distribution of reports
Rolling Forecast: Rolling forecasts are a popular alternative to the traditional approach to annual budgeting. It is a methodology that uses historical data run rates modified with the latest influences to predict future budgets, expenses, and other financial data. Rather than managing forecasts based on a static prior year budget, rolling forecasts update budgeting assumptions periodically throughout the year. This enables adapting quickly to shifts in resource allocations impacted by changes in the economy, industry, or business plans. Rolling forecast uses a combination of actual monthly, actual year-to-date, original budget, updated revenue, and expense forecasts to calculate future periods out the standard 4 to 6 quarters (or more if required).
SAC – SAP Analytics Cloud: built on SAP’s Hana Cloud platform, SAC solution supports Business Intelligence, Analytics, Reporting, Enterprise Performance Management/Planning, and Predictive Analytics/AI – all in one. This tool is currently SAP’s flagship solution that serves as a front-end interface for numerous SAP cloud products. Notably, SAC provides the ability to work with a wide range of live data sources and connections through the cloud. SAC is an essential component within xP&A’ s toolkit.
SaaS - Software as a Service: a pricing policy where customers acquire rights to use software for a subscribed timeline for a specific price. SaaS is likened to an office lease, where a recurring payment is made to have rights to use software without ownership. Column5’s Darwin, for example, can be sold via a SaaS license or perpetual. Other examples are SAPs SAC and IBP; which are sold exclusively via SaaS licensing.
S&OP – Sales & Operations Planning: This is a monthly integrated business management process that empowers leaders to focus on key capacities suck as production, supply chain, inventory, and link this with sales, marketing, demand management, and new product introduction. With an eye to evaluate the financial and business impact of changing operational activities, the goal of S&OP software is to enable executives to make better-informed decisions through a dynamic connection of operational plans and strategies across the business. Often repeated on a monthly basis, S&OP enables effective production and supply chain management while focusing the resources of an organization on delivering what their customers need while staying profitable.
Serial Analytics: Serial analytics focuses on a one-way process to extract data from sources, transform the data to a standard, and then display the data in prescribed formats for distribution to human analysts The time expectation for Serial analytics is calibrated by the due date of the output – a report or completed plan or budget. This process is the current standard for EPM and even xP&A solutions, and its effectiveness is diminished by redundant content, manual activities, and latency throughout. A new process called Dynamic Networked Analytics has been defined to rethink performance management by leveraging advanced technologies such as xP&A products, artificial intelligence, and the scalability of big data platforms. Prior to the advent of DNA, the benefit of improving serial analytics was limited to labor savings by reducing manual activities via automation. See also DNA.
Spreadsheet: A spreadsheet is a file that exists of cells in rows and columns and can help arrange, calculate, and sort data. Data in a spreadsheet can be numeric values, as well as text, formulas, references, and functions.
Strategic Planning: Strategic planning is an organizational management activity that is used to set priorities, focus energy and resources, strengthen operations, ensure that employees and other stakeholders are working toward common goals, establish agreement around intended outcomes/results, and assess and adjust the organization’s direction in response to a changing environment. It is a disciplined effort that produces fundamental decisions and actions that shape and guide what an organization is, who it serves, what it does, and why it does it, with a focus on the future. Effective strategic planning articulates not only where an organization is going and the actions needed to make progress, but also how it will know if it is successful.
Time Series: Time series analysis evaluates data trends to enable algorithms and analysts to forecast future KPI values. The intent is to estimate where outcomes will land as influenced by key business drivers with an established relationship to outcomes.
UI/UX: Abbreviation for User Interface/User Experience. User Interface (UI) is the series of screens, pages, and visual elements such as buttons and icons that enable a person to interact with a service, such as screens, keyboards, and sounds. User Experience (UX) is the internal experience with the UI, such as a sequence of actions and impressions interacting with it.
Visualization: Data visualizations can be used to understand relationships within variables of data. Typically statistical metrics (explanatory variables) are used to help represent data in determining impact as characteristics change. Normally graphical, it is a representation of information and data through charts, graphs, and maps. Data visualization tools assist in making business decisions by providing an accessible way to observe and understand trends, outliers, and patterns in data.
Workflow: Workflow describes the steps in a business work process, through which a piece of work passes from initiation to completion; and how these steps can be executed and automated according to a set of procedural rules. Organizations use workflow to coordinate tasks, improve organizational efficiency, add responsiveness, and drive profitability. The workflow may be sequential, with each step contingent upon completion of the previous one, or parallel, with multiple steps occurring simultaneously. Big data has helped to automate workflows through the development of AI and ML systems. Workflow software and apps help businesses manage projects across different teams, locations, and time zones.
xP&A –Extended Planning & Analysis: This is a solution for managing an enterprise's business performance. Extending IFP (Integrated Financial Planning with focus on finance) and IBP (Integrated Business Planning with focus on operations), xP&A combines both into a single data platform. In addition to automating data integration, data collection, and reporting activities, xP&A includes Artificial Intelligence (AI) and predictive analytics functionality. While xP&A is a Gartner defined term for a technology solution category, only base requirements are defined to have a product eligible for inclusion in this category. No specific business process is provided or required to be performed. Typical serial analytics processes like planning, budgeting, and forecasting could be the process. Or an advanced process like DNA could be supported by this technology. See also - What is xP&A?