Ad-hoc Analytics: A process where an end user of a BI tool can create their own visualizations, analyze and transform their own data sources, and make custom reports, without interacting with the SQL layer of their BI system or getting permission from IT. Ad-hoc analytical tools are a major part of most competent self-service BI systems. 

Alerting: Many BI systems can notify users when metrics change in unexpected ways. Businesses set up these alerts so that they can respond in real time to changes and solve problems as soon as they occur. Alerts can include emails, text messages, push notifications, or instant messages to a services such as Slack.

Analytics: A systemic approach to analysing data. Broadly speaking, it's what a BI tool does to data to find insight. A business intelligence program uses standard analytical techniques to discover and communicate trends in an organization's data; the process of applying all these different techniques is called 'analytics'.  

Automation: Using cloud computing, many cloud-based BI systems can do time-consuming data tasks for their users. For example, most BI tools cleanse and normalize their data automatically, with minimal user input. This helps users save time and streamline their workflows. 

Big DataAn industry term used to refer to data sets that are larger and more complicated than regular ones. Large businesses collect massive amounts of data, and these large data sets can overwhelm the capabilities of less powerful BI tools. Big data is often harder to manage and store than regular data, so businesses need to implement specific big data solutions to make full use of it. Businesses can leverage their big data towards deeper insight and use it as the basis for machine learning algorithms. 

For more information on big data software, read our full report here

Business Intelligence: A general term to refer to all of the processes and techniques a business uses to manage and analyze their data. Usually, software that helps businesses use their data in an effective way is called 'business intelligence' software. Not all data analysis tools are business intelligence tools; business intelligence software tends to act as an all-in-one solution for a business's data needs, while smaller data analysis tools may not have complete data functionality.

Some features common to most business intelligence software are: 

  • - Data management features which collect information from other software and collate it into coherent data streams.
  • - Data storage features like data warehouses and data marts which help users access data quickly.
  • - Dashboarding and reporting tools which help users draw conclusions from their data at a glance.
  • - Data visualization tools which allow users to build charts, graphs, and maps from their data.
  • - Predictive analytics which let a business make forecasts and projections about future data trends.
  • - Data mining tools which use complex algorithms to find insights that'd be invisible to human eyes.
  • - Benchmarking and KPI tracking which can help employees track their progress and diagnose problems.

Cloud-based: A piece of software that's not downloaded to company servers or stored on premise. Instead, cloud-based tools connect over the Internet to banks of servers across the globe where the software is actually stored. Businesses don't need to invest in computing power or storage space for their software if they use cloud-based solutions. 

Dashboard: A type of user interface that uses common data visualizations like charts and graphs, along with scorecards and other statistics, to present data in a way that's easy to understand and simple to internalize at a glance. Business intelligence tools almost always have some sort of dashboarding tool, since dashboards make data analysis much easier for the end user. 

To learn more about dashboards, read our ranking of the industry's best dashboarding tools here

Data Cleansing: Sometimes, data sets can have flaws that make them ineffective for data analysis. Often, these flaws are corrupt, inaccurate, or incomplete data entries that the data analysis algorithm doesn't know how to handle. Before a BI tool can use this data, it has to remove, correct, or normalize all these incorrect data entires. This process is called 'data cleansing'. In most BI tools, it's a semi-automated process. 

Data Governance: The strategies a BI tool uses to decide who can access and view what data, when. It'd be impractical to let everyone in an organization see all of that organization's data. BI tools can limit what data certain users or groups of users can see, and what data streams they can access and analyze, so that each employee only needs to focus on the data that's important to them. 

Data Lake: A strategy for storing and managing structured and semi-structured data for later consumption. In a data lake, all of a business's data is stored in one central location, similar to a data warehouse. In a data lake, though, data is stored in its raw, unprocessed form. When a BI tool pulls data out of a data lake, it then has to transform it using an ETL process into structured data that the BI tool can readily consume. This is in contrast to a data warehouse, which stores information that's already been transformed. 

Data Mart: In many cases, it'd be impractical to have users access a business's whole data warehouse every time they need data. If every user is querying the data warehouse, it'll quickly get overloaded with requests and run slower than it should. With a data mart, system administrators can give users access to just a small portion of a business's data. When a user needs data, they'll query their data mart, not the data warehouse, which saves them time and conserves computing power. Data marts also have some data governance applications. 

Data Mining: A process that uses programming and other statistical techniques to analyze data in new and unique ways. It can help discover new connections and trends within already analyzed data sets, which allows businesses to extract more value from the data sets they already have. While data analysis often attemps to prove or disprove theories about a given data set, data mining isn't hypothesis-driven. 

Data Pipeline: The process by which data gets from where it's originally collected to employees who will use it to drive insight. There are many steps in the data piepline, from data collection to data transformation to data visualization. Some of this process takes place outside of a business's BI tool, but BI tools do help to simplify and streamline the process. 

Data Scientist: A person who has specific experience in working with data. Often, these employees have a deeper understanding of the code that drives most of the major BI systems. With legacy BI systems, data scientists are crucial to every part of the data process. In more modern, self-service BI systems, they're less important, since end users can do more of theirown data work. They're still useful for more complex data operations. 

Data Silo: A term used to refer to situations where data is trapped in one piece of software or within one department. With legacy business intelligence systems, data doesn't often flow easily. It can easily get trapped within poorly connected software tools or by departments that don't collaborate well. Employees can't easily access data in data siloes, so they can't use it to drive insight. Modern, cloud-based BI solutions can help break down data siloes. 

Data Source: The original source system that the data originated from. In a BI tool, data sources are often other pieces of software. A BI tool extracts data from the other software in an organization and transforms it so that it can be used for analysis. 

Data Stream: The term used to describe a data set that's being collected, analyzed, and visualized in real time (or close to real time) as part of a dashboard or other end-user analytical tool. Often, a data stream is a combination of data sets from a few different sources. 

Data Visualization: An umbrella term for many different methods of presenting data in a visual way. Most people can understand data much easier when it's presented as graphics like charts, maps, and graphs. BI tools often have tools which allow users to build data visualizations and share them with others. Data visualizations are an important part of dashboarding, since they allow users to interpret data much quicker than other solutions. 

Read our report on the best data visualization tools here

Data Warehouse: A strategy for storing and managing business data. In a data warehouse, all of a business's data is stored in one central location. A BI tool extracts data from other software it's connected to, transforms data into a form that can be used easily, and loads it into a business's data storage system. In a data warehouse, that data is loaded into a central repository where it can be accessed easily. Some data warehouse structures may also have data marts to minimise loading speeds and make data governance easier. For some businesses, a BI tool is powerful enough to act as their data warehouse, while larger companies tend to use third-party data warehousing services. 

Database: A collection of data tables where a business stores its data. There are a few different models for databases, like cloud data warehouses or on-premise transactional datamarts. Some businesses use BI tools as databases, while others use third-party data storage tools or cloud-based storage solutions. 

Descriptive Analytics: A type of data analysis that focuses on collecting and summing up the data that a business needs. Usually, descriptive analytics are very simple and just provide some basic information about the data situation. They describe the data, they don't draw any conclusions from it. For example, a bar chart that shows total sales per month is a descriptive analytic. It's not providing any guesses as to how or why sales changed from month to month, it's just presenting the data.  

Diagnostic Analytics: This type of data analysis tries to find relationships and trends within data to answer specific business questions. It attempts to diagnose the business problem by suggesting relationships between different data points. For example, a business whose sales are dropping may use diagnostic analytics to find out why. Their BI tool will suggest different data sets that may have a relationship to sales figures, in an attempt to isolate the metrics that might affect sales. 

Drill-Down: The process of narrowing a data set to find out more information about a certain data point or set of points. This can be done manually through SQL queries or OLAP analysis, but many BI tools allow users to drill down using interactive features like visual analytics. 

Embedded Analytics: A technique where businesses build their BI tool directly into public-facing apps, sites, and services. Businesses can embed a visualization, dashboard, set of dashboards, or other analytical tools into the tools that their employees and customers already use. This means employees don't need to log into BI systems to see their data, and customers get a more valuable experience. Embedded analytics are becoming very popular, with many companies embedding BI content. 

For a full list of the best embedded analytics vendors, check out our report here

ETL: An acronym that stands for 'Extract, Transform, Load', ETL is a term used to describe the process that most BI tools use to get data out of other pieces of software. A BI tool extracts data from the other software on a network, transforms that data from its raw form into a consistent, uniform style that the BI tool can utilize, and loads that data into the BI tool itself so that it can all be analyzed using one tool. 

Forecasting: Using predictive analytics, BI tools can make educated, data-driven guesses about what a business's data will look like in the future. This process is called forecasting. Businesses can use forecasts to plan ahead with much greater accuracy, which helps to manage a business's risk and let them operate with more efficiency. Forecasting is an essential part of many modern supply chains. 

Interactivity: In the BI space, interactivity usually describes to what degree a visualization reacts to user input or changes in the underlying data. If users can click through data visualizations to learn more, narrow or broaden the scope of the visualization, or the visualization changes over time, the visualization is described as interactive. If a user can't interact with it, and it doesn't change when the data is updated, the visualization is static. Usually, most companies prefer interactive visualizations. 

KPI: Key Performance Indicator. A metric that a business uses to gauge company health. They specifically focus on one project or application, and provide metrics to judge how close that project or application is to success. Businesses can use KPIs to track progress, gauge employee productivity, and find and solve business problems. They're an essential part of project management, but are usually only possible to track effectively using a BI tool. 

Location Analytics: This type of analytics provides an additional layer of insight into business data, by attaching geolocation data to other data sets. With location analytics, businesses can see how their data is arranged spatially. Many businesses track location data in addition to other data points, and then map this location data using data visualization techniques. For example, a retail chain could separate their stores by region, and then look at regional data instead of company-wide data. 

Machine Learning: A paradigm of computer programming where a model is created and then trained on a large set of initial data, using past results to learn for the future. Once properly trained, machine learning algorithms can find insight that might be invisible to a human data scientist, but they usually require big data sets and BI tools that can handle the additional computer load. 

Marketing Analytics: This is a term for analytical techniques tailored for marketing applications. A marketing team may have its data split up among any number of different CRMs, collaboration tools, and other pieces of business software. BI software can analyze the data from all these different marketing tools and provide a marketing team with greater insight than they would have otherwise. 

Read our ranking of the best marketing analytics tools here.

Metadata: Data about data. When an application collects data, it usually also collects some data points related to that data, like when that data was collected, where it was collected, and so on. BI tools can use this related data to find additional relationships that might otherwise be invisible. 

Metrics: Data points and sets that help a business diagnose problems, compare with benchmarks, and track progress on a project. In BI systems, users can combine metrics with other performance indicators and visualizations into dashboards. 

Mobile Analytics: The term for a a BI tool's mobile performance, whether within a phone browser or in its own separate app. Most cloud-based BI tools have some sort of mobile presence, allowing users to work with their BI tool from anywhere. Embedded tools can also help bring BI computing power to a business's apps and mobile sites. 

Click here to see our ranking of the industry's best mobile experiences. 

Natural Language Processing: The technique a BI tool uses to turn user queries typed in plain language into code-based commands. BI tools store data in ways that are somewhat complicated to understand. A BI tool has to communicate in programming language with its database, so that it can properly find the data a user wants to see. Natural language processing, abbreviated NLP, automatically translates user searches into code commands so that users don't have to use code themselves. Many self-service BI tools have powerful NLP algorithms. 

Neural Network: Algorithms that are designed to process information in the same way that a human brain does. They can often see trends that a simple machine algorithm might miss. Many BI tools use them in conjuntion with machine learning algorithms to find novel insight. 

Normalization: The process of turning differently formatted data streams into one consistent data source within a BI tool. For instance, different tools may use different abbreviations for commonly abbreviated words. A BI system can transform all these different abbreviations into one consistent one so that all the data can be analyzed the same way. 

OLAP: An acronym for 'online analytical processing'. An analytical technique that leverages the computing power of cloud-based BI solutions to analyse data with large amounts of computing power. With this computing power, businesses can see the relationships between their data from multiple different dimensions. It can make it easier to see how three or more data sets are related to one another. 

On-premise: Describes when an organization keeps their data, software, and other digital assets downloaded to their own servers. In contract to cloud computing, on-premise is a traditional approach to running BI solutions. It offers advantages in data governance, but most businesses prefer cloud-based solutions. Many BI tools don't even offer an on-premise option. 

Predictive Analytics: This type of analysis uses the data that a business has already generated, plus current trends, to project how those trends will continue in the future. It allows businesses to use their data to make educated guesses about future operations. This is one of the more complicated forms of analysis, but one of the most useful. Many different businesses in many different industries can benefit from predictive analytics. 

For a ranking of the best predictive analytics tools on the market, read our full report here

Qualitative Data Analysis: An umbrella term for all the different analytical techniques necessary to analyze qualitative data, like text. Computers can't analyze text very well. They need to use special techniques, like sentiment analysis and topic modeling, to draw any insight from text and other pieces of qualitative data. These tools used to be very complex and ineffective, but modern, self-service BI tools have made them more accessible. 

Check out our ranking of the best qualitative data analysis tools on the market here

Query: Interacting with a database or BI tool to answer a business question. A Structured query language (SQL) is used to communicate with the database to extract, write, and update information. Queries are also frequently used in BI tools to further clean, transform, and visualize the data.

Self-Service: An approach for building software tools and features where end users without any technical expertise or data experience can still use all of the software's features. By making workflows simpler and more intuitive, software designers can make even the most complex BI procedures accessible to those without any specific training. The term can also describe tools that don't need a lot of support from the vendor to use properly. 

Sentiment Analysis: A type of qualitative data analysis that attempts to gauge the overall emotion of a piece of text by looking for certain keywords and sentence structures. Using this analysis, businesses can better utilize things like customer reviews, by figuring out what percentage of them are positive, which are negative, and so on, without any human input. 

Topic Modeling: An analytical technique for qualitative data that parses pieces of text for certain keywords and content indicators, in an attempt to pick out sections that discuss certain topics. This analysis helps businesses make better use of things like customer reviews, since they can focus on reviews that discuss certain topics without making someone read them all to figure out what they're about. 

Visual Analytics: A technique for data analysis found in many BI tools, where users can query their BI systems by interacting with some sort of data visualization. These interactions can streamline the data analysis process. For example, a user could click on one section of a data visualization to drill down on one aspect of the data set, without manually querying their BI database for that data. 

White Labeling: A technique where an organization buys a BI system, but uses their own branding and sites to run it. White labeling allows companies to create insights for their customers and integrate them into their already-existing tools, such as a web portal or software application. End users won't be able to tell that the business is using an off-the-shelf BI solution because all vendor branding is removed. 

Writeback: The term for inserting new data directly into a source system using a BI tool, without going through another piece of software or an application. Often, embedded analytics and dashboards incluse some sort of writeback functionality to make them more convienent to the end user.