Select the Right Business Intelligence and Analytics Tool for the Right User

Summary

Data and analytics leaders want to reduce the complexity of their BI and analytics tool portfolios by using fewer vendors, but no one vendor offers all the best capabilities, and innovations challenge a single-vendor approach. Instead, their strategy should match the right tool to the right user.

Overview

Key Challenges

• No one vendor offers the best capabilities across the full spectrum of business intelligence (BI) and analytics capabilities. Nor does any one vendor meet all associated strategic requirements, such as for global presence, long-term vision, and attractive pricing and packaging.

• Small, niche vendors may offer some of the best functionality, but they may be riskier investments in terms of viability, market presence and availability of skills.

• Industry innovations often come from startups, which means companies have to expand their technology portfolios or risk falling behind.

• New BI buying patterns, such as "land and expand," as well as BI marketplaces, support more decentralized buying of software, which is in contrast to the traditional purchasing of software by central IT departments.

Recommendations

Data and analytics leaders should:

• Stop pursuing a single-vendor approach for all BI and analytics capabilities when this jeopardizes achievement of business benefits.

• Segment users and deploy corresponding BI and analytics tools, based on each segment's use cases and required capabilities.

• Be prepared to use capabilities from multiple vendors to provide the best functionality and keep pace with innovation.

• Make the most of technology investments by reducing overlap within a tool category or use case.

Contents

• Introduction
• Analysis
  • Stop Pursuing a Single-Vendor Approach for All BI and Analytics Capabilities When This Jeopardizes Achievement of Business Benefits
  • Segment Users and Deploy Corresponding BI and Analytics Tools, Based on Each Segment's Use Cases and Required Capabilities
  • Be Prepared to Use Capabilities From Multiple Vendors to Provide the Best Functionality and Keep Pace With Innovation
  • Make the Most of Technology Investments by Reducing Overlaps Within a Tool Category or Use Case
• Gartner Recommended Reading

Tables

• Table 1. Categories of Business Intelligence and Analytics Tool, With Representative Products

Figures

• Figure 1. Specialty Vendors Show Higher Achievement of Business Benefits Than Megavendors
• Figure 2. Funding for BI Projects Is Often Driven by the Business
• Figure 3. Match the Business Intelligence Tool to the User Segment
• Figure 4. Levels of Skill and Capability Required by Four Popular Business Intelligence and Analytics Tool Modules
• Figure 5. Decision Matrix for Business Intelligence and Analytics Tools

Strategic Planning Assumptions

By 2017, virtually all new analytic software purchases will begin as free or low-cost proofs of concept, enabling buyers to try the software before they buy.

By 2018, smart, governed, Hadoop-, search- and visual-based data discovery will converge into a single set of next-generation data discovery capabilities as components of a modern business intelligence and analytics platform.

Introduction

In the early days of BI, vendors primarily delivered one tool: an ad hoc query module that helped non-SQL-savvy users generate their own reports. Then came online analytical processing (OLAP), dashboards, mobile BI, data discovery and cloud BI. Today, most leading BI and analytics vendors offer these capabilities as part of a platform, with varying degrees of integration and maturity per module.

New capabilities continue to emerge, including self-service data preparation, natural-language query and search, and Hadoop-based data discovery: Some of these capabilities are available only from startups. Larger vendors may look to these startups for inspiration and mimic their innovations, partner with them or eventually acquire them. Alternatively, they may decide not to invest in certain innovative capabilities that they think will not gain market traction.

Three examples of these innovative capabilities:

• Self-service data preparation capabilities are available from startups such as Trifacta and Paxata. In addition, these capabilities are a differentiator for Alteryx, which partners with Tableau, Qlik and Microsoft (for Power BI). At the same time, each of these partners has introduced and enhanced its own self-service data preparation capabilities, while Alteryx has continued to enhance its interactive visual exploration capabilities. The differences are in the depth and robustness of these vendors' offerings (see "Critical Capabilities for Business Intelligence and Analytics Platforms" and "Market Guide for Self-Service Data Preparation for Analytics" ). At some point, the self-service data preparation capabilities of Tableau, Qlik and Microsoft (Power BI) might remove the need for partnership. Or self-service data preparation might become a focus of industry convergence, as we have recently seen IBM partner with Datawatch to boost the data preparation capabilities of Watson Analytics.

• Natural-language query and search is an area in which some vendors have made major investments. SAP introduced the concept in SAP BusinessObjects Explorer and plans to bring the capability to Lumira. Microsoft offers Q&A as part of Power BI. However, Tableau, MicroStrategy and SAS currently lack such capabilities. This is giving rise to search-based data discovery vendors such as AnswerRocket, Incorta and ThoughtSpot.

• Hadoop-based data discovery enables users to query and analyze data within Hadoop and NoSQL data sources. This sector includes vendors outside the top 15 (by market share), such as Platfora and Datameer, that began in the big data discovery sector and have since expanded to include support for relational data sources. Some data discovery products can now access data in Hadoop as well. However, they all have different approaches in how they access the data and whether they extract it into their own in-memory analytical engine (their visualization and data preparation capabilities also vary). At the same time, traditional BI vendors like MicroStrategy are adding direct Hadoop Distributed File System (HDFS) connectivity to their modern BI platforms, which reduces the need to use specialist providers.

With all these innovations, data and analytics leaders must assess when to buy, whom to buy from, and how long to wait for their preferred or incumbent vendor to provide a solution. Often, the large BI and analytic vendors are the last to deliver robust products in emerging areas. Data and analytic leaders who wait for a big BI vendor to provide a solution will not have an innovative set of capabilities.

In some areas of life we have a good understanding of why we need different tools: different knives for cooking; different tools, such as hammers and screwdrivers for home repair jobs; and different modes of transport for different journeys. But we lack this understanding for BI and analytics tools. In part, this is due to a disconnection between the users of the tools (who understand the different capabilities they need) and the senior managers (who make the investment decisions).

For companies to realize the full potential of data and analytics, they need multiple capabilities, and currently, that means multiple tools. Whether you can get those multiple tools from a single vendor depends on cost, benefit, timing and capabilities.

Regardless of which tools and technologies you invest in, if you do not also invest in user enablement, evangelism, continued innovation, and organizational approaches that foster best practices in BI and analytics, then your BI and analytics will not achieve their full impact.

Analysis

Stop Pursuing a Single-Vendor Approach for All BI and Analytics Capabilities When This Jeopardizes Achievement of Business Benefits

Megavendors and large independent vendors offer multiple BI tool modules. Buying multiple modules from a single vendor may give you more licensing leverage, but it does not necessarily mean you will get a lower cost of ownership. Cost of ownership is also impacted by ease of use and the degree of integration between modules. Furthermore, cost should be a secondary consideration to the achievement of business benefits.

Figure 1 shows that certain types of product deliver more business benefit than others (see also "Survey Analysis: Customers Rate their BI Platform Ownership Costs" ). The offerings of data discovery vendors, small independents and cloud BI vendors do better in this regard than those of megavendors and large independents, according to Gartner's customer survey data.

Figure 1. Specialty Vendors Show Higher Achievement of Business Benefits Than Megavendors BI = business intelligence; BIPOC = business intelligence platform ownership costs

Source: Gartner (May 2016)

The problem, of course, is that BI costs are more readily identifiable than the benefits they deliver. And the IT department, to which many BI teams report, is often measured in terms of cost containment. It may, therefore, be preferable to allow the business to make the BI tool investments and to determine the potential benefit of bringing in capabilities from an additional supplier, or, conversely, the opportunities liable to be missed by sticking with a single-vendor solution.

Judging from an analysis of Gartner's ITScore maturity assessments, the business already drives the majority of BI and analytics investments either directly from the business unit (8% and 17%) or via a steering committee (17%); in only 45% of cases do the investments come primarily from the IT budget as an IT cost center (see Figure 2).

Figure 2. Funding for BI Projects Is Often Driven by the Business ITScore reports are based on Gartner self-assessment surveys. Data from 2015 derives from 511 assessments.

Source: Gartner (May 2016)

Although cost is often identified as a reason for buying from a single vendor, the promise of integration and the ability to navigate seamlessly from one module to another are also compelling reasons. However, integration is often a "work in progress" for many vendors. The extent to which modules are integrated often depends on whether a vendor has acquired technology or whether a different development team has brought new capabilities to market. Consider the following:

• Microsoft Power BI could not initially access content from Microsoft Analysis Services cubes. Microsoft later added support for this source, but, at the time of writing, drill-down within a table visual is not supported.

• Oracle Big Data Discovery cannot consume content from Oracle Business Intelligence Enterprise Edition (OBIEE) or those data models.

• IBM Watson Analytics initially could not access any content from Cognos. It now supports content from list-style reports only. Support for content from more complicated reports and Framework Manager models is on IBM's roadmap.

• SAP Lumira content can be published to an SAP BusinessObjects BI Launchpad portal, but the scheduling and mobile capabilities available to SAP BusinessObjects Web Intelligence are not available to Lumira content.

Customers, therefore, should understand the degree of integration between modules and not assume that integration is present when buying a range of capabilities from a single vendor.

Segment Users and Deploy Corresponding BI and Analytics Tools, Based on Each Segment's Use Cases and Required Capabilities

Different users require different tools, depending on the following factors:

• Technical expertise: The degree to which a user can write or customize SQL, as opposed to building queries using a semantic layer that hides the complexities of the physical data model.

• Data knowledge: The degree that a user understands nuances in the data, such as sales based on invoice date, versus sales based on orders, not yet paid or net of discounts, for example.

• Governance: The degree to which access to data should be tightly controlled. Salary data and personal identifiers, for example, require more governance than anonymized, aggregated and/or public data, which can be shared more openly.

• Detailed versus summary data: Whether a user needs access to detailed data in near real time, or only to summarized data with a degree of latency.

• Types of data source: Transactional data generated by order and accounting systems accounted for many early BI and analytics efforts, but users increasingly want to analyze new types of data, such as clickstream, sensor-generated, social and semistructured data, as well as internal and external data sources.

• Predictability of questions: External stakeholders and front-line workers may have predictable questions, such as what are the total sales this month or what orders to ship today. However, managers and information workers who are trying to understand why a metric is performing in a certain way, or who want to explore an opportunity or test a hypothesis, have less predictable questions. Thus, a predefined dataset or report will rarely meet their needs.

• Analytic robustness and job content: Some jobs come with a responsibility for data and analysis; others are more concerned with consuming a limited set of data. Some tools are good for list-style reporting; others include advanced analytics, such as for forecasting and clustering, without necessarily leading to a full advanced analytics platform.

• Mobility: Field workers or traveling executives may need access to data via smartphone or tablet. A data scientist, on the other hand, may mainly work from a desktop PC.

• Cost and availability of skills: Broadly used BI modules that serve the masses may be easier to find, internally and externally; specialist modules that require a higher skill set may be scarce, internally and externally, and may be higher-cost.

Individual tools can, however, be used for multiple purposes and multiple classes of user. Figure 3 shows some core modules and their positioning for a spectrum of users. IT developers, for example, will use production reporting tools to author standard reports published to an information portal. Front-line workers, customers, suppliers and regulators will access data through interactive fixed reports. However, an ad hoc query tool could also be used by an analyst or information worker to author such reports. Likewise, executives and managers will want information delivered via dashboards, so they can readily track trends and problem areas; front-line workers, such as call center operators, may also use dashboards to track current call center performance. Dashboards can be authored using a data discovery tool, or via a specialty dashboard-authoring tool with more robust key performance indicator tracking and alerting.

Figure 3. Match the Business Intelligence Tool to the User Segment

Source: Gartner (May 2016)

Figure 4 shows the levels of skill and capability required for four popular BI and analytics tool modules.

Figure 4. Levels of Skill and Capability Required by Four Popular Business Intelligence and Analytics Tool Modules

Source: Gartner (May 2016)

Be Prepared to Use Capabilities From Multiple Vendors to Provide the Best Functionality and Keep Pace With Innovation

Although customers would like to standardize on a single vendor for multiple solutions, vendors innovate and acquire technology at different paces.

Table 1 provides a sample list of vendors and positionings for the most popular BI and analytics tool modules (for fuller definitions and listings, see the "Recommending Reading" section). Table 1.   Categories of Business Intelligence and Analytics Tool, With Representative Products

BI = business intelligence; OBIEE = Oracle Business Intelligence Enterprise Edition; OLAP = online analytical processing Add/Remove Columns

 

Source: Gartner (May 2016)

Customers should develop a use case matrix that guides users through a decision-making process for when to use a particular tool.

Figure 5 shows a sample decision tree based on timeliness, granularity and source of data.

Figure 5. Decision Matrix for Business Intelligence and Analytics Tools EDW = enterprise data warehouse

Source: Gartner (May 2016)

The number of tools used may pose challenges to data integration at the back end, as well as to the presentation of results at the front end. Customers can minimize such challenges by putting the intelligence and reusable data manipulations in a model that can be consumed and shared across the various tools. Sometimes, though, the use case does not support that approach and more flexibility is required. At the front end, consider the degree to which either the BI and analytics tool or a third-party portal allows content from multiple tools to be stored and accessed from a central area.

Make the Most of Technology Investments by Reducing Overlaps Within a Tool Category or Use Case

Minimizing the number of competing BI tools offers several potential benefits:

• Lower cost of ownership

• Improved internal support as experts in a BI competency center or analytics center of excellence have fewer tools to support

• Reduced burden on business users as they have fewer tools to learn

• Seamless integration for easy navigation from one module to another

Organizations can achieve these benefits by limiting the number of tools per category and use case. For example, in the production reporting tool category, if you have both Microsoft Reporting Services and SAP Crystal Reports, pick just one of these products for use in new initiatives. In the ad hoc query category, if you have both IBM Cognos and OBIEE, again, settle on just one. The legacy tool then enters maintenance mode — new content is not developed in it.

Minimizing overlap within a category can be difficult if particular business units have a strong preference for a particular vendor or product. Companies may have multiple tools per category for a number of reasons, such as:

• The company acquires a business that had standardized on tools from a different vendor.

• Business units buy their own BI and analytics tools, so many such tools are already deployed.

• BI capabilities are embedded in a purchased analytics application.

One large media company, for example, traditionally used SAP BusinessObjects for its ad hoc query needs, but its finance department decided when implementing Oracle E-Business Suite that the integration of financial reporting would be better if it used OBIEE, which was being offered at a substantial discount. In this use case, OBIEE could be the standard tool for financial data, but this could also force users to use multiple tools, depending on the data they wish to analyze. It would be better to select a single vendor for this category.

In another example, a large medical manufacturing company was an early adopter of TIBCO Spotfire, but acquired a company that uses Tableau software for its data discovery capabilities. Users of both companies are satisfied with their respective products, in which they have several years of content and expertise. In this case, too, it would be better to choose a single product, but the likelihood of resistance to change, disruption and user dissatisfaction suggests that the more practicable way forward is to support both products and specify the standard by business unit, for as long as the acquired company works autonomously. The company could also decide to use a single product for new content, based on functional requirements. For example, if advanced analytics is a requirement, this is currently an area of strength for TIBCO Spotfire. If mobile capability is a key requirement, Tableau would be a better alternative. (For more information, see "Critical Capabilities for Business Intelligence and Analytics Platforms." )

Having multiple tools per segment can increase support costs. When a single user has to learn and use multiple tools, this may impair productivity and the achievement of business benefits. But migrating all content to a new standard can also be costly in terms of redevelopment and disruption to users. For these reasons, companies may decide to build new content in a designated strategic tool, but continue to support existing content in multiple BI tools. When multiple tools are supported within a particular category, it's important to provide users with a decision tree to help them identify which product to use when.

Gartner Recommended Reading

Some documents may not be available as part of your current Gartner subscription.

"Magic Quadrant for Business Intelligence and Analytics Platforms"

"Critical Capabilities for Business Intelligence and Analytics Platforms"

"Magic Quadrant for Advanced Analytics Platforms"

"How to Architect the BI and Analytics Platform"

"Eight Steps to Picking the Best Self-Service BI and Data Discovery Tool"

"Think Twice Before Changing Business Intelligence Tool"

"How to Implement a Modern Business Intelligence and Analytics Platform"

"Survey Analysis: Customers Rate Their Business Intelligence Platform Ownership Cost"

"Market Guide for Self-Service Data Preparation for Analytics"