The BI Survey 7 - Summary of chapters

1 - Introduction

Here we explain how to get the best out of The BI Survey 7, before looking forward to the range of analyses you will find in later chapters and presenting a number of headline findings from the report.

2 - The sample

The value of a survey like this depends on having a sufficiently large, well-distributed and unbiased sample. This section describes the characteristics of the people who took part in the study and how we recruited them.

3 - The products

An overview of the products covered by the remit of the Survey. This section also contains notes on the 15 vendors whose products were found to be most widely used.

4 - Age profiles

One of the questions in the survey asked how long ago respondents had purchased their BI software. This chapter presents some interesting analyses of the age profiles cross-referenced with the products used.

5 - The purchase cycle

This section analyzes how people buy BI tools and how satisfied they are. It considers the various influences on their decisions, what factors they feel are important and the resultant success rate with different tools.

6 - The BI ownership experience

This section looks at several aspects of BI deployments, including how widely BI is deployed, the on-going admin/maintenance effort, the cost of ownership and the benefits seen.

There are many ways to evaluate BI products technically, but measuring their customers’ success focuses more on the real purpose of implementation. This allows products that are not technically comparable to be evaluated side by side in a meaningful way.

We therefore asked several questions that helped us measure how successful the deployments had been. The combined results were then used as a calibration tool that is widely used throughout the Survey to assess the correlation between many factors and project success.

In particular, we asked to what extent a range of potential business benefits had been achieved. Achieving business benefits is, after all, the whole purpose of any BI deployment, so it makes sense to compare deployments based on the extent to which this has been done. This is more important than specific technical achievements.We call this combined weighted score the BBI — Business Benefits Index.

This section describes the successes and failures, with a few comments on how uncontrollable factors like organization size, location and data volumes appear to be linked to success rates. The links between controllable factors and success are investigated in each of the relevant sections that follow.

7 - Vendor effectiveness

A software vendor's marketing department does more than run campaigns and design the web site. It is responsible for a whole set of product management, positioning, pricing, packaging and promotion activities, the purpose of which is to help increase sales while keeping selling costs down. We do not try to evaluate the individual activities, but their overall effectiveness, measured in terms of how buyers and prospects behave.

This section measures the effectiveness of BI vendors against a variety of criteria, including marketing, sales, deployment, product support and loyalty. Note that higher scores for vendors in several of these criteria are not necessarily in their customers’ best interests.

8 - Implementation

The success of a project depends at least as much on how well it is implemented as on what product is used.We therefore asked respondents which resources had been used to implement their systems, what the external consulting costs were, and the length of time to the initial rollout.

In this section, we also examine the correlations between those factors and overall project success.We did not ask questions to quantify the amount of internal resources used. There are many aspects of a BI implementation, from basic needs assessment, through product selection, project management, design, application development, testing, tuning, documentation, user education and administration to continuing enhancements. These tasks are often shared between in-house and external resources.

9 - Timescales

Slow (and late) implementations are one of the biggest traditional problems with data warehouse and BI applications. We test this hypothesis with a range of analyses looking at the success and problem rates by time taken and discover which products typically implement fastest.

10 - What goes wrong?

BI deployments depend on a complex choreography of people, data and technology. If any of these fails to perform, the problems soon mount up — and over 70 percent of our respondents identified at least one major problem that had occurred. This section analyzes these problems and how they relate to other factors. It also looks at what would deter people from deploying their solutions more widely.

11 - Applications

BI products can be used to implement a wide variety of applications, and most buyers use them for more than one application. On average, we found that each respondent reported 4.2 applications. This section looks at the application mix across a number of dimensions.

12 - Web BI

For several years, it has been confidently assumed by many observers that BI applications were mainly deployed via Web browsers. Through a series of analyses of web deployment and extranet usage, this section looks at what our respondents actually reported, which turns out to be rather far from the received wisdom.

13 - Server platforms

BI applications have to live in, and be compatible with, a range of other technologies, including platforms and databases. In this section, we report only what respondents say about their current environment, rather than their future plans, which often change.

14 - Client/server combos

Several OLAP servers allow a choice of front-end tools from both the server vendor and third parties, and most BI front-ends can be used with both relational databases and other sources, such as ERP applications and OLAP servers. This section analyzes the usage of different front-ends with OLAP servers, and the data sources used with BI front-ends.

15 - Source databases

Most BI applications draw their data from relational databases, so they need to have good connectivity with standard relational databases. In addition, Oracle and Microsoft are also database vendors more than they are BI vendors, so their BI products are sold as extensions to their relational databases. Most organizations are more likely to have chosen a database than a BI standard, so their BI choices may be influenced by their database standard. This section examines the use of databases in conjunction with BI software.

16 - Data volumes

Most BI vendors enjoy boasting of their products’ scalability. They frequently fail to define what they mean by ‘scalable’, but one common definition is based on the amount of data they can handle. We therefore try to cut through this boastful fog to find out how much data real users report handling in their applications. This section seeks to analyze reported data volumes by a number of parameters before addressing that perennial question: "Is bigger better?"

17 - Performance at the speed of thought?

Poor query performance is by far the most frequently reported product-related problem in both The BI Survey 7 and all previous editions of The OLAP Survey.

By having a better understanding of which products perform well or badly, buyers will be assisted in their product choice, and will know what problems are likely to occur if they choose a product that is known to be slower. They will also have more realistic expectations for performance than they can get from a slick demo with trivial amounts of data and no complex architectures. And if they do choose a product that is known to have performance problems, they should be prepared to invest in faster hardware.

Purely and simply, this section is 'must-have' is you are evaluating any of the products covered in The BI Survey 7.

18 - Appendix: Survey questionnaire

The Survey ran from June 15 to September 4, 2007. All data was captured on-line, from a total of 2746 respondents. This section shows all the questions that were asked.

	Home Product overview FAQ About us Contact us

Why buy? Purchase Register Free Chapter Video	The BI Survey 7 - Summary of chapters 1 - Introduction Here we explain how to get the best out of The BI Survey 7, before looking forward to the range of analyses you will find in later chapters and presenting a number of headline findings from the report. 2 - The sample The value of a survey like this depends on having a sufficiently large, well-distributed and unbiased sample. This section describes the characteristics of the people who took part in the study and how we recruited them. 3 - The products An overview of the products covered by the remit of the Survey. This section also contains notes on the 15 vendors whose products were found to be most widely used. 4 - Age profiles One of the questions in the survey asked how long ago respondents had purchased their BI software. This chapter presents some interesting analyses of the age profiles cross-referenced with the products used. 5 - The purchase cycle This section analyzes how people buy BI tools and how satisfied they are. It considers the various influences on their decisions, what factors they feel are important and the resultant success rate with different tools. 6 - The BI ownership experience This section looks at several aspects of BI deployments, including how widely BI is deployed, the on-going admin/maintenance effort, the cost of ownership and the benefits seen. There are many ways to evaluate BI products technically, but measuring their customers’ success focuses more on the real purpose of implementation. This allows products that are not technically comparable to be evaluated side by side in a meaningful way. We therefore asked several questions that helped us measure how successful the deployments had been. The combined results were then used as a calibration tool that is widely used throughout the Survey to assess the correlation between many factors and project success. In particular, we asked to what extent a range of potential business benefits had been achieved. Achieving business benefits is, after all, the whole purpose of any BI deployment, so it makes sense to compare deployments based on the extent to which this has been done. This is more important than specific technical achievements.We call this combined weighted score the BBI — Business Benefits Index. This section describes the successes and failures, with a few comments on how uncontrollable factors like organization size, location and data volumes appear to be linked to success rates. The links between controllable factors and success are investigated in each of the relevant sections that follow. 7 - Vendor effectiveness A software vendor's marketing department does more than run campaigns and design the web site. It is responsible for a whole set of product management, positioning, pricing, packaging and promotion activities, the purpose of which is to help increase sales while keeping selling costs down. We do not try to evaluate the individual activities, but their overall effectiveness, measured in terms of how buyers and prospects behave. This section measures the effectiveness of BI vendors against a variety of criteria, including marketing, sales, deployment, product support and loyalty. Note that higher scores for vendors in several of these criteria are not necessarily in their customers’ best interests. 8 - Implementation The success of a project depends at least as much on how well it is implemented as on what product is used.We therefore asked respondents which resources had been used to implement their systems, what the external consulting costs were, and the length of time to the initial rollout. In this section, we also examine the correlations between those factors and overall project success.We did not ask questions to quantify the amount of internal resources used. There are many aspects of a BI implementation, from basic needs assessment, through product selection, project management, design, application development, testing, tuning, documentation, user education and administration to continuing enhancements. These tasks are often shared between in-house and external resources. 9 - Timescales Slow (and late) implementations are one of the biggest traditional problems with data warehouse and BI applications. We test this hypothesis with a range of analyses looking at the success and problem rates by time taken and discover which products typically implement fastest. 10 - What goes wrong? BI deployments depend on a complex choreography of people, data and technology. If any of these fails to perform, the problems soon mount up — and over 70 percent of our respondents identified at least one major problem that had occurred. This section analyzes these problems and how they relate to other factors. It also looks at what would deter people from deploying their solutions more widely. 11 - Applications BI products can be used to implement a wide variety of applications, and most buyers use them for more than one application. On average, we found that each respondent reported 4.2 applications. This section looks at the application mix across a number of dimensions. 12 - Web BI For several years, it has been confidently assumed by many observers that BI applications were mainly deployed via Web browsers. Through a series of analyses of web deployment and extranet usage, this section looks at what our respondents actually reported, which turns out to be rather far from the received wisdom. 13 - Server platforms BI applications have to live in, and be compatible with, a range of other technologies, including platforms and databases. In this section, we report only what respondents say about their current environment, rather than their future plans, which often change. 14 - Client/server combos Several OLAP servers allow a choice of front-end tools from both the server vendor and third parties, and most BI front-ends can be used with both relational databases and other sources, such as ERP applications and OLAP servers. This section analyzes the usage of different front-ends with OLAP servers, and the data sources used with BI front-ends. 15 - Source databases Most BI applications draw their data from relational databases, so they need to have good connectivity with standard relational databases. In addition, Oracle and Microsoft are also database vendors more than they are BI vendors, so their BI products are sold as extensions to their relational databases. Most organizations are more likely to have chosen a database than a BI standard, so their BI choices may be influenced by their database standard. This section examines the use of databases in conjunction with BI software. 16 - Data volumes Most BI vendors enjoy boasting of their products’ scalability. They frequently fail to define what they mean by ‘scalable’, but one common definition is based on the amount of data they can handle. We therefore try to cut through this boastful fog to find out how much data real users report handling in their applications. This section seeks to analyze reported data volumes by a number of parameters before addressing that perennial question: "Is bigger better?" 17 - Performance at the speed of thought? Poor query performance is by far the most frequently reported product-related problem in both The BI Survey 7 and all previous editions of The OLAP Survey. By having a better understanding of which products perform well or badly, buyers will be assisted in their product choice, and will know what problems are likely to occur if they choose a product that is known to be slower. They will also have more realistic expectations for performance than they can get from a slick demo with trivial amounts of data and no complex architectures. And if they do choose a product that is known to have performance problems, they should be prepared to invest in faster hardware. Purely and simply, this section is 'must-have' is you are evaluating any of the products covered in The BI Survey 7. 18 - Appendix: Survey questionnaire The Survey ran from June 15 to September 4, 2007. All data was captured on-line, from a total of 2746 respondents. This section shows all the questions that were asked.