Tag Archives: Analytics

[repost ]22个免费的数据可视化和分析工具推荐

original:http://news.cnblogs.com/n/99297/

本文总结推荐22个免费的数据可视化和分析工具。列表如下:

数据清理(Data cleaning)

当你分析和可视化数据前,常需要“清理”工作。比如一些输入性列表“New York City” ,同时其他人会说”New York, NY” 。因此你需要标准化这些工作,使看到统一的样式。下面的两个工具被用来帮助使数据处于最佳的状态。

1、DataWrangler

斯坦福大学可视化组(SUVG)设计的基于web的服务,以你刚来清理和重列数据。点击一个行或列,DataWrangler 会弹出建议变化。比如如果你点击了一个空行,一些建议弹出,删除或删除空行的提示。它的文本编辑很cooool。

clip_image002

2、Google Refine

Google Refine。用户在电脑上运行这个应用程序后就可以通过浏览器访问之。这个东西的主要功能是帮用户整理数据,接下来的演示视频效果非常好:用户下载了一个 CSV 文件,但是同一个栏中的同一个属性有多种写法:全称,缩写,后面加了空格的,单数复数格式不一的。。。但是这些其实都代表了同一个属性,Google Refine 的作用就是帮你把这些不规范的写法迅速统一起来。

clip_image003

统计分析(Statistical analysis)

有时,你需要你的数据的图形化的表达。

3、R 项目

R语言是主要用于统计分析、绘图的语言和操作环境。虽然R主要用于统计分析或者开发统计相关的软体,但也有人用作矩阵计算。其分析速度可比美GNU Octave甚至商业软件MATLAB。

clip_image004

虚拟化应用与服务(Visualization applications and services)

这些工具提供了不同的虚拟化选项,针对不同的应用场景。

4、Google Fusion Tables

Google Fusion Tables 被认为是云计算数据库的雏形。还能够方便合作者在同一个服务器上分享备份,email和上传数据,快速同步不同版本数据,Fusion Tables可以上传100MB的表格文件,同时支持CSV和XLS格式,当然也可以把Google Docs里的表格导入进来使用。对于大规模的数据,可以用Google Fusion Tables创造过滤器来显示你关心的数据,处理完毕后可以导出为csv文件。

Google Fusion Tables的处理大数据量的强大能力,以及能够自由添加不同的空间视图的功能,也许会让 Oracle,IBM, Microsoft传统数据库厂商感到担心,Google未来会强力介入数据库市场。 

clip_image005

5、Impure

clip_image006

Impure,允许点击、拖曳来连接模块,由西班牙分析公司Bestiario 创建。

6、Tableau Public

clip_image007

7、Many Eyes

clip_image008

clip_image009

8、VIDI

clip_image010

9、Zoho Reports

clip_image011

10、Choosel

clip_image012

11、Exhibit

clip_image013

12、Google Chart Tools

clip_image014

13、JavaScript InfoVis Toolkit

clip_image015

14、Protovis

clip_image016

15、Quantum GIS (QGIS)

clip_image017

16、OpenHeatMap

clip_image018

17、OpenLayers

clip_image019

18、OpenStreetMap

clip_image020

19、TimeFlow

clip_image021

20、IBM Word-Cloud Generator

21、Gephi

clip_image022

22、NodeXL

clip_image023

[repost ]MySQL Conference & Expo 2011:Open Source reporting tools

origianL:http://en.oreilly.com/mysql2011/public/schedule/detail/17288

David Stokes (Oracle)

There is an ever growing needs for reports from the data in your datbases. And there are many open source projects that will let you quickly produce ad-hoc reports with sharp graphics from you MySQL instances.

So what do you need to download? How do you connect it to your MySQL or other database? Can I add graphics, charts, and templates to the reports? What do these reports look like? And will they look better than something dumped into Microsoft Excel? If you are novice to the reporting tools, this session will guide you through their commonalities and difference and help you decided which one is best for your environment.

There are several open source reporting tools from Penatho, Actuate, Jaspersoft and others that can make a DBAs lies much easier.

This session is designed for novices in the business intelligence world who are novice to intermediates in the mySQLDBA world.

David Stokes

Oracle

I am a long time user of MySQL and was the MySQL Certification manager.

[repost ]大数据时代的数据管理

original:http://www.infoq.com/cn/articles/big-data-management

不知怎么地,大数据,Big Data这个词就变得流行起来。

处理大数据惯常是属于商业智能(BI)的事情。抽取数据、挖掘数据,制成报表、OLAP、仪表盘、挖掘模型,作为辅助决策之用。不过在BI领域都不这么叫法,大伙儿都说海量数据,Large-scale Data。这听起来还是略显学术气,不如Big Data来的通俗——大数据。这大概是因为如今随处可见的数据,一种爆炸效应带来的结果,已经脱离某种专业的范畴,人们需要用更简单的术语来命名这种数据爆炸。这给不温不火的BI带来一些新的刺激,让BI人看到一些希望。

以前,不说国内,就算是国外,做BI也大多是局限在几个大行当,电信、金融、零售、政府,他们需要数据来帮助自己理性决策。在国内很长一段时间里,更是仅限于电信和金融两个行当。可是尴尬的地方在于,决策者有时候更愿意相信自己的直觉,而非数据。这种意识虽然逐渐在变化,可从来没有发生过根本的变化。意识的变化是艰难的。当一些新兴行业的介入,他们对数据的利用方式,价值的榨取,让人看到数据分析不仅仅用于辅助决策,而是可以从数据中获得收益了,它已经不再是一种锦上添花的东西了,那正是因为大数据时代的到来。这得感谢互联网以及还未兴起的物联网,在这些行当里面,数据在爆发,不断增长。他们不甘心只是如报表、OLAP、仪表盘之类的分析应用。数据分析部门可以按照推荐系统的点击效果利润分成;交易的数据可以包装成分析服务销售给商户,让他们自己去洞察市场商机;根据用户的点击流行为和上网内容,个性化广告布放等等。

就在刚过去的9月,TDWI(数据仓库学院)发布了2011年第四季度最佳实践报告,而这份最佳实践的主题正是大数据分析。TDWI会通过调查问卷的方式,对全球范围的企业调查,目标对象既有IT人,有业务单位的人,也有咨询顾问。问卷的问题一般都会询问企业应用BI技术的实际情况,现在如何,计划如何。所以,这类最佳实践报告可以反映出当下某项技术的现状和趋势。报告的内容也遵循一定结构,一下定义,二看现状,三分长短,四谈趋势,最后再来个厂商介绍。同样,这份大数据分析的最佳实践报告也是如此结构。

其中关于“大数据”的定义,值得关注。如果我们仅仅从字面上看,大数据似乎跟海量数据差别不大,仅仅是变得更加通俗?并非如此,这份报告给出一些区别,TDWI赋予这个术语更多的含义,更多符合目前数据爆炸时代的含义。

大数据的3V

Big Data的3V。大数据有3V的特性。

Volume、Variety、Velocity。这3V表明大数据的三方面特质:量大、多样、实时。对,不光是数据量大了。对TB、PB数据级的处理,已经成为基本配置。还能处理多样性的数据类型,结构化数据和非结构化数据,能处理Web数据,能处理语音数据甚至是图像、视频数据。实时。以前的决策支持时代,可以用批量处理的方式,隔夜处理数据,等决策者第二天上班,可以看到昨天的经营数据。但现在的互联网时代,业务在24小时不间断运营,决策已经不是第二天上班才做出,而是在客户每次浏览页面,每次下订单的过程中都存在,都会需要对用户进行实时的推荐,决策已经变得实时。

这个定义非常完美,形式上也很漂亮,3个V。

可细细想想,这每个方面的V,难道不是传统BI一直在试图征服的嘛?也许所谓大数据时代,是新瓶装旧酒。只是换了一个称呼,而具体要解决的问题,仍是那些存在已久的问题。可毕竟大数据时代轰轰烈烈地,踏着旧的海量数据浪潮而来,而且这将是更高一浪。平常人站在下面,是否会腿脚发软,或是识破浪头的力度,来个漂亮的转身冲浪呢?

大数据管理的需求与挑战

在这样的大数据时代,数据仍然是最关键的。如何将大数据管理好,仍然是对企业的考验。

无处不数据。手机通话、移动在产生数据,ATM在产生数据,商品上的RFID在产生数据,包裹从一个城市到另一个城市在产生数据。就算是一个小小的店铺,当它销售出去一瓶水,也可能会记录到Excel里面,产生了数据。数据记录这世界的存在和变化。

当企业的某项资产非常重要,数量巨大时,就需要有效管理。如今,数据已经成为这种资产。以前人们还不会将它看做是资产,而是一种附属物。客户来办理业务,在系统中产生了这种附属物。而现在,发现在客户办理业务这条信息中,蕴含这一些客户的需求,成千上万条这类信息累积下来,就能洞察客户所需,为设计新产品,为客户个性化营销产生新的价值。数据变成一种资产了,需要被管理起来。

数据仓库是管理数据的工具。在近二三十年里,以某种类似蜗牛的速度爬行,它始终还是贵族家的玩具。只有那些多金的买主才会为它买单。这让数据管理变得高高在上,数据当做资产只是停留在理念层面。人们还在争论着,数据仓库能够给我们带来什么?

我自己曾总结过一句话,体现数据仓库的六项价值——“能快速、及时、方便、准确而安全地访问整合过的数据。”现在看看,发现这个描述还蛮符合大数据时代,对数据管理的需求。

而这六方面价值也对应了不同的技术领域。

  • 数据仓库硬件、软件、模型要保障对数据的快速访问。比如专用设备,按照数据温度选择数据是否高速存储,采用特殊存储技术;
  • DW模型确保数据的整合性,当你需要企业视图的数据,需要以年为周期的数据,需要数据模型的支持;
  • ETL保障数据及时性。批量的ETL已经不足够,需要准实时,甚至是数据流式处理;
  • 元数据管理让数据访问更方便,不仅仅将数据以表、字段的方式管理,要将数据切分地更小,可管理;
  • 数据质量管理保障数据的准确一致,让数据可信;
  • 数据仓库架构、权限管理保障数据访问安全。

大数据时代对六项价值之一——快速访问数据的性能,有明显推动。人们最迫切的希望还是从无到有,从慢到快吧。让数据唾手可得。

数据库技术在变化

传统数据库并未专为数据分析而设计,数据仓库专用设备的兴起(Data Warehouse Appliance),如Teradata、Netezza、Greeplum、Sybase IQ等等,正表明面向事务性处理的传统数据库和面向分析的分析型数据库走向分离,泾渭分明。数据仓库专用设备,一般都会采用软硬一体,以提供最佳性能。这类数据库会采用更适于数据查询的技术,以列式存储或MPP(大规模并行处理)两大成熟技术为代表。另外,新兴的互联网企业也在尝试一些新技术,比如MapReduce技术(这得感谢Google将它发扬光大),Yahoo的开源小组开发出Hadoop,就是一种基于MapReduce技术的并行计算框架。在2008年之前,Facebook就在Hadoop基础上开发出类似数据仓库的Hive,用来分析点击流和日志文件。几年下来,基于Hadoop的整套数据仓库解决方案已日臻成熟。目前在国内也有不少应用,尤其在互联网行业的数据分析,很多就是基于这个开源方案,比如淘宝的数据魔方。而在一些商业性的产品中,也已经融入MapReduce技术,如AsterData。

低廉的数据仓库解决方案降低了数据管理的门槛,长尾的中小企业不一定非得去跟Oracle、IBM这样的大公司去谈高高在上的价格。开源的产品,配置足够的硬件存储,有一支专业的服务团队,就可以架构一个数据仓库平台。在去年,就曾有多位朋友向我咨询的数据仓库方案,他们有一个不约而同的期望,价格不要太高。他们有服务团队。我没有其他推荐,只有推荐Hadoop。

还有一些其他的技术可以让数据访问性能提高,比如数据温度技术,可以区分经常被访问和很少被访问的数据,经常访问的就是高温数据,这类数据将存储在高速存储区,访问路径会非常直接,而低温数据则可以放在非高速存储区,访问路径也可一些相对复杂一些。近两年,存储访问的技术也在变化着,比如Teradata前几年推出固态硬盘数据仓库,用接近闪存的性能访问数据,比原来在磁盘上顺序读取数据快很多。后来又兴起一批内存数据库产品,这类产品在DBMS软件上进行优化,规避传统数据库(数据仓库)读取数据时的磁盘IO操作,再次大大节省访问时间。比如SAP的HanaBI、Oracle的TimesTen、SolidDB、extremeDB、Altibase。

文本、语音、图像、社交网络、地理位置…大数据时代的数据类型如此丰富。用关系型数据库存储这类数据,再深入去分析挖掘这些数据,开始有些负累。

于是,越来越多的NoSQL数据库涌现出来,其中很大一部分是用于分析用途。比如西班牙有个小厂商,叫illumnate,他们拥有一个叫Correlation DBMS的数据库产品。它不像关系数据库那样按照表、字段存储,那样冗余很大。CDBMS的做法是,针对每个不同的值,只有一个地方存储,而所有对这个值的引用,都在索引中记录。比如有个客户的姓名叫“张三”,而还有一个公司名字也叫“张三”,那么在CDBMS里面,只存有一个“张三”这个值,但在索引里面记录了有两个地方引用它。这种数据库是专门为分析而设计的。因为不存储冗余数据,所以它对于海量数据,非常节省空间。如果说这个有点不太吸引人的话,另一个据称的优点就是做ad-hoc查询非常快捷。

社交网络很火热,Facebook、Twitter、QQ、MSN,甚至是普通的电信通话、邮件,都构成社交网络。人们决策的一个重要依据其实就跟社交群体相关,周围人的决策会带动你的决策,用社交网络理论来做决策支持是一个重大方向。

用关系型数据库来存储社交数据有点吃力。我跟你打电话,“我”是一个“用户”的实体,“你”是另一个“用户”的实体,我们之间存在了“通话”的关系;“你”还可能跟“她”发生了关系。但社交网络的分析还需要关注圈子、关系紧密度…… 人们想从中找到人与人之间的关系、圈子,是不是一个家庭的,是不是一个公司的,是不是情侣关系。甚至还要去发现一个人的重要程度,是否具备某种影响力。用实体关系来表述这种社交网络需要绕些弯路转换。所以,自然出现了一种图数据库(Graph DBMS)。数据按照节点、关系和属性键值存储。开源产品Neo4j就是这类GDBMS。基本上这也是一种键值数据库,也就是说其最底层数据存储都是按照key-value存放的,这种存储方式是比较适合并行处理,适用于分析。而graph database的重要特点就是内置了常见的graph算法,它的存储结构让这类算法性能倍增。可想,未来也许会出现专为图像分析而出的数据库,专为视频分析的,等等。

数据的量越来越大,种类越来越丰富,大数据时代需要新的数据管理手段。列式、MPP的关系型数据仓库在改变着,NoSQL的CDBMS、GDBMS也试图在改变着。关系型数据库是企业IT建设时代的数据管理基石,而在Big Data时代,也许需要一种新的,正在探索中的数据管理基石。

作者简介

刘庆(网名:Q),定居合肥,BI独立顾问,兼职于Teradata,从事电信业的BI咨询服务工作,入BI一行10余年,早期研究BI架构,近些年偏重业务分析。另一身份为ttnn BI论坛创办人,写写文章,编编杂志。

[resource ]IBM 软件试用版:IBM Cognos Business Intelligence , IBM Cognos Business Intelligence Developer Edition V10.1

original:http://www.ibm.com/developerworks/cn/downloads/im/cognosbi/

试用 IBM®Cognos® BI,它通过面向服务架构(SOA)交付全面的商业智能功能,为您做出更好的决策提供所需的功能和信息。使用报告、分析、指示板和记分卡来监控业务绩效、分析趋势和度量结果。面向服务架构让 Cognos BI 的部署和管理变得非常容易。使用所有数据源构建报告、OLAP 数据集、指示板和记分卡。 为成千上万的用户提供可靠的伸缩性。模块式部署允许您满足直接需求,并根据需求进行扩展和修改。

Tab navigation

Tab navigation

下载试用 IBM Cognos Business Intelligence Developer Edition V10.1 ,该产品是试用版产品,提供 180 天的试用期。

操作系统 版本 大小 下载方法 下载
Windows® XP SP3
或更高版本
V10.1 589MB-966MB HTTP |
Download Director
下载

注意:在带宽为 1.5M/秒的网络环境中下载时,所需时间大约为 50 分钟至 1 小时 23分。我们推荐您使用 Download Director(需要支持 Java 的浏览器),它为下载大文件提供暂停-重新开始功能,支持高速传输并且能够通过防火墙访问文件。

相关说明

功能组件

此下载试用版包含以下组件:

  • IBM Cognos BI 10.1 – Query Studio, Report Studio, Analysis Studio, Event Studio
  • IBM Cognos BI Framework Manager
  • 事例 – IBM Cognos BI 的 DB2, Informix, Oracle, SQL Server 事例
  • 数据源 – Derby as Content Store, Clients for LUW DB2, DB2 on System z, CS, IDS
  • 应用服务器 – Tomcat

约束条件

下列约束条件只适用于试用版:

  • 不得用于生产环境
  • 只支持一名管理员和 5 个用户
  • 只能够运行在一台设备上
  • 要求安装 Internet Explorer 7 、Firefox 3.5 及其更高版本

Cognos 相关资源

[repost ]IBM cognus business insight Analyst reports

original:http://www-01.ibm.com/software/analytics/cognos/business-insight/library.html

Analyst reports

The Forrester Wave: Enterprise Business Intelligence Platforms, Q4 2010

To assess the state of the enterprise BI platforms market, Forrester has evaluated the strengths and weaknesses of top enterprise BI platform vendors. IBM Cognos Business Intelligence is one of the leaders.

Last updated: 12 Jan 2011

View PDF (635KB)

View all Cognos Business Insight Analyst Reports

Back to top

Data sheets

IBM Cognos Business Insight

Learn about the revolutionary BI workspace that lets business users assemble, personalize, analyze and interact with virtually all types of data in any time horizon through a dashboard-style interface.

Last updated: 12 Jan 2011

View PDF (630KB)

View all Cognos Business Insight Data Sheets

Back to top

Demos

Cognos Business Intelligence for Analysts

This demo shows how an analyst can use IBM Cognos Business Intelligence to find out why some products are underperforming in a specific market and to investigate solutions to increase revenue.

Last updated: 17 Aug 2011

View demo (SWF, 16MB)

Cognos Business Intelligence for Executives

This demo shows how an executive who is responsible for the well being of the organization can use Cognos Business Intelligence to get quick access to the pulse of the organization.

Last updated: 17 Aug 2011

View demo (SWF, 16MB)

IBM Cognos Business Insight for Self-Service Reporting

See how Business Insight can transform your business through actionable analytics and self-service reporting.

Last updated: 16 Aug 2011

View demo (SWF, 55MB)

Full Demo: IBM Cognos Business Intelligence

See how IBM Cognos Business Intelligence enables an enterprise-wide integration with reporting, plus analysis, scorecarding and dashboards – as well as planning, scenario modeling, real-time monitoring and predictive analytics.

Last updated: 25 Apr 2011

View demo (SWF)

What’s New for Executives in IBM Cognos Business Intelligence V10.1

See how to get an at-a-glance view of operational and financial performance using the latest capabilities of IBM Cognos Business Intelligence.

Last updated: 12 Jan 2011

View demo (SWF)

View all Cognos Business Insight Demos

Back to top

White papers

The Power of Analytics Maturity: Achieve better business outcomes by raising your analytics quotient

Read about the technologies in IBM’s comprehensive, unified business analytics system and how they work together to help organizations grow in analytics maturity – that is, their ability to use analytics to improve business outcomes.

Last updated: 16 Jun 2011

View PDF (700KB)

Business Intelligence for Marketing Organizations

This paper introduces you to the business intelligence solutions that can meet the needs of marketing-from CMOs to brand managers and marketing operations analysts to search engine marketing managers and campaign strategists.

Last updated: 21 Jan 2011

View PDF (680KB)

Business Intelligence for Business Users

In this paper, learn how important it is to address the different business intelligence needs of your business constituents, so they have insight when and where they need it.

Last updated: 11 Jan 2011

View PDF (832KB)

IBM Cognos Business Intelligence Query and Reporting: Gain insight and outperform

Learn how IBM Cognos Business Intelligence Query and Reporting provides you with the information you need to drive fact-based, comprehensive business decisions in this white paper.

Last updated: 11 Jan 2011

View PDF (2MB)

View all Cognos Business Insight White Papers

[repost ] BI Platforms User Survey, 2011: Customers Rate Their BI Platform Vendor Cost of Ownership

original:http://www.gartner.com/technology/media-products/reprints/oracle/article190/article190.html

Overview


Cost has become a more important factor in choosing a business intelligence (BI) platform, and survey data from BI professionals shows what BI platforms cost and what factors raise or lower these costs. These results can guide BI leaders when they make buying decisions.

Key Findings
  • Organizations are increasingly considering what they believe to be low license price alternatives solutions such as Microsoft, software-as-a-service (SaaS) and open-source vendors to augment or replace highly-priced and often incumbent platforms. However, low license price is the smallest fraction of the overall cost and does not always translate into equally low BI platform ownership costs over time.
  • The survey results show that of the three main components of the cost of a BI platform — software licenses, implementation fees and administrative costs — administration costs make up as much as 75% of the overall BI platform ownership cost and are recurring, so BI leaders need to focus considerable attention here.
  • Three factors affect the cost of software licenses: vendor pricing models, product packaging and product scalability.
  • In general, ease of use, lowers implementation costs, but also affects factors such as the complexity of analysis performed by users and breadth of product use across BI platform capabilities.
Recommendations
  • When evaluating the cost of BI platforms, extend your analysis beyond initial license fees to include implementation and administration costs.
  • For any deployment, try to maximize the number of people who use the BI platform to bring down per-user costs such as by combining projects where possible.
  • Balance any consideration of cost with functional requirements. Low-cost tools that do not meet requirements will not deliver the expected business benefits.
  • Use a similar rigor on the selection of services providers as they make up a significant component of the overall cost.
  • If you are considering moving from a high-priced vendor to one with lower overall costs, make sure you include switching costs as part of your evaluation.


Table of Contents

Analysis
Total Cost of BI Platforms From Different Types of Vendors
Megavendors, Excluding Microsoft
Microsoft
Large Independent Vendors
Data Discovery Vendors
Open-Source Providers
Small Independent Vendors
Software-as-a-Service Vendors
The Factors That Contribute to the Total Cost of BI Platforms
License Costs
Results for Individual Vendor Products
Megavendors, Excluding Microsoft
Microsoft
Data Discovery Vendors
Open-Source Tools
Large Independent Vendors
Small Independent Vendors
SaaS Vendors
Factors That Affect License Pricing
Implementation Costs
Results for Individual Vendor Products
Megavendors, Excluding Microsoft
Microsoft
Data Discovery Tools
Open-Source Tools
Large Independent Vendors
Small Independent Vendors
SaaS Vendors
Factors That Affect Implementation Costs
Administration Costs
Results for Individual Vendors
Megavendors, Excluding Microsoft
Microsoft
Data Discovery Tools
Open-Source Tools
Large Independent Vendors
Small Independent Vendors
SaaS Vendors
Factors That Affect Administration Costs
Recommendations
Appendices
Appendix 1: Factoring in Complexity of Deployment
Appendix 2: Composite Integration Score
Appendix 3: Number of Administrators Per Vendor
Appendix 4: Types of Data Discovery


List of Tables

Table 1.
Categories of Business Intelligence Platform Vendors
Table 2.
Administration Efficiency by User and Data
Table 3.
Mean Number of Administrators Per Platform
Table 4.
Data Discovery Types

List of Figures

Figure 1.
Elements of the Total BI Platform Ownership Cost
Figure 2.
Total Business Intelligence Platform Ownership Cost
Figure 3.
Total Business Intelligence Platform Ownership Cost Per User
Figure 4.
Total Cost Per User, Breadth of Use, and Business Benefits
Figure 5.
Total Cost Per User, Complexity of Deployment and Business Benefits
Figure 6.
Total License Cost Per Product
Figure 7.
Total License Cost Per User
Figure 8.
Implementation Cost Per User
Figure 9.
Ease of Use, Complexity of Deployment, Implementation Cost Per User, and Business Benefits
Figure 10.
Composite Ease of Use, Breadth of Use, and Complexity of Analysis Done By Users
Figure 11.
Ease of Use Developers, Integration, Implementation Cost Per User, and Product Quality
Figure 12.
Total Administration Cost Per Platform
Figure 13.
Total Administration Cost Per User Per Platform
Figure 14.
Efficiency of Business Intelligence Platforms Measured by Users, Data and Administrators
Figure 15.
Data Volume vs. Number of Users vs. Average Administration Cost Per User
Figure 16.
Business vs. IT Administrators and Administration Cost Per User
Figure 17.
Complexity of Deployment by Platform

Analysis

A recent Gartner survey of 1,225 BI professionals worldwide found that cost has become a bigger factor in buying decisions. From discussions with clients, Gartner finds that most BI leaders evaluate the cost of a BI platform based primarily on license fees. However, license fees represent less than one-quarter of the total cost of a BI platform; implementation costs and especially administration costs account for three quarters, whether measured absolutely or per user (see Figure 1). The results of our survey show the average license, implementation and administration costs of offerings reported by customers from the different BI platform vendors as well as how factors such as complexity of deployments, including the number of users and size of data (Appendix 1) and ease of use influence total costs. BI leaders can use these results to evaluate the costs of the particular BI platforms they are considering.

Figure 1. Elements of the Total BI Platform Ownership Cost

Figure 1.Elements of the Total BI Platform Ownership Cost

Chart represents customer perception and not Gartner’s opinion.
Total administration costs are annually recurring. They are based on the number of IT and business administration full-time equivalents reported by respondents. This number was multiplied by an average annual salary cost. This chart shows the first-year cost only.
N=718
Source: Gartner (March 2011)


Total Cost of BI Platforms From Different Types of Vendors

Survey respondents report different total costs for the BI platforms of different types of vendors. Table 1 shows how Gartner categorizes the different vendors.

Table 1. Categories of Business Intelligence Platform Vendors

Category
Vendors
Megavendors
IBM, Microsoft, Oracle, SAP
Large Independent Vendors
Information Builders, MicroStrategy, SAS
Data Discovery Vendors
Advizor, QlikTech, Tableau, Tibco Software
Open-Source Providers
Actuate (BIRT), Jaspersoft, Pentaho
Software-as-a-Service Vendors
Birst, PivotLink
Small Independent Vendors
Actuate (e.Report), arcplan, Bitam, Board International, Corda Technologies, LogiXML, Panorama Software, Quiterian, Salient Management, Targit
Source: Gartner (March 2011)

The results of this survey show that license cost is a small fraction of overall BI platform ownership cost. Those organizations looking to lower cost should look beyond license to implementation and in particular to administration costs for efficiencies. The findings below are based on Figure 2 that shows total BI platform ownership costs per cost category per product and Figure 3 which shows total BI platform ownership costs per user per product.

Figure 2. Total Business Intelligence Platform Ownership Cost

Figure 2.Total Business Intelligence Platform Ownership Cost

Chart represents customer perception and not Gartner’s opinion.
Cost is adjusted for Complexity (see Appendix 1).
Total administration costs are annually recurring. They are based on the number of IT and business administration full-time equivalents reported by respondents. This number was multiplied by an average annual salary cost. This chart shows the first-year costs only.
N=718
Source: Gartner (March 2011)

Figure 3. Total Business Intelligence Platform Ownership Cost Per User

Figure 3.Total Business Intelligence Platform Ownership Cost Per User

Chart represents customer perception and not Gartner’s opinion. Cost is adjusted for Complexity (see Appendix 1).
Total administration costs are annually recurring. They are based on the number of IT and business administration full-time equivalents reported by respondents. This number was multiplied by an average annual salary cost. This chart shows the first-year cost only.
N=718
Source: Gartner (March 2011)

A summary of total cost findings by vendor type is discussed below.


Megavendors, Excluding Microsoft

Large, complex megavendor deployments (except Microsoft) cost the most overall, but cost less than the average on a per-user basis. (The average megavendor deployment covered in the survey had 3,230 users.) Small and less complex megavendor deployments will likely realize smaller economies of scale because they will receive smaller license discounts and cannot extend support resources over as many users. A previous Gartner survey where the average deployment size was 500 users found substantially higher per-user license costs for all megavendors, except Microsoft.


Microsoft

Microsoft has below-average license costs both on an absolute and a per-user basis. However, the total cost per user exceeds that of the other megavendors due to initial implementation and administrative costs.


Large Independent Vendors

SAS has the highest per-user cost across all cost categories. SAS and MicroStrategy have higher license and implementation costs per user than Information Builders, while MicroStrategy and Information Builders have lower per-user administration costs. Survey respondents report lower per-user costs for Information Builders across all cost categories, while MicroStrategy supports the most complex types of analysis conducted by its users. As with the megavendors, large, complex deployments will be the most costly overall, but cost less on a per-user basis, while small and less complex deployments will likely realize smaller economies of scale.


Data Discovery Vendors

Data discovery vendors show higher than average cost per user because deployments support small numbers of users. These tools also require a higher proportion of administrators, many of whom are business users — a key value proposition of these types of tools. QlikTech has the lowest cost per user profile of the data discovery vendors, although Tibco Spotfire supports the most complex types of analysis including predictive analytics. Cost considerations increase as these types of tools are more broadly deployed. Therefore, it is important to focus on negotiating strategies that account for bigger discounts for larger user volumes and on deployment strategies that maximize leverage of IT and business administrators across the user base.


Open-Source Providers

In general, open source vendors have below average license costs (though licenses for Actuate’s BIRT cost as much as many commercial alternatives). But a high number of administrators per user puts the total cost of open-source deployments on par with those of commercial vendors. Jaspersoft is the only open-source vendor with a below average total cost per user.


Small Independent Vendors

Virtually all the small departmental vendors in the survey show below average total cost on an absolute basis due to smaller and less complex deployments. However, on a per-user basis, only LogiXML and Bitam fall below the survey average. This is particularly true for Quiterian and Board, which tend to have very small deployments (among the smallest in our survey). However, both Quiterian and Board also provide extended capabilities. Quiterian provides both data discovery capabilities and predictive analytics, while Board provides both BI and corporate performance management capabilities in a single package.


Software-as-a-Service Vendors

Like the small independent vendors, SaaS vendors show lower total cost in absolute terms due to smaller and less complex deployments, but small numbers of users lift per-user costs above the average. BI leaders typically believe that SaaS deployments cost less because they require less hardware and infrastructure, and provide for easier (at least in theory) product upgrades. However, based on research conducted in April 2010, Gartner has found that cumulative license costs can exceed those of on-premises alternatives after three and five years so it is important to carefully consider the cost benefits.

Figure 4. Total Cost Per User, Breadth of Use, and Business Benefits

Figure 4.Total Cost Per User, Breadth of Use, and Business Benefits

Chart represents customer perception and not Gartner’s opinion.
Cost is adjusted for complexity (see Appendix 1).
“Breadth of product use score” is the sum of user activity percentages across reporting, ad hoc analysis (all levels of complexity), dashboards, scorecards, and predictive analytics for each vendor.
The Business Benefits score is an average of scores on 10 different benefit areas scored by respondents on a scale of 1 to 7 where 1-2 = poor, 3-5 = avgerage, 6-7 = outstanding.
Orange dots represent above-average and blue dots represent below-average business benefits score.
N=718
Source: Gartner (March 2011)

Moreover, vendors whose customers report realizing above average business benefits tend to have less complex deployments (see Appendix 1). Figure 5 shows each of the vendors products’ complexity of deployment score on the horizontal axis versus its total cost per user on the vertical axis. An orange color dot indicates that respondents reported achieving above average business benefits for that product. Only MicroStrategy, Microsoft and Information Builders have both complex deployments and above average business benefits achieved. While only Microsoft and Information Builders have below average total BI platform ownership costs per user as well.

Figure 5. Total Cost Per User, Complexity of Deployment and Business Benefits

Figure 5.Total Cost Per User, Complexity of Deployment and Business Benefits

Chart represents customer perception and not Gartner’s opinion.
Cost is adjusted for Complexity (see Appendix 1).
The business benefits score is an average of the scores on 10 different benefit areas scored by respondents on a scale of 1 to 7 where 1-2 = poor, 3-5 = average, 6-7 = outstanding.
Orange dots represent above-average and blue dots represent below-average business benefits score.
N=718
Source: Gartner (March 2011)


The Factors That Contribute to the Total Cost of BI Platforms

The total cost of a BI platform has three main components: software licenses, implementation fees and administrative costs. The survey results show that of the three main components of the cost of a BI platform — software licenses, implementation fees and administrative costs — administration costs make up as much as 75% of overall BI platform ownership costs and are recurring, so BI leaders need to focus considerable attention here.


License Costs

Results for Individual Vendor Products

Licenses represent the most visible component of BI platform cost, and most enterprises (shortsightedly) focus the bulk of their vendor comparison effort here. The different pricing and packaging models of BI platform vendors make direct comparisons difficult for BI leaders evaluating products. Using our survey data, we look at total costs and average costs per user, adjusted for the complexity of deployment (Appendix 1), to create a basis for comparison across vendor products (see Figures 6 and 7).

Figure 6. Total License Cost Per Product

Figure 6.Total License Cost Per Product

Chart represents customer perception and not Gartner’s opinion.
Cost is adjusted for complexity (see Appendix 1).
N= 718
Source: Gartner (March 2011)

Figure 7. Total License Cost Per User

Figure 7.Total License Cost Per User

Chart represents customer perception and not Gartner’s opinion.
Adjusted for complexity of deployment (see Appendix 1).
N=718
Source: Gartner (March 2011)


Megavendors, Excluding Microsoft

Megavendors tend to have higher total license costs due to more complex deployments. However, all megavendors are below average in per-user costs (Microsoft is the only megavendor in the fourth, or bottom, quartile).


Microsoft

Despite complex deployments, Microsoft’s total license costs fall significantly below average, and its per-user license costs are lower than those of any other vendor in the survey.


Data Discovery Vendors

Per user, licenses for the data discovery vendors cost above average for two main reasons. First, deployments tend to be smaller so costs are spread across a small user base. Second, data discovery vendors appear to be able to charge a premium over traditional BI tools (see Appendix 4), which lack these functions. Inquiries from Gartner clients show that while the average discounts on BI platforms in 2010 was approximately 60%, discounts on data discovery tools averaged around 20%. Moreover, in absolute terms, the licenses for Tibco and QlikTech deployments cost more than those for Tableau and Advizor deployments, in part because Tibco and QlikTech have larger deployments. Moreover, a higher mix of Tibco customers deploy its analytic applications, which tend to be more expensive.


Open-Source Tools

For total license costs, all open source vendors fall below the survey average. All of these vendors tend to have smaller, less complex deployments. Jaspersoft and Pentaho fall into the bottom quartile; Actuate BIRT into the second quartile. On a per-user basis, Pentaho moves into the third quartile, and BIRT moves higher than the survey average.


Large Independent Vendors

Large independent, traditional BI platforms tend to have high total license costs mostly due to large deployments, although SAS has much smaller deployments than either MicroStrategy or Information Builders. On a per-user basis, license costs for both SAS and MicroStrategy remain above average while Information Builders falls in the bottom quartile.


Small Independent Vendors

For total license cost, the small independent vendors all fall below the survey average, mostly in the third and fourth quartile. However, per-user costs lift most of these vendors, with only LogiXML, Bitam and arcplan remaining in the bottom quartile.


SaaS Vendors

Both BI SaaS vendors in the survey, Birst and PivotLink, which both tend to have very small departmental deployments, appear to follow the dynamic of the small, independent vendors. They have below average total license costs, but these costs look higher than average when calculated per user.


Factors That Affect License Pricing

Three factors have the biggest affect on the cost of software licenses: vendor pricing models, product packaging and product scalability. BI leaders should consider these variables to evaluate the license costs of the specific products and projects they are planning.

Pricing Models: BI platform vendors offer a number of different pricing models:

  • Per named user.
  • By role.
  • Per concurrent user or session.
  • Per CPU or core.
  • Per server.
  • Subscription (such as SaaS).
  • Open-source annual maintenance fee model (a form of subscription).
  • Enterprise licensing.
  • A hybrid of the models mentioned above.

When negotiating license fees, minimize license cost and risk over time by employing best practice negotiating strategies.

Packaging: The kinds of product packages that BI platform vendors offer include:

  • Complete BI platform functions in one package.
  • Packages of analytic applications and/or packages of core BI platform functions.
  • Capabilities sold “a la carte.”

Scalability: BI platform scalability drives hardware size and CPU-based (core) license costs. The most scalable BI platforms require less hardware and maintenance effort while addressing the needs of all BI users (report consumers to power user and analysts). Assuming a well designed data model, scalability generally comprises support for:

  • Concurrent users and user types.
  • Workload or complexity of applications.
  • Size and type of data.

The cost of scaling BI platforms can be reduced by BI platform features such as:

  • In-memory and caching capabilities.
  • Multi-pass SQL.
  • Aggregate awareness.
  • Optimized SQL for different databases.
  • Function shipping to different databases.
  • Multi-threaded processing, parallel processing and load balancing.


Implementation Costs

Results for Individual Vendor Products

On average, implementation costs, such as for system integrators, equal roughly 60% of license costs, so they represent a significant factor in total cost. Ease of use translates into lower implementation costs in part because easy-to-use tools allow IT developers and other BI authors to develop BI content more quickly and it allows more business users to create their own reports and analysis, thereby saving the cost of IT specialists to design them (see Figure 8). In addition, tools that support less complex deployments tend to cost less to implement.

Figure 8. Implementation Cost Per User

Figure 8.Implementation Cost Per User

Chart represents customer perception and not Gartner’s opinion.
Adjusted for complexity of deployment (see Appendix 1).
N=718
Source: Gartner (March 2011)


Megavendors, Excluding Microsoft

All megavendors, except for Microsoft, have below average per-user implementation costs. This is likely because megavendor deployments tend to be large with significant numbers of report consumers over which to spread costs.


Microsoft

Microsoft customers report above average per user implementation costs, perhaps for these reasons:

  • The need to deploy three products — Office, SQL Server and SharePoint — to meet BI platform requirements.
  • The use of these three components for non-BI functions.
  • The need to build Analysis Services cubes, which are often part of a Microsoft deployment.
  • The use of partners for implementation.


Data Discovery Tools

Data discovery tools enable business users to develop more of their own analytic content without the assistance of the IT department. For example, Tableau developers tend to be business analysts, not IT developers while most traditional BI platforms require more specialized and often higher cost IT skills.

QlikTech and Tableau have below average implementation costs per user. Tibco and Advizor have higher Implementation costs per user in part because these vendors tend to support more complex analytic applications, including predictive analytics. QlikTech and Tableau most often support interactive dashboards and ad hoc analysis, albeit often with complex types of queries.


Open-Source Tools

Actuate BIRT and Jaspersoft fall below the survey average for implementation cost per user while Pentaho implementations cost just above average. Pentaho scores poorly on both ease of use and BI platform integration, which can affect deployment costs.


Large Independent Vendors

SAS and MicroStrategy have higher implementation costs per user than Information Builders. MicroStrategy and Information Builders tend to have much more complex deployments than SAS.


Small Independent Vendors

Implementation costs vary widely among the small independent vendors. High ease of use, particularly for developers, places LogiXML in the bottom quartile for implementation cost per user. Board, Targit and Salient have low total implementation costs, but small numbers of users lift these vendors into the first quartile for per user implementation costs.


SaaS Vendors

Birst and PivotLink have above average implementation costs per user. Like the small independent vendors, they tend to have small deployments and numbers of users over which to spread costs.


Factors That Affect Implementation Costs

In general, ease of use (for developers), in part a function of BI platform integration and BI platform developer productivity features (particularly for the full range of simple to complex types of analysis), lowers implementation costs, but ease of use (for end users) also affects factors such as the complexity of analysis performed by users and, breadth of product use across platform capabilities. Ease of use can also reduce the cost of training and change management as users adopt intuitive tools more easily. BI platforms with integrated tools rather than multiple user interfaces and tools tend to require less training, even for diverse user groups.

A number of factors drive ease of use:

  • Many data discovery tool offerings (such as those of QlikTech, Tableau and Tibco Spotfire) do not use a traditional IT modeled semantic layer. These tools provide easy-to-use capabilities for business analysts to access, mash up and manipulate data with minimal IT assistance. This approach reduces the deployment and maintenance costs associated with a semantic layer, but it increases the potential for creating personal, workgroup or departmental silos, which can cost more to manage.
  • Intuitive BI content authoring tools include a graphical user interface and design environment, and out-of-the-box objects and wizards, which reduce the coding required for all levels of analytical complexity.
  • Widely available skills make it easier and often less costly than hard to find skills to develop analytic content.

Figure 9 maps the relationship between ease of use and complexity of deployment, per user implementation cost and business benefit. The horizontal axis rates vendor platforms by ease of use, the vertical axis by complexity of deployment, with the dotted lines showing the average scores for each measure. The size of the bubbles represents implementation costs per user while the orange bubbles show platforms whose customers reported above average business benefits on our survey. A large number of platforms in the lower right-hand quadrant show low cost and above-average business benefits correlated with high ease of use and low complexity. But vendor platforms with complex deployments can also deliver good business value although they tend to be easier to use than complex deployments from other vendors that are perceived as delivering less business value. Among complex deployments, business value does not necessarily correlate with low per-user costs. Only Information Builders achieves above-average ease of use, below-average implementation costs per user, above-average complexity of deployments and above-average business benefits.

Figure 9. Ease of Use, Complexity of Deployment, Implementation Cost Per User, and Business Benefits

Figure 9.Ease of Use, Complexity of Deployment, Implementation Cost Per User, and Business Benefits

Chart represents customer perception and not Gartner’s opinion.
Adjusted for complexity of deployment (see Appendix 1).
Ease of use score is a combined measure of ease of use for end users and ease of use for developers, each scored on a scale of 1 to 7 where 1-2 = poor, 3-5 = average, 6-7 = outstanding.
The size of the bubbles represents implementation costs per user. The orange bubbles show vendors whose customers reported above-average business benefits on our survey.
N=718
Source: Gartner (March 2011)

The survey results also suggest that ease of use expands the breadth of product functions used. Figure 10 shows ease of use scores (horizontal axis) versus breath of use scores (vertical axis) with the orange color of the dots showing vendors that support above-average complexity of analysis (such as data discovery, moderate to complex ad hoc analysis and predictive analytics) by users. Enterprises tend to use BI platforms with higher scores on ease of use for a broader range of activities (for example, reporting, ad hoc analysis and dashboards) rather than for a single function. Data discovery vendors, Tibco Spotfire, Tableau and Advizor are also used for more complex types of analysis. This paradox — ease of use combined with support for complex types of analysis — gives them their momentum in the market.

Figure 10. Composite Ease of Use, Breadth of Use, and Complexity of Analysis Done By Users

Figure 10.Composite Ease of Use, Breadth of Use, and Complexity of Analysis Done By Users

The orange dots show vendors that support above-average complexity of analysis.
Chart represents customer perception and not Gartner’s opinion.
Adjusted for complexity of deployment (see Appendix 1).
Ease of use score is a combined measure of ease of use for end users and ease of use for developers, each scored on a scale of 1 to 7, where 1-2 = poor, 3-5 = average, 6-7 = outstanding.
“Breadth of product use score” is the sum of user activity percentages across reporting, ad hoc analysis (all levels of complexity), dashboards, scorecards, and predictive analytics for each vendor.
Complexity of analysis/usage is a weighted average score based on the percentage of respondents reporting use of the platform. Activities are weighted as follows: viewing static reports = 1, monitoring performance via a scorecard = 1, viewing parameterized reports = 2, doing simple ad hoc analysis = 3, interactive exploration and analysis of data = 4, doing moderately complex to complex ad hoc analysis = 5, using predictive analytics and/or data mining models = 5.
N=1,225
Source: Gartner (March 2011)

BI platforms with strong integration also tend to have low implementation costs (see Appendix 2). Figure 11 shows ease of use scores for developers (horizontal axis) versus platform integration scores (vertical axis). The size of the bubbles represent implementation cost per user. The orange color means the vendor scored above average in product quality. Vendors with above-average platform integration scores tend to also have above-average product quality scores. Moreover, vendors, such as LogiXML, Tableau, Information Builders, Bitam, QlikTech with high ease of use for developers and high integration scores tend to also have below average implementation cost per user scores. These results suggest a correlation between integration and product quality and a correlation between ease of use for developers, platform integration and implementation cost per user.

Figure 11. Ease of Use Developers, Integration, Implementation Cost Per User, and Product Quality

Figure 11.Ease of Use Developers, Integration, Implementation Cost Per User, and Product Quality

Chart represents customer perception and not Gartner’s opinion.
Adjusted for complexity of deployment (see Appendix 1).
Ease of use is scored on a scale of 1 to 7, where 1-2 = poor, 3-5 = average, 6-7 = outstanding.
Product quality is scored on a scale of 1 to 7, where 1-2 = poor, 3-5 = average, 6-7 = outstanding.
The size of the bubbles represent implementation cost per user. The orange color means the vendor scored above average in product quality.
N=718
Source: Gartner (March 2011)


Administration Costs

Results for Individual Vendors

The survey results show that administration makes up as much as 75% of overall BI platform ownership costs, and these costs are recurring, so BI leaders need to focus considerable attention here. Figure 12 and Figure 13 shows the total and per-user administration costs for individual BI platforms.

Figure 12. Total Administration Cost Per Platform

Figure 12.Total Administration Cost Per Platform

Chart represents customer perception and not Gartner’s opinion.
Adjusted for complexity of deployment (see Appendix 1).
Total administration costs are annually recurring. They are based on the number of IT and business administration full-time equivalents reported by respondents. This number was multiplied by an average annual salary cost. This chart shows the first-year cost only.
N=1,169
Source: Gartner (March 2011)

Figure 13. Total Administration Cost Per User Per Platform

Figure 13.Total Administration Cost Per User Per Platform

Chart represents customer perception and not Gartner’s opinion.
Adjusted for complexity of deployment (see Appendix 1).
Total administration costs are annually recurring. They are based on the number of IT and business administration full-time equivalents reported by respondents. This number was multiplied by an average annual salary cost. This chart shows the first-year cost only.
N=1,169
Source: Gartner (March 2011)


Megavendors, Excluding Microsoft

Megavendors generally have high overall administration costs, and all score below average in administration costs per user. These results stem in part from the enterprise administration features of traditional BI platforms for supporting large numbers of users and large deployments. Departmental BI platforms and data discovery tools that are trying to move away from their departmental heritage typically have more immature management features. Also, very large user volumes allow economies of scale for administration.


Microsoft

Microsoft customers report below average administration costs per user. Gartner inquiries and other Gartner research suggest that in addition, Microsoft BI skills are widely available.


Data Discovery Tools

Tableau and Tibco have above average total administration costs while all data discovery vendors have above average administration costs per user. This can in part be explained by the fact that in data discovery deployments, more and more end users are also administrators. For Tableau, Tibco and Advizor, a higher proportion of administrators are business users rather IT professionals — not the case for most other vendors.


Open-Source Tools

Open-source tools still require administration. The administration costs per user for open source vendors is consistent with those of commercial vendors, although Jaspersoft is below the survey average.


Large Independent Vendors

MicroStrategy has below average total administration costs while both MicroStrategy and Information Builders have below average administration costs per user. Both vendors have features in their development environments that boost productivity. SAS continues to require specialized skills and is more difficult for developers to use, according to the survey.


Small Independent Vendors

Platforms from the small independent vendors all fall below the average in total administration cost reflecting their smaller deployments, but per-user administration costs are mixed. Bitam is the only small vendor in the survey with below-average administration costs per user.


SaaS Vendors

Birst and PivotLink have among the lowest total administration costs, but they rise above the survey average for per-user costs.


Factors That Affect Administration Costs

The complexity of a deployment in terms of volume of data, the number of users and pervasiveness of the deployment across an enterprise affect administration costs, as does BI platform integration and ease of use. In Figure 14, the size of the bubbles represent the total number of IT and business administration full-time equivalents reported by respondents, while the horizontal and vertical axes represent the amount of data and number of users, respectively (see Appendix 3 for more details on administrative resources per vendor). Actuate e.reports had the largest number of users for the average deployment, Quiterian the fewest. MicroStrategy managed the largest datasets, Advizor and Birst the smallest.

Figure 14. Efficiency of Business Intelligence Platforms Measured by Users, Data and Administrators

Figure 14.Efficiency of Business Intelligence Platforms Measured by Users, Data and Administrators

Chart represents customer perception and not Gartner’s opinion.
The size of the bubbles represent the total number of IT and business administration full-time equivalents reported by respondents.
N=1,169
Source: Gartner (March 2011)

On average, megavendors require fewer administrators per 1,000 users — they support four times the users and two times the data volumes with less than half the number of administrators, compared with the average of data discovery vendors (see Table 2). However, the megavendors’ advantage has shrunk compared to 2009, when megavendors supported 11 times the number of users and five times the data volumes with less than three times the number of administrators. Thus, data discovery vendors are not only supporting larger deployments, administrative efficiency is improving. Nonetheless, data discovery vendors still score on the high end of administrators per 1,000 users compared to other vendor categories.

Table 2. Administration Efficiency by User and Data

Vendor Category
Total Number of Administrators
Average Data Size per Deployment
Average Number of Users per Deployment
Administrators per 1,000 GB
Administrators per 1,000 Users
Megavendor
34.25
1,777
3,457
19.27
9.91
Large Independent
24.56
1,687
2,070
14.56
11.87
Data Discovery
19.3
765
805
25.17
23.9
Open Source
8.9
572
582
15.45
15.18
Small Independent
9.3
313
538
29.68
17.28
SaaS
5.8
79
365
73.07
15.74
GB = gigabyte; SaaS = software as a service
Source: Gartner (March 2011)

Average administration cost per user tends to go down as data and user size increases. Figure 15 shows data volume versus number of users for each vendor. The size of the bubbles represents average administrative costs per user.

Figure 15. Data Volume vs. Number of Users vs. Average Administration Cost Per User

Figure 15.Data Volume vs. Number of Users vs. Average Administration Cost Per User

Chart represents customer perception and not Gartner’s opinion.
Cost is adjusted for complexity of deployment (see Appendix 1).
Total administration costs are annually recurring. They are based on the number of IT and business administration full-time equivalents reported by respondents. This number was multiplied by an average annual salary cost. This chart shows the first-year cost only.
The size of the bubbles represents average administrative costs per user.
N=1,169
Source: Gartner (March 2011)

In general, the proportion of business users who act as administrators does not lower per user administration costs (see Figure 16). Tableau and Tibco have a higher proportion of business administrators than IT administrators, consistent with their business user value proposition. With both Tableau and Tibco Spotfire, business users can mash up data, thereby reducing the need for IT support. By contrast, another data discovery vendor, QlikTech, relies more on IT professionals as administrators because data integration still requires scripting and therefore IT support.

Figure 16. Business vs. IT Administrators and Administration Cost Per User

Figure 16.Business vs. IT Administrators and Administration Cost Per User

Chart represents customer perception and not Gartner’s opinion.
Adjusted for complexity of deployment (see Appendix 1).
Total administration costs are annually recurring. They are based on the number of IT and business administration full-time equivalents reported by respondents. This number was multiplied by an average annual salary cost. This chart shows the first-year cost only.
The size of the bubbles represents average administrative costs per user.
N=1,169
Source: Gartner (March 2011)


Recommendations

BI leaders can use the results of the Gartner survey to guide their analysis for BI platform costs and to identify the best product for the enterprise’s particular needs:

  • When evaluating the costs of BI platforms, extend your analysis beyond initial license fees to include implementation and administration costs. By far, administration represents the most significant portion of BI platform costs. Managing them is the key to lowering total costs.
  • In addition to the three main components of BI platform costs, consider the factors that affect them, such as ease of use, complexity of deployment, and BI platform integration. For example, ease of use correlates with lower per user implementation costs and broader use of the BI platform with higher achieved business benefits.
  • For any deployment, try to maximize the number of people who use the BI platform to bring down per-user costs such as by leveraging resources projects where possible.
  • If you are considering moving from a high-priced vendor to one with lower overall costs, make sure you include switching costs as part of your evaluation.
  • Balance any consideration cost with functional requirements. Low-cost tools that do not meet requirements will not deliver the expected business benefits.


Appendices

Appendix 1: Factoring in Complexity of Deployment

Vendors have different customer mixes across large and small companies and simple to complex deployments. We account for these differences by adjusting cost numbers based on an index of deployment complexity derived from a measure of deployment complexity shown in Figure 17. We calculate the scores in Figure 17 by considering vendors’ average number of users, data volume, performance score, departmental versus global deployment, complexity of workload, and breadth of use. We then create our deployment complexity index by dividing each vendor’s complexity of deployment score by the survey average. A vendor with a high complexity of deployment (customers with larger, more complex deployments) would have an index greater than 1 whereas a vendor that primarily serves smaller, departmental deployments would have an index less than 1. Where we present costs adjusted for complexity, we divide the reported numbers by the complexity of deployment index. Adjusting for complexity moves the costs of vendors with less complex deployments upward and those of vendors with the more complex deployments downward by that factor.

Figure 17. Complexity of Deployment by Platform

Figure 17.Complexity of Deployment by Platform

N=1,225
Source: Gartner (March 2011)


Appendix 2: Composite Integration Score

How well a BI platform’s components are integrated with each other directly influences integration costs and the skills needed to develop and deploy BI products. The level of a BI platform’s integration depends on how unified the semantic layer is, how many servers must be deployed and how seamlessly integrated the BI platform is with the rest of the information technology stack and complementary BI technologies. We calculated the composite integration scores for BI platform vendors based on survey participants’ responses on a scale of 1 to 7, where a score of 1-2 = poor, 3-5 = average, 6-7 = outstanding, to the following questions:

  • The product is well integrated both within the platform itself and with complementary BI technologies.
  • The front-end tools have a consistent user interface and menus, and users can easily move authored content from one tool to the next.
  • The BI platform semantic layer is unified and fully integrated and used across BI platform tools.
  • The BI platform is integrated with complementary BI capabilities and other parts of the software stack such as data integration, search, content management, enterprise applications, collaboration, business activity monitoring and business process management.
  • The BI platform uses common security and a single administration application across components.


Appendix 3: Number of Administrators Per Vendor

Table 3 shows the actual numbers for Figure 14.

Table 3. Mean Number of Administrators Per Platform

Vendor/Product
Mean Number of Combined IT and Business Administrator FTEs
Microsoft
42.4
SAP Business Objects
37.8
IBM Cognos 8
33.4
SAP Netweaver BW
31.0
Oracle (OBIEE)
26.7
SAS
24.7
MicroStrategy
24.6
Information Builders
24.4
Tableau Software
24.1
Average
18.9
Tibco Spotfire
18.2
Actuate e.reports/e.spreadsheet
16.8
QlikTech
15.4
Board
12.0
Salient
11.5
Actuate BIRT
10.9
Arcplan
10.8
Quiterian
10.6
Corda
10.1
Panorama
9.6
Pentaho
8.5
Bitam
8.5
Targit
8.0
Jaspersoft
7.1
LogiXML
6.3
Birst
5.8
PivotLink
5.7
Advizor
5.7
FTE = full-time equivalent
Source: Gartner (March 2011)


Appendix 4: Types of Data Discovery

We refer to traditional versus data discovery platforms in this report. The table below shows the high-level distinction between the two types.

Table 4. Data Discovery Types

Market Segment
Traditional Enterprise BI Platforms
Data Discovery Platforms
Key Buyers
IT
Business
Main Sellers
Megavendors, large independents
Small, fast growing independents
Approach
Top down
Bottom up
IT modeled (semantic layer)
Business user-mapped (mashup)
Query existing repositories
Move data into dedicated repository
User Interface
Report/KPI dashboard/GRID
Visualization
Use case
Monitoring
Analysis
Deployment
Consultants
Users
BI = business intelligence; KPI = key performance indicator
Source: Gartner (March 2011)