You are currently browsing the monthly archive for August 2011.

We cite collaboration as one of five key technology influences on the business intelligence (BI) market, and I get many questions about collaboration and BI from end users and vendors alike. The rise of social media websites such as Facebook and Twitter has raised awareness of collaborative platforms and created a critical mass of participants, which is a necessary ingredient for successful collaboration. However, I have to point out that consumer-oriented social media tools do not provide all the necessary components for collaborative BI. 

There’s a big difference between analyzing social media data and collaborative BI. Social media provides a rich new source of data that many organizations found difficult or impossible to capture in the past. Analyzing various aspects of social media interactions can help an organization better understand its customers and how its products are faring in the market. Text analytics and sentiment analysis are useful techniques to analyze the actual content of the social media interactions. More traditional structured data analysis such as number of followers and number of mentions can be useful also. I refer to these aspects as social media analytics. 

On the other hand, collaborative BI uses collaborative processes and technology to support and enhance an organization’s business intelligence activities. This concept has been around for some time. I worked on products 15 years ago that used workflow, annotations and email to create a collaborative environment. On a large scale, however, that never took off. These techniques persist today in software products, but they are not pervasive in the BI landscape. The rise in social media may change that, impacting BI in the same way that mobile technology has crept into BI infrastructure from the consumer side. 

In fact, it already seems to be doing so. I was surprised when twice in one day I heard two large enterprises, Boeing and eBay, describe their uses of collaborative BI. Both occurred at the Teradata third-party influencers meeting. Boeing presenters told of consolidating a number of systems using Teradata and IBM Cognos 10. As they migrated from a number of disparate systems to a consolidated system, they used collaboration to gather input from users to validate that the new system was accurate and operating properly. The watchful eyes of the end users spotted issues that slipped through the formal testing process; this is a simple yet excellent example of using collaboration to enhance BI. 

Oliver Ratzesberger of eBay shared a more complicated story that is the best enterprise instance of collaborative BI I have seen. Ratzesberger has spoken previously at Teradata events about eBay’s analytics as a service. The company employs a private cloud approach for its analytical data marts. Any authorized user can request a data mart with up to 250GB of storage by filling out a simple Web-based form. Within minutes of approval, the data mart is not only available, but it is populated with the appropriate data so the user(s) can begin working with it immediately. The collaborative aspect, referred to as eBay DataHub, allows users to publish any analysis from MicroStrategy or Tableau BI applications with the click of a button. Users can comment on these, follow each other, follow different topics, join analytical groups and participate in discussions – think LinkedIn for analytics. DataHub provides a centralized place to create and monitor all analytical activity a user might be involved in.  

I’ve seen other aspects of collaborative BI previously. For instance, I commented on the collaborative aspects of IBM Cognos 10 which support collaborative discussion of decision-making processes and data lineage. I also like the way’s Chatter can be used to embed analytic dashboard objects directly into a chat stream. However, both of these examples were vendor-driven demonstrations. No longer just an ivory tower exercise or a twinkle in a software developer’s eye, the examples of eBay and Boeing show that collaboration can provide value to enterprise BI. 

Consumerization has impacted mobile BI by bringing more mobile devices into the enterprise. Similarly, social media is impacting BI by bringing tools as well as knowledge of collaboration into the enterprise. We’ve tried for years to create a collaborative environment for BI using notes, annotations, workflow and email as a poor man’s substitute. With tools like Facebook and Twitter in the hands of consumers and with tools like Chatter and Tibco tibbr facilitating enterprise collaboration, perhaps we’re on the verge of a breakthrough in collaborative BI.  

Ratzesberger mentioned that eBay is considering making the code behind DataHub open source. I hope it chooses that route because it would provide a great starting point for others to begin to think about or embrace collaborative BI. 


David Menninger – VP & Research Director

In preparation for conducting a benchmark research study on predictive analytics, I’ve been speaking with vendors in this market segment to gather background. One of them is Opera Solutions, which I was not familiar with despite having worked in predictive analytics for more than 10 years. I was surprised to learn the company claims to be generating $100 million in revenue annually. Founded in 2004, Opera provides predictive analytics as a service, employing a staff of approximately 150 data scientists along with hundreds of other employees. It claims this is the largest private group of scientists outside IBM. (I’ve written previously about IBM’s significant efforts in the predictive analytics market.) 

I mention that for a reason. Opera Solutions is positioned at the intersection of several interesting industry trends: the rise of predictive analytics, the rise of big data, the rise in cloud computing – and a trend I will explore in our upcoming predictive analytics benchmark: the lack of skilled resources to work on the others. Academic institutions are recognizing the shortage of analytical skills also and are adding degree programs to address this need. 

Software as a service (SaaS) implies some level of staff to develop, maintain and operate systems that companies run for clients, but the Opera Solutions model goes beyond that. Its data scientists provide domain expertise lacking in many organizations to create and apply predictive models to specific business activities. Once the models are created, Opera continues to operate them and refine them on an ongoing basis. This services component caused me to question whether Opera is really just a consulting service, but it appears to have developed a reusable library of software integral to its cloud-based offering.   

Opera focuses on developing applications in conjunction with a partner or customer, then generalizing those applications for broader use. It focuses primarily on six markets: global investments, credit and risk, marketing, supply chain, custom applications and data transformation. It currently has a portfolio of nine applications on the market and another six in development. This concept of jointly developing applications has advantages and disadvantages. If you believe you have some unique way of doing business or analyzing data, you may not want to share that with others. On the other hand, if one of Opera’s existing products fits your needs, it may help you get into production more quickly and more cost-effectively than developing from scratch. If you do need to start from scratch and will allow Opera to allay some of the cost of developing your application by selling it to others with whom you do not compete, you should be able to reduce your costs.  

The applications are a mix of Opera’s proprietary components and best-of-breed commercial or open source components. For instance, to manage large-scale data Opera works with standard relational databases, in-memory databases and Hadoop. Opera hosts the analytic applications in an Amazon cloud configuration, but they can be installed on-premises if the customer prefers. Generally each product pulls data together from standard data sources and processes the data through Opera’s analytic stack to produce insights or actions. These insights or actions feed into decision-making processes via APIs and interfaces to existing applications. These applications may provide automated decision-making or present data to individuals to use in their daily activities. Opera Solutions then monitors what has happened as a result of these decisions and continuously repeats the process. Often this process can be labor-intensive, but Opera has found ways to automate many of the ongoing operations with its software components. 

Much of the company’s intellectual property relates to its signal libraries, which help determine what is “noise” and what is a good predictor. In addition Opera has invested in visualization capabilities and machine learning. However, the company does not want to be in the software infrastructure market. It sees its value in providing the combination of services and software. Given the underdeveloped state of the predictive analytics market and an apparent lack of expertise, that’s not necessarily a bad place to be. Opera can ensure the success of its customers by staying involved with their projects. It’s also realistic to recognize the role that skilled individuals play in predictive analytics projects today and use a cadre of such people to generate business. 

Opera Solutions takes an interesting and unique approach. Given an apparent dearth of predictive analytics talent in enterprises and the potential value of those analytics, it could be a way to bring more forward-looking analysis into your organization. You will need to be comfortable with the model of sharing the resulting intellectual property with others, but if that isn’t an issue, this could be a viable option.    


David Menninger – VP & Research Director

Follow on

Enter your email address to follow this blog and receive notifications of new posts by email.

Join 22 other followers

RSS David Menninger’s Analyst Perspective’s at Ventana Research

  • An error has occurred; the feed is probably down. Try again later.

David Menninger – Twitter

Top Rated

Blog Stats

  • 46,527 hits
%d bloggers like this: