Q. What distinguishes an organization that is successful with business intelligence (BI)?
A. Successful BI implementations are not just about tools, but are about people and processes. Focusing too soon on tool selection is not going to provide the best result. A BI project is not a one-shot thing, but is a journey that takes time and patience. Data is extremely valuable, but hard to manage. Companies that want to reap the value of data need to treat it as a critical asset and resource and adopt a systematic, rigorous approach to managing it.
To succeed, BI projects need a strong leader who is knowledgeable about both technology and business and can straddle both worlds, translating between the two. Since the ultimate goal is to achieve significant business value, it’s usually better to have a technically-oriented business person lead the team.
Q. What steps should an organization take to establish a BI program or strategy?
A. The first step is to outline the expected business outcome and to make sure that there is buy-in around the ultimate goal. This is much more important than technology, which is a secondary matter at this early stage.
The second step is to focus on the data. Most companies, especially larger companies, do not have agreed definitions for very basic concepts like “customer” or “sales”. It’s vitally important to standardize on definitions of key terms before proceeding with more detailed planning.
The third step is to select technology and implement it. It’s important here to understand in detail the information requirements of different constituencies in your organization. One size doesn’t fit all when it comes to data analytics tools.
Fourth, it’s critical to get a quick win. Find a project of significant value to the business and deliver it quickly. Once the business gains confidence in the technical team, it will eagerly invest in additional projects. With momentum, the technical team can then lay the foundation for an enterprise-wide program.
Q. What trends are driving the technology landscape today?
A. The first major driver of change is the fact that there are now new deployment options – primarily cloud and mobile. Cloud deployment options are particularly attractive to small and medium-size companies and departments that do not have access to IT resources. Large companies are sometimes reluctant to invest in the cloud because of regulatory and privacy issues. However, once companies move their operational applications to the cloud, analytical applications soon follow.
Many companies are deploying BI applications on mobile devices to meet the needs of executives and managers who travel a lot as well as regional sales people or traveling technicians and operations managers who manage a store or facility and are constantly walking around supervising activity. Mobile makes a lot of sense for workers on the go.
The second major shift is the increasing importance of non-traditional data sources. Unstructured data like server logs, machine data, text documents like email, images and other forms of non-traditional data are becoming much more important and are beginning to have a profound effect on the technology required to report against this data.
Another major shift is the importance of self-service. Business users increasingly want access to data analysis that goes beyond a series of canned reports built by the IT department. They want to be able to do ad-hoc analysis, querying multiple data sets to build their own easily shareable reports
The final upheaval in the BI world has been the emergence of powerful, but very easy to use data visualization tools like Tableau and QlikView built using much more modern technology like in-memory data processing, in conjunction with a powerful drag-and-drop user interface. These tools are much easier and cheaper to get started with and can produce real business value with no IT involvement at all.
Q. Are these bottom-up visualization and data discovery tools going to ultimately replace top-down full-stack BI tools, as some are predicting?
A. These newer technologies don’t do everything, but their biggest customers are pulling them in the direction of offering comprehensive suites. Today, they use modern technologies to offer faster, better, cheaper BI tools than established players. However, as they move from serving individual analysts to departments to enterprises, they will inevitably become bigger, slower, and much more expensive. This will make them vulnerable to the next generation of BI tools that use the latest technologies and approaches. This is the inevitable software lifecycle.
Q. The emergence of NoSQL databases like Hadoop to handle large volumes of unstructured data has generated a lot of buzz. Will Hadoop generate a whole new category of BI tools, or will the current toolsets adapt?
A. It’s really too soon to say, as this is an emerging area. We have already seen the emergence of two very strong Hadoop-specific vendors: Datameer and Platfora. Both of these new vendors have achieved rapid success which indicates that a whole new category is being developed. However, the big data movement has certainly caught the attention of the traditional BI vendors and I would not be surprised to see them enter this space. We shall see.
Q. What is your view of vertical specific BI, like Sales PRISM for sales pipeline data, or Adaptive Insights for corporate performance data?
A. There is certainly real value in focusing on a specific domain area. Data is ultimately all about the business, and the more focused and nuanced you can be, the better you will be able to return value to the business. Packaged applications focused on specific vertical markets or niches certainly make a lot of sense, but they are not easy to build and support: They require a considerable amount of domain expertise and continued investment to keep up with changes in the operational tools on which they are based. There is a whole host of small companies that has done exactly that; these companies have differentiated themselves by supplying solutions specific to a certain industry or operational technology. A very good example of this is Entrinsik which has developed a solution –Entrinsik Informer – crafted to meet the needs of institutions of higher education, and other vertical markets, still using so-called “value-pair” databases.