Both BI and DWH on the Open Source software are set to grow, says N. K. Subramaniyam, Executive Director, Operations and Technology, Saksoft Ltd, Chennai (www.saksoft.com). SME customers are especially wary of the costs and the deployment time involved in a BI and DWH implementation using commercial software, he adds, during a recent interaction with Business Line. The alphabet soup, for starters, is about business intelligence, data warehousing, and small and medium enterprise.
Technology is a mainstay of any DWH implementation, explains Subbu, even as the early morning joggers go past us in Nageswara Rao Park. There are three recommended approaches to tool selection, viz. single vendor, best of breed, and Open Source, he outlines. Our conversation moves from the grassy lawns to the cyberspace, in the form of an email exchange…
Excerpts from the interview
Can you elaborate on the three approaches to DWH tool selection?
Single vendor: A single BI vendor provides better interoperability, and tends to be cost-effective in terms of licensing and maintenance costs. However, the customer may lose out on the emerging technologies when tied down to one vendor.
Best of breed: A best-of-breed technology platform provides the flexibility to mix and match tools that best suit requirements and budget. However, long-term costs and vendor stability may be an issue.
Open Source: Such software has a great opportunity to benefit, because more and more organisations realise the importance of BI as a key business strategy, even while striving to keep a check on the huge licensing and maintenance costs associated with commercial software.
What do you see as the key drivers behind the growth of BI and DWH in India?
The key driver for the growth of BI is the demand by decision makers for accurate, timely information. Financial information has been available even without a formal BI infrastructure; what has been missing is information on the customer and portfolio performance. BI and DWH have grown as enterprises have realised the importance of this missing information in taking business decisions.
The DWH and BI industry is growing at a fast pace and is one of the top 10 business and technology priorities. In India, BI is the next wave in the BFSI (banking, financial services and insurance) sector after core banking and branch-level automation. Financial services organisations in India are now looking at making investments in the areas of identification of the right customer segments and marketing right products to them.
Companies are looking at a complete BI solution and not just a BI tool set. Traditionally, the banking and financial services industry has been at the forefront of establishing a BI infrastructure; now, industries such as the airlines, telecom, retail, hospitality, and healthcare are also realising the importance of BI.
Are there any patterns of significance that you notice among the enterprises that adopt BI and DWH?
We have seen two types of organisations moving on to adopt BI and DWH: Gen 1 organisations that do not have a DWH, and currently therefore obtain information from isolated islands of data. Such organisations start small, and tend to look at low-cost solutions while ensuring that they derive maximum value, and are therefore willing to look at Open Source software, and also bottom-up approach in building DWH solutions.
A bottom-up approach starts with identifying what is required first from a reporting perspective and building the DWH initially with only those data elements, so that there is quick availability of the required information. A top-down approach works in the opposite manner – data across the enterprise are integrated in a DWH so that all possible MIS can be generated. While this approach ensures there are minimal data level iterations, it obviously takes longer to build.
The other organisations are the Gen 2 enterprises, which already have a DWH and are looking more at technology upgrades and other advancements such as formal information governance policies and regular data quality audits. Such organisations are also more open to innovative technologies such as data warehouse appliances.
Do you find SMEs taking to BI and DWH?
SMEs have begun to realise how BI contributes to making more informed decisions, and therefore to profitability. Mid-sized financial services organisations in India are now looking at making investments in the areas of identification of the right customer segments and marketing right products to them.
However, there is a perception that a typical BI and DWH implementation is a multi-million dollar investment and takes years to build. But with new technology and resource availability, one can build a good DWH solution in 180-360 days. The overall awareness of the information infrastructure may be lesser in a mid-sized organisation and hence user education tends to constitute a large part of the requirements analysis phase.
In what areas do large enterprises derive the maximum value from BI and DWH?
The most important area where enterprises derive maximum value is in obtaining a single source of truth – large enterprises typically have multiple systems in multiple technology platforms, where there is extensive data duplication. An enterprise-wide data warehouse provides a single integrated platform for all data across the business, standardised definitions of data and therefore a single version of the truth of the business.
A single source of all data also results in a unified customer view, which ensures that the business can track profitability by customer, create well-defined cross-sell strategies and track risk and exposure at the customer level. Another important business benefit from a unified customer view is the ability to offer relationship pricing.
Is there a potential to leverage the cloud and SaaS models for BI and DWH?
Today, the market for BI as a service is a miniscule percentage as compared to the overall BI platforms market. However, BI as a service is set to grow and will largely be driven by cost, time and accessibility factors.
In the on-demand model, the cost shifts from being a capital expense to an operational expense, and the solution can be delivered quickly and efficiently. Industries such as retail and manufacturing are more likely to move to the BI as a service model. Even among these industries, mid-sized organisations with limited IT budgets and resources which need a quick deployment of the BI platform will move to the SaaS (software as a service) model faster than the larger organisations.
While the financial services industry has been the early adopter of BI as a practice, moving to the on-demand model from the on-premise model will need a paradigm shift in the business intelligence model, as there are concerns on data security and regulatory and compliances issues associated with putting data in the cloud.
How important is information governance in the overall BI and DWH implementation?
Information governance is a key strategic activity in the overall BI and DWH implementation. It consists of a set of approved formal rules applied to the management and control of management information within an organisation for the benefit of all. These rules are applied for the effective management and control of information assets. An information asset is any management information component that adds business value like data models, data definitions, reporting tool sets, external policy and process documents and so on.
Information governance aspects must be identified during the initial BI assessment phase. This will typically consist of defining the data principles and policies, the data management process, identification of all the data assets, and nominating the data owners and data stewards.
Information governance is one of the key success factors of a DWH and BI implementation. It is not sufficient to aggregate all data in an enterprise-wide data warehouse, but ongoing data governance and quality assessment are just as important for deriving maximum business value.
Would you like to talk about the emerging technologies in the BI and DWH domain?
Historically, customers were focusing on BI tool set and tools for ETL (extract, transform and load) processing. Even after spending a lot of time and money in developing a large DWH, customers still needed to spend time in performance tuning, repeat iterative development and it took considerable amount of effort.
Today, DWH appliances are available which are pre-tuned for large DWH implementations. Customers are actively looking at appliances such as Netezza and Teradata. Currently the price of these appliances is high. Over time, prices will come down and one can see large-scale adoption.
Similarly, new tools are emerging in the BI space using in-memory processing and column-based processing. Usage of these tools is also catching up.