Go slow to go fast

Keeping the CFO, business users, and IT organisations smiling while building a high value business intelligence environment. By Dennis Samuel, area vice president SEA and India, Teradata division of NCR Corporation.

As businesses accumulate massive amounts of data, experts have continuously lauded the concept of enterprise business intelligence (BI). Having rapid access to a constantly growing understanding of your customers, operations, and risk environment - complete with analytical capabilities that provide answers to the most pressing business questions - is surely an enticing vision. Under pressure to reduce operating costs, mitigate risks, respond to regulatory demands, and improve profits, banks are especially attracted to this vision.

Potholes and stumbling blocks
The challenge has been to get from "here" to "there", as the many potential stumbling blocks can make this a difficult journey.

Some have been tripped up by trying to do too much at once. They want all of their enterprise data, available now, for any of a range of potential inquiries. That''s a huge investment of time and resources, and these companies have struggled to simultaneously absorb costs, gain buy-in, and deliver incremental value.

Others have been too cautious - invested too little and moved too slowly. Consequently, there is so little value produced that business users end up believing they can accomplish more by maintaining their own "private" silos of data for their very specific business problems. There is virtually no cross-organisational buy-in for an enterprise BI environment.

Another common mistake is that IT organisations focus on chasing the data down, rather than developing a detailed understanding of how business users will put the data to work and produce real business value. Again here, business users, in their disappointment, wind up more dependent than ever on disparate, "home-grown" BI solutions to meet their evolving business needs. And when the business side ignores their efforts, the IT organisation feels under appreciated and underused.

These are not the scenarios the many proponents of enterprise BI have in mind when they spin their visions. In contrast, a few companies have carefully mapped out and travelled along a common-sense route that avoids the potholes and wrong turns and leads systematically to a powerful and relevant BI environment.

Communication and common sense
From both the successes and the flawed implementations some best practices have emerged for building a BI environment. These practices are grounded in two core principles.

1. There must be a strong, ongoing dialogue, with a common language among business users and the IT organisation.
Too often business users and the IT people who construct the BI environment live in different worlds, barely aware of each other''s needs or processes. Either there are too few avenues of communication or they simply don''t speak the same "language" and so struggle to make their needs known to each other.

2. There must be a step-by-step approach to creating the BI environment.
To avoid doing too much or too little, companies should focus first on addressing one high priority project, the success of which will support user adoption and whet an appetite for the next project. Typically, these first projects need to fit one of two criteria: they can garner a rapid return on investment or they can have a significant "political" value in the organisation. Either way, the principle is the same: build on success.

Three steps to a true BI environment
These two principles underlie a three-step process that ensures maximum user adoption and maximum value.

Step One: Senior management, in discussions with the IT organisation, develops hypotheses regarding the greatest value drivers within an organisation where BI can play a role.
The idea here is that the BI construction process must be rooted in the most urgent strategic concerns of the organisation. In banking, this might mean the need to cut customer attrition by 25 per cent, or to grow the home equity unit by 10 per cent and cut home equity and first mortgage default rates by 10 per cent.

Once senior management identifies the business drivers, they can then identify the key business users, anyone from marketing managers and data analysts to customer service representatives and relationship managers.

Step Two: Clarify and expand the hypotheses through sessions with business users. In the next step, senior management and the IT organisation meet or have a series of meetings with business users to test and refine the original hypotheses.

The meetings focus first on understanding the business users'' unique goals and objectives. For example, business users might say that in order to improve the mortgage business they need a picture of all current mortgage holders who are not in default, but who also have additional deposit and loan products. Perhaps credit card delinquency and a checking account in which they have just discontinued their direct deposit, for example, might signal a potential mortgage loan default.

From there, the business users and IT group can use business questions to ferret out the current capabilities that can be used to provide that information. What data does the organisation already have? What data does it need to answer that business question? How quickly must the information be available to be effective? Are the applications in place to make it happen? This process allows both groups to understand the existing information and application gaps and decide if they have efficient ways to close those gaps. It also helps both groups understand each other''s primary concerns and capabilities and focuses dialog on the ultimate use and value of newly gained information.

Finally, both groups can use these meetings as a platform from which they can expand the original idea to other improvement opportunities. IT might say, for example, that they can also provide more detailed customer history and customer profile information in near real time. Would that add value to how mortgage group uses the information about a discontinued direct deposit? Would the fact that the customer has never previously been delinquent on their credit card payments be of use? Through this dialogue, the original concept gains in value and, better still, the IT and business sides engage in a joint process that has a shared goal. This process - and recognition of each other''s issues - can help bridge the "language" gap.

Step Three: Construct and test the model
Once the needs and current capabilities are clear and the project defined, IT must take the business-based guidelines and construct the BI planning and analysis environment for this particular project. The first task is to link the business questions to the required data and ultimate sources from which the data originates. Companies should organise their data within an appropriate logical data model, one that mimics the way a particular industry works, not just the current organisational structure. This will ensure a focus that quickly facilitates the transformation of data into valuable business information.

From there, the system can link the business questions (eg which mortgage holders have discontinued a direct deposit?) to an array of specific business improvement opportunities (decrease defaults, improve retention of high-value deposit and/or mortgage holders, increase home equity cross sale), which are linked to the original overarching goals and objectives that senior management laid out (grow the home equity business by 10 percent and cut home equity and first mortgage defaults by the same amount).

Where appropriate, companies can model the anticipated business impact of this effort, analysing the combination of revenue growth, cost reduction, and / or risk mitigation it is likely to produce. This will help prioritise contending business opportunities to ensure more effective resource deployment.

But banks must weigh constructing such a model against its potential to slow the overall project. In lieu of a modelling process, some banks might roll out initial pilots in select markets and test results, before extending this idea nationwide. Focus by both IT and business on an ultimate roadmap for integrated, enterprise data will help reduce the likelihood that one-off pilots don''t become "productionised," thereby increasing the cumulative enterprise cost of unproductive data movement and data redundancy within silos.

A Foundation to Build Upon
A number of thought-leading organisations have successfully used this three-step approach, which works for three simple reasons:

• The entire process is grounded in strategic business goals.
• It facilitates an ongoing dialogue between IT and business users.
• It greatly enhances organizational buy-in.

The benefits are, of course, considerable. First, the dialogue enables projects to grow holistically beyond the original thinking to tap the full potential of BI. Second, by building on successes, banks create a constantly refined "template" that allows for continuous improvement. And, third, banks increase enthusiasm for continuously growing the BI environment by clearly demonstrating that business user requirements are being effectively addressed in a timely manner.

In turn, banks can realise the vision of a fully functioning and relevant BI environment.