Data and Analytics – An Iterative Process

Step by Step

Many people out there believe that a data and analytics program can be developed with a single purchase, hire, or action. Like installing Microsoft Excel or some other “plug and play” software, these individuals believe their data and analytics program will be installed and voilà – a data and analytics program exists.

Let me be very clear: data, analytics, and business intelligence is not a “plug and play” model. To develop a truly functioning and sustainable data and analytics practice, your organization must take an iterative approach towards building out your BI program.

The Dangers of Going “All In” at Once

Some organizations believe that they should “go all in” on business intelligence and build out everything at once. This is an expensive, time-consuming, and resource-intensive process that is more than likely doomed to fall short of its long term potential.

Why?

Most financial institutions have at least a half dozen high priority data sources they would need to draw data from to ensure high analytic value. For example, a credit union may need data from their core, CRM system, personal LOS, mortgage LOS, digital banking platform, and card servicing vendor. These six applications hold massive amounts of valuable information. Integrating these data sources in a data warehouse/data mart model all at once would require a significant up-front capital investment and would require multiple FTEs to be nearly completed devoted to ensuring those projects are developed accurately. From development to testing and validation to delivering the proper analytics, reports, and dashboards to business users, the amount of work needed to be completed at once would be overwhelming.

Let’s say you did take this route and build out everything at once – what happens if your business users don’t buy into the business intelligence program? With so much change at once, your business users might be overwhelmed and user adoption would be minimal. This also assumes your organization has the necessary data skills in-house. How quickly can you train your staff to use data properly or build the necessary reports and models to make use of the data you now have? With such a significant up-front investment required in this model, your yielded value would be but a fraction of its potential and your ROI would be minimal if existent at all.

Let’s talk about the right way to go about it.

The Iterative Data, Analytics, and Business Intelligence Process

The iterative data and analytics development model is without a doubt the most effective strategy. Start by focusing on your most valuable source of information; in the credit union example, this is most likely the core banking platform. Develop the necessary analytic structures for that most valuable application first. This includes developing a data warehouse/data mart to structure the data properly, creating a reporting infrastructure, developing some type of BI portal, training internal staff on how to effectively use the data and analytics applications, and beginning to deliver incremental value to those business users from that single application. Then, once a certain level of development has been reached, your organization should start to focus on bringing another high value data source. Building off of the work completed when developing data and analytics from the first application, integrating this second data source will benefit from the work completed previously. Users will already start to see the value that data and analytics provides to them and the organization providing increase user support throughout the development process.

The iterative strategy allows you to be more agile and flexible as you go through the development lifecycle. Learning the challenges faced in creating data and analytic infrastructure for the first data source allows you to be more efficient when integrating a second data source into the business intelligence program. As your organization continues to develop the program, the time and effort required to integrate each subsequent data source is reduced and the value yielded from each additional data source grows exponentially.

How long will it take for a data and analytics program to be built?

As part of our business intelligence strategy engagements, we typically recommend an 18 to 36 month roadmap to get your data and analytics program to an effective and sustainable level. This allows for your organization to develop the necessary infrastructure, adopt a data-driven culture, and train and nurture data skills in an adequate time frame. Budgetary and resource restraints are always a potential limiting factor, but we have found that the 18 to 36 month timeframe is sufficient and effective.

So What is the Point?

The one line summary of this whole article is this: data, analytics, and business intelligence programs need to be developed step-by-step – iteratively – over the course of 18 to 36 months to reduce adoption and development risk, minimize unnecessary up-front capital expenditures, and maximize long-term, sustainable value.

Whether you develop your program internally or use external consultants to build out your data and analytics vision, taking an iterative approach is critical for your organization’s business intelligence future.

Posted in Uncategorized and tagged , , , .