Our Approach & Business Analytics Methodology
Business Analytics Methodology (BAM) is WiseWithData’s proven methodology for getting the most value from analytical projects.
Stage 1 – Business Problem Definition
This step is often skimmed over by analytics professionals, but is critical to success of any analytics project. WiseWithData will work with your organization to come up with a clear definition of the business problem you are facing. During working sessions, we break down the areas of most concern by your business into their most simple expressions. By doing this, we can simplify how we fit analytics into the business problem, ensuring successful outcomes.
Stage 2 – Strategy Definition
Once WiseWithData has a clear idea of the business problems you are facing, we develop a strategy to attack the highest value target that can be solved within a small well defined scope. We plan out time-lines and what data, documentation, people and technical resources will be required to achieve success.
Stage 3 – Data Management
Data Profiling, Quality and Cleansing
Many analytics projects have less than optimal outcomes, largely because too little time was spent assessing and dealing with data quality issues. Data is the foundation of all analytics projects; as such you want it to be of the highest quality possible. At WiseWithData, our core focus is on ensuring the data is meaningful, accurate, and in the right format for the right use case.
Data Integration and Transformation (ETL)
Looking in data silos, you miss out on the interactions between different parts of your organization, leading to inefficiencies and lost opportunities. By bringing together disparate data sources from across your organization, you are able to see a complete picture. At WiseWithData, our Data Integration experts have many years of experience at joining data from disparate systems to solve business problems.
Analytical Data Warehouse
Depending on the nature of the project, is is often recommended that the target for data cleansing and integration activities is a highly structured data repository known as a analytical data warehouse (ADW). There is a robust and proven industry standard methodology for defining the structure of the warehouse, called Dimensional Modelling or star-schema approach. Many analytics projects do not leverage this best practice, leading to complex and inflexible designs, lengthy implementations and cost overruns.
Stage 4 – Data Discovery Visualization and Reporting (BI)
Data discovery and BI are core to most analytical solutions, as they are generally the key interface through which analytics is consumed. Modern BI and data visualization tools are intended to be self-serve in nature, and only require experts during deployment, configuration and administration. There are a wide variety of technologies that are built to address this need. If you have existing BI / Data visualization tools, WiseWithData can integrate outputs from work-streams into that tool. If you don’t have an existing tool, WiseWithData can provide assistance in choosing the right tool for your organization.
Stage 5 – Advanced Analytics
Analytical Base Tables
Performing advanced analytics requires that the data be transformed into a very specific format. Analytical Base Tables (ABT’s) are a fully de-normalized flat table structure containing many columns of data at a specific granularity. These tables can have 1,000’s of columns and millions or billions of rows. It is very important that the data transformations to arrive at the ABT be as transparent as possible, so that the model development teams can fully understand the data attribute meanings. Our resources have many years of experience developing ABT’s for a variety of model development projects.
Analytical Model Development
The development of analytical models is intensive and requires very specialized resources with many years of experience. There are many pitfalls in the process, and just having a theoretical background in statistics is often not enough. What is required is a deep understanding of business problem, the data, the underlying theory behind the techniques, and a wealth of practical experience. At WiseWithData, our resources have been mentored by some of the best applied statisticians in the world.
Having an analytical model means nothing if the insights it provides are not actionable. WiseWithData works with you to provide the right integration points with operational systems in order to maximize the value of the deployed models.
Integrated Decisioning and the Net Present Value framework
Highly evolved analytical organizations often use a framework for integrating multiple dimensions of a business problem together. By integrating multiple dimensions along with leveraging knowledge of the business impacts associated to individual analytical model scores, you can understand the long term impact a decision has to your organization. The Net Present Value (NPV) framework incorporates both short and long term impacts to your organization and weighs them according to a defined mathematical formula.
A prime use for the NPV framework is with organizations that have both credit risk and marketing departments. The risk department often wants to limit exposure to high risk customers, while the marketing department often wants to go after those same customers because they are highly responsive. By integrating the decision points, the marketing department can go after the most profitable customers, while taking into account the negative implications of poor credit quality. The NPV framework necessitates collaboration between different parts of the organization and breaking down operational silos, which is often very difficult to do. Organizations which have implemented this analytical framework however, are more successful and teams are often much happier because deep structural conflicts within the organization have been eliminated. If your organization already uses analytics extensively, this may be your best next step on the analytics journey.