In the buzzword intensive space of analytics, it’s easy to get enamored by technology. All too often the latest and greatest “eye candy” catches the attention of both analysts and even business decision makers. The coolest new trend in technology can distract from the only way which analytics actually makes an impact: optimizing organizational decision making.
At Aviana we see clients struggle with prioritizing their investments in analytics. Many departments have interest in analytics – and deciding whose initiative is most important is challenging. We also see a significant worry that investing in tools before being assured that success is indeed possible prevents many from delving too deep into breakaway approaches, including predictive analytics.
Aviana helps clients deal with this problem. We work closely with organizational decision makers to identify the decisions which matter most in their organizations, and then rapidly evaluate how – or if — analytics can make an impact. Our rapid feasibility analysis approach provides a practical perspective in less than 30 days, reducing the risk of investing in people and technology.
The Aviana eXploratory Modeling Appraisal
The Aviana eXploratory Modeling Appraisal (XMA™) is a results oriented rapid prototyping methodology, designed for feasibility analysis of advanced analytics principles on departmental business use cases.
The endeavor proceeds through a set of highly integrated steps designed to evaluate the inherent challenges and potential of fully realizing an analytics solution.
An Aviana engagement manager works with a business or technical sponsor to identify between one and three focus areas or “ideas” of how advanced analytics could impact the organization. Once these ideas have been pinpointed, a tight collaboration between our engagement manager and analytics practitioners with data and business subject matter experts begins.
We look to gain a full understanding of the business process upon which we will be doing analytics. At the conclusion of the process review, we validate that we can in fact identity a solution target. The solution target takes the form of a decision which can be impacted using analytics.
Aviana’s expert consultants then engage in a three stage process to evaluate potential and risks.
Evaluate the Data (The Data Audit)
The first stage of the XMA™ is data evaluation. We assess data integrity including the completeness and quality of it. Our practitioners use advanced analytics to examine key issues which in our experience impact program effectiveness: data cleanliness, sparsity, outliers and extreme values, appropriate granularity, and available targets and predictors. The data audit helps us ascertain the distance from “usable in analytics programs” the client’s data is.
Often the data audit reveals a process or data generation gap. For example, an XMA™ for a large bank revealed that key information critical to loan application evaluation was captured in highly inconsistent unstructured forms – and shared in manual review sessions. Bankers learned to game the system and knew how to hide “key information” which might prevent a loan from being approved.
Attack the Problem (Flash Modeling)
Once data has been evaluated, Aviana experts familiar with the category of business problem begin to unwind the data and weave it into potential solutions. Exploratory predictive modeling on client data is done in IBM SPSS Modeler using a highly compressed and modified version of the industry standard “CRISP-DM” methodology.
Appraise the Solution (XMA™ Scoring)
The XMA™ Scoring phase of the project is when our consulting team evaluates the full potential of the endeavor, should it be moved to production. The results are descriptive and are designed to help drive ongoing discussion on program design. We provide observations on our experience during the modeling process, and considerations moving forward.
The two axis XMA™ Prioritization Matrix, assists the client in prioritizing further investment:
Potential Return of Focus Area: Based on the anticipated value of deploying the model into production, should the client anticipate a HIGH or LOW return for a fully deployed project?
Anticipated Difficulty of Production Deployment: Based on fully developing the project into a decision making system (by, for example deploying results into a BI System) should the client anticipate a HIGH or LOW difficulty of transitioning the project into production
The Aviana eXploratory Modeling Appraisal is ideal for any organization considering advanced and predictive analytics – or the potential value of a “big data strategy.” At the conclusion of the project you will have a clear picture of how – or even if – an investment in people or technology makes sense. All work done in the tools including SPSS modeler streams, charts, output files etc will be accessible to client at the completion of the project should it acquire licenses for SPSS Modeler software.
We welcome the opportunity to learn about your business challenges. Please contact us today to Request A Meeting.