The discovery of Shadow IT functions is a clear indicator that business groups are trying to solve a problem taking a more (even If you hate me for saying it) streamlined approach.  A common symptom is discovering business unit driven transactions acquiring licenses for a new software package or analytic tools. Shadow IT groups have become very prevalent in large enterprises. This means two things. The good news is that there is an existing culture of innovation within the enterprise where users understand the need for speed, lack of complexity and the value of analytics. The bad news is that these users often believe the incumbent IT process and eco-system cannot meet their flexibility, price point and time to market needs. Addressing these needs often requires re-thinking the SDLC for Information  and setting data free for exploration.  Read the Capgemini whitepaper titled An Effective Architecture for Setting Data Free within the Enterprise.
Information Management executives often are challenged with balancing their production needs with their Analysis /Analytics needs and which takes priority. Many techniques including work-load prioritization, agile development etc have been tried and have failed because ideation is iteration. It means that you start off with the seed of an idea, continue to iterate as long as you see a path to a positive outcome and abandon it as soon as you see no path to a positive outcome. We see this occurring organically within shadow IT groups albeit with partial and at times incorrect or “dirty” data and often with a limited perspective and often with limited infrastructure and ability to execute.
Large enterprise organizations are bringing Shadow IT organizations into the light, supporting and super charging them with the full power of Enterprise Infrastructure that only IT can deliver. This is a major shift in the IT model.

Many times, the most effective model for Big Data Analytics is a combination of decentralized services for Business Intelligence and Stand Alone Shared Services for Analytics”- AT Kearney[i]  In this model, shadow IT teams are empowered and organized into “analytic concierges” whose function is to work very closely with line-of-business stake holders in fulfilling their analytics needs. In this model, each analytic concierge is populated with the roles that are required for the specific division or segment they are serving. Each analytic concierge includes many roles. Common roles include Data Scientists, Statisticians, Business Analysts, Data Analysts, Business Strategists / Architects, Trainers, Facilitators etc and is designed and dedicated to serve one specific division or combination of business units. The model is based on raising the affinity and effectiveness of the concierge based on its dedication to serving the needs of one business function although individual resources may be rotated across concierges to enable synergies. This model is very convergent with Capgemini’s BI Service Service model (BISC) as illustrated in the diagram below and the referenced white paper.

Fig: from Capgemini POV titled “Applying Federated Governance and Architecture to Improve Information Delivery”

To raise the effectiveness, organizations often compliment each concierge with a fulfillment group within the IT BI Service center that is aligned to their needs.  This group supports the concierge with data acquisition, data cleansing, annotation, data modeling, data transformation, data aggregation, meta data, master data, report creation, advanced statistics and a host of other technical and production functions that enable the concierge center to take their ideas to production once its value is proven in a sandbox area. 
For more go to: