Data Discovery, Data Mapping and Data Warehousing in the Cloud

Start with the Report, not the Data Format

One of the key reasons DMA managed analytics deliver such meaningful insights is because they are based on business behavior, not data formats and types. In contrast to traditional enterprise data integration approaches, where the first priority often is to identify all enterprise data and design a data warehouse schema to house it, the DMA methodology starts at the output – the business reports required to drive business performance.

The Solution Drives the Data Analysis, Mapping & Integration

Based on the client’s needs, Deloitte Managed Analytics identifies the locations within your source data systems where the required data resides. In many cases, we employ prebuilt data extractors or work with your IT team to create extraction scripts to push encrypted data to our secure FTP site. For the bulk data load, we use a broad-swath extraction approach to help ensure inclusion of required and contextual data. This methodology also helps eliminate delays due to multiple trips back into the data sources to find additional data elements.

Once the initial data arrives, it is analyzed, pattern-matched and mapped to the rich business behavior model in our DMA platform. We call our data integration approach Model-Driven Data Integration™ (MDDI™) because it recognizes your data by comparing it to our Intelligent Data Schema (IDS), a robust, predefined data model optimized for business analytics. Relationships in our DMA platform enable us to work backward to find more concepts and business behaviors in your data to complete the array of mappings.

With a large percentage of the target data mapped, we use your solution requirements to configure the analytics engine which in turn drives the visualization tools. These same specifications also help reduce data volumes -- since we know which data elements are needed in your solution, we use a subset of the MDDI mappings and we populate only the required data elements.

Agile Data Integration: Business Relevance Guides Quality Assurance

Typically, in just weeks from solution design, you and your business users can be performing QA on your data in your solution's reports. Each QA cycle will add more user features – starting with the core reports, then adding advanced analytics, and then dashboards and key performance indicators (KPIs). Quality issues are identified and fed back into the MDDI and incorrect or incomplete mappings, or business rules are rapidly revised.

This rapid-cycle, or agile Integration/QA process provides both the context and content which makes gaining user confidence and user adoption much more likely. It is not uncommon for our customers to begin to identify business opportunities within the QA process – prior to “go live.”

What if you cleansed all the data, but nobody cared?

Data integration challenges such as data cleansing and Master Data Reconciliation (MDR) are where it is possible for BI systems get lost in complexity and indecision. Without the focusing effect of a targeted business solution, too much effort, cost and time are expended cleansing data and resolving master data conflicts for data that is simply never used. “Clean data” is not an absolute concept -- it must be whether the data adds insight to a specific business solution.

DMA analytics use business-relevance as a guiding compass. Data is cleansed only if there are relevant values in it that are required for the solution. Our agile QA cycle is designed to identify these issues precisely and naturally side-step irrelevant data cleansing operations. Similarly, we perform master-data reconciliation only at the levels of detail which matter for the solution. Data integration is completed faster and the solution deployed sooner because no time is spent cleansing and reconciling unnecessary data.