Extending Boundaries with Collaborative Business Intelligence - Page 2

Sundar Balasubramaniam & Paul Nannetti

Integrated Business Process Analytics

In the earlier example of ATM withdrawal, business process analytics, besides monitoring events in isolation, also determine security situations based on a series of and/or repetition of events. The security monitoring solution should be able to correlate such events and raise appropriate alerts. The correlations can be defined in terms of

  • Event origin
  • Event volume
  • Event type
  • Business process
  • Time

The data points listed above are readily available but the ability to process them in a particular way to address a business concern is not. The ability to use pre-defined correlation processes or a self-correcting algorithm to analyze the data and initiate appropriate action, has become essential now that electronic consumer activity has increased phenomenally. Organizations are increasingly required to anticipate bad events and deploy systems that will prevent them.

In a recent survey, CIOs have identified Business Process Improvement as their number-one concern. And the number-one technology initiative that they plan to undertake is Business Information Management (BIM). While the link between the two is not apparent, when you look at it closely, in order for business process improvements to take place, you first need a business process monitoring solution. By analyzing information gathered from business process monitoring, one can develop innovative solutions that provide a quantum leap in process improvement. This is achieved best by applying the science of process analytics and integrating the results of such analytics into the processes.

Business Process Analytics-An Example

Extending business process analytics and integrating it with operations will provide tremendous business benefits. One such example is in the retail industry where stock-outs are a $10 billon problem (estimates vary between $6 billion to $12 billion). While most BI programs have provided information on lost opportunities, retailers are still struggling to prevent such losses. In our view, the solution has three pieces that are essential to addressing this problem.

The first piece is to gather demand data at the store level and extrapolate or forecast that demand into the future. Predictive algorithms can take into consideration geography, demography, regional consumer behavior and preferences, seasonality, weather conditions, historical buying patterns and immediate off-take from the store to determine short-term demand (next 3-6 weeks) and medium-term demand (8-12 weeks).

The second piece is to provide a composite and integrated view of demand and supply. Demand information, as provided by predictive analytics and including POS data, invoices, shipments, etc., needs to be integrated with supply information such as store inventory, DC inventory, In-transits, ASN and vendor inventory. This provides immediate visibility of potential gaps at an aggregate level and possible sources of supply.

The third piece of our solution is optimization. Knowing what the demand is likely to look like 3-6 weeks out, we need to find the best way to address this projected demand, based on different supply choices. There could be many ways to address this short-term demand and some of the options are:

  • Move stock from a nearby store where it is in surplus of projected demand
  • Advance planned replenishments to that store
  • Re-route planned replenishments from areas of surplus to areas in demand
  • Create new routings and/or replenishments from the DC
  • Move from vendor warehouse direct to store (DSD or drop shipments)
  • Create new work orders (WIP) for increased demand and release for production

Each of these options comes with a cost and using optimization routines one can make trade-offs to identify the optimal solution to address the projected demand. These are the demand fulfillment choices that need to be presented to key decision makers. Where standardized, they can be integrated into the appropriate application and automated for fulfillment.

Achieving the above solution would require bringing multiple technology pieces together, such as:

  • Real-time data integration platform to source POS data and other demand information and integrate it with supply chain information.
  • Means of making sense of unstructured data from social networking sites to understand interest and propensity.
  • RFID and/or GPS technologies to track inventory and shipments.
  • Statistical tools to develop and run statistical models to predict demand.
  •  Optimization algorithms and tools to determine optimized fulfillment choices.
  • BI presentation tool to present the findings to the user and initiate action.
  • SOA/EAI solution to integrate the findings into the operational application.

Thus solving a particular business problem involves the orchestration and synchronization of multiple tools and technologies. A clear problem definition and a solution that is driven by the business problem (not by a tool or technology) are critical. A solid solution architecture, with appropriate tools and technologies that can be seamlessly integrated, is essential. In order to fully exploit the power of BI, one has to deploy additional technologies and the knowledge of different physical sciences. Such a collaborative approach to BI will ensure that problems are anticipated and successfully addressed, delivering greater benefits to businesses and to the public at large.

Add Comment      Leave a comment on this blog post

Post a comment





(Maximum characters: 1200). You have 1200 characters left.



Subscribe to our Newsletters

Sign up now and get the best business technology insights direct to your inbox.