Enterprise Information Management

If your wife adds (if you are the lucky one) additional 1/2 tea spoon sugar in your bed tea, can you imagine what would be the impact on your blood sugar level and to maintain the ideal balance with calorie count how many extra steps you’ll be running. Mass penetration of internet and ever increasing health consciousness in society has ensured that everyone has easy access to appropriate diet charts and general health tips such as above are addressed to some extent. However, healthcare providers around the world who are facing the pressure of patients and healthcare authorities to improve quality care with no complications, seek an analytical solution to empower them with robust tool to analyze data and regularly identify patients at risk.

From a healthcare providers’ perspective, this is addressing a business issue. From patient’s perspective, this is advanced care. Whereas from the solution provider’s angle it’s prediction of future (likelihood of a patient developing a complication) based on history data, which has gone through statistical modeling technique. To me, this complete process, starting from problem identification to the solution implementation is appropriate exemplar of adding value in society through Predictive Analytics.

Until recently, predictive analytics solutions were serving niche business domains and within that certain specific business processes, however the intensity of investments planned by the BIG IT service providers hints at another “commoditized” IT service in the pipeline. Another exciting side of the story is that the industry is looking forward to explore the benefits for daily & intra-day operational decision making. Traditionally analytics applications have supported strategic & tactical business decision making, which did not put ‘low latency’ pressure on data integration but operational decision making needs data refresh as close to real time as possible. As far as operational analytics solutions are related, debates are already on embedded and stream analytics but the real challenge would be obviously on data feed. And it would be interesting to watch the trend, how solution providers align with this demand – whether it is much advanced data integration through typical Operational Data Store (ODS)/Data warehouse/Data mart route or Analytics Solution embedded in Business Process Workflow itself.   

Posted by Piyush Kumar at 06:32 AM | Permalink | Comments (3) | TrackBacks (0)

You can notice Information Management (IM) space is quite hot these days.  There are multiple tools coming into the market to put the power of knowledge in the hands of business users.  These are also coming to the aid of IT Teams, in reducing their workloads, and enabling them to focus more on the data and information governance aspects.  But quite often we lose track of the important principles that we need to follow.  Albeit these are not quite difficult to follow, but are essential. 

One can follow a particular sequence like:

1.       The objectives have to be listed first, like: the business drivers, what should attract the attention of senior management, speed of availability, and done with lowest possible cost

2.       Then we need to understand the important business and technology hurdles.  Business hurdles are usually faced: standardisation of definitions, clarity on information / data ownerships, agreements on master data hierarchies, and achieve common understanding on business case.  Technology hurdles are like: data standardisation, data cleansing, enterprise data model building, manage metadata, and other important items like scalability, reliability and availability

3.       Since the IM programs are usually extensive, it is essential to give Business Centricity to establish quick wins, and continue the goodwill gained to further the program

4.       Look at ways and means of reducing the effort wastage, through re-use of methodologies, frameworks and documentation templates

Finally, the IM is supposed to have and provide:

1.       A pretty good base has to be built to serve as a foundation.  This can be multi-layered data bases, data marts etc., but should have proper rules controlling the basic data management

2.       To improve the standardisation, master data management (MDM) has to find a place in here.  Business needs should dictate whether this should be a centralised MDM or decentralised MDM

3.       Data privacy and information access is paramount, in addition to the information availability.  Hence, access controls have to be properly set for the different business users, based on their span of control and information life cycle exposed to them

4.       The last but one is the availability of play area(s) for business users.  This could be either in the form of departmental data marts, just sand boxes, or some play pens.  This area could be controlled in the form of persistent areas, aggregate areas, and non-persistent areas.  This should also allow business users to bring in all their tools and expertise to explore data, fit the statistical modelling, and bring out information for informed decision making

5.       To make all this work in tandem, and still provide freedom to the business users, we need better IM Governance. 

Allow me to spend some more time on the important aspects that should be considered as part of IM Governance.  First of all, IM Governance should be a body of some permanent folks, and some semi-permanent folks.  The permanent folks are the ones who are more associated with the technology aspects and the data modelling aspects.  The semi-permanent folks are the ones who are business users, who act as business analysts for the whole build.  The semi-permanent folks are constantly replaced with a new set of people, depending on which business function is being addressed at any point of time.  The permanent folks will also take care of the overall program budgeting, spend, re-usability aspects etc.  They will also take care of the complete information lifecycle, its Governance and Stewardship.  The Governance should handle Corporate Goals and Principles, define Roles and responsibilities, establish a working structure and along with it processes.  The Information stewardship focuses mostly on the: Identifying vital / sensitive data, exposing metadata models for business users, define business rules on data manipulation and sharing, data rules on transformations, cleansing, creation of applicable metrics at different management hierarchies. 

More on the IM Guidelines later....

Posted by Narasimha Rao Kopparapu at 12:58 PM | Permalink | Comments (4) | TrackBacks (0)

Advanced Analytics (AA) at best, can easily be mentioned as Interactive Decision Making, with the help of business focused subject areas.  In fact, these Advanced Analytics are the focused Decision Support Systems, that help make decisions thru established interactive methods like OLAP, Sandboxes and other Statistical Analysis. 

However, for these Advanced Analytics to be made possible, we need the have the appropriate Business Domain Knowledge and the Frameworks to quickly deploy and make them adoptable.  Hence, we may say that customers can easily leverage their existing BI installations for Advanced Analytics, as they would have the necessary cleansed, meaningful and well arranged data.

Most of the BI usage, at the most, is around reporting only.  Very few have ventured into AA, although we are seeing some such stuff in the Performance Management or CPM areas.  Deploying Advanced Analytics will be easy if the customer already have some kind of BI DW in place, as this becomes the very foundation of it.  In that case, deployment of AA in the absence of established BI environment, will be challenging in timelines and efforts.  Quicker deployment of BI environment can be achieved through deployment of ready-made BI solution frameworks, which are best addressed through mature domain BI data models. 

Some vendors are already offering AA applications, especially in the areas of Procurement Spend Analytics, Value Chain, Activity Based Costing, Fraud, Production Planning, Customer Churn, ALM, Warranty Budgeting etc.  There is another school of thought to include the real time analytics / decision making also as part of AA.

AA is very popular nowadays, but will be interesting to note the percentage of revenues, compared to the regular development & maintenance revenues.  Will be continued in my future blogs.

Posted by Narasimha Rao Kopparapu at 07:12 PM | Permalink | Comments (3) | TrackBacks (0)

Everyone is aware that BI is a Competitive differentiator.  We know that BI plays a bigger game, be it in marketing for cross / up selling opportunities, bring in collaboration through EPM / CPM Applications for Planning and Budgeting, MDM CDI for 360 degrees view on customers, dashboards for quick outlook on performance, ad hoc environment for self analysing capabilities etc., etc.

Most of the big companies have already invested in BI in a big way.  It is quite evident that big companies thrive on multiple BI tools and technologies.  Small companies have started plans to invest in BI in big way. Credit Crunch and economic downturn is causing heartburns for the CIOs who are interested in investing, but not having enough funds on hand.  BI Programs being obviously high expenditure programs suffer the most. 

Take heart that innovations are happening in BI space too.  These new innovations let you play hard in the game, even though your hands are tied on expenditure.  Some people call them as BI Mashups, others call them as BI Killer Tools.  These have different flavours too.  Some are most IT users oriented and others are business user oriented.

I don't want to talk about a new tool in the form of BI Mashups or BI Killer tools that are going to replace the existing BI / Reporting tools.  Then what is this?

Consider this: More than 70% of the BI is involved in standard reports.  Rest of the BI is spread across ad hoc reporting, analytics, OLAP, Dashboards etc.  So, savings can be brought if there is a good answer to address the 70% of the problem. 

In case of large organisations, where multiplicity of BI tools exist, there is always a need to bring the entire organisation under one umbrella of Enterprise Reporting.  This can be achieved through different means.  Either by standardising the reporting tool or by introducing a new thin umbrella on top of all these multiple reporting tools.  Standardising the reporting tool will be a huge investment, instead the second seems more suitable.  Tools like InfoFlow are making waves in this space.  These tools provide single user interface for most of the standard report users.  It can leverage the existing reports that are developed on different tools for its own reports.  It can also schedule the existing reporting tools.  This drastically reduces the need for expensive extra licenses for different reporting tools. 

But for smaller organisations, where they have just embarked on BI programs, or contemplating for the same, it is definitely a good idea to put more power in the hands of business users.  IT Team size will be small in these organisations.  Reporting / BI Applications can be quickly developed by the business users themselves using tools like QlikView, Tibco Spotfire etc.

I can see a great future for both these types of tools in the market.

Posted by Narasimha Rao Kopparapu at 03:25 PM | Permalink | Comments (1) | TrackBacks (0)

We cannot say that one type of tool / technology is superior compared to the others.  There has to be a judicious mix of technologies to meet the business drivers.  At the same time, senior management and the line management should be able to draw their attention where it is necessary at the right time.  Information Access Platform(s) should keep in mind the necessary parameters like speed, Cost and Risk Mitigation to service different kinds of information requirements of the Corporates.  Most importantly, do-more-with-less should be borne in mind always, with the proliferation of tools and technologies, where their borders have started overlapping with each other. 

Corporates feel that their best data sources are their agreements and the mail server databases, where the entire correspondence and details of their third party arrangements are available.  But at the same time they cannot let their transactional information attain any lesser importance.  Providing information, trends and insights in the transactional information is also necessary.  Business users are also interested in building small but quick applications to understand aberrations in their data, and take fast decisions.  All these things call for data parsing, data arrangement, data transformation, and data analysis techniques using Search, BI and Interactive Visualisations tools / platforms.  Sometimes, Corporates also need to access external / streaming-in data to understand the public opinion about their products / brands and services.  All these play an important role in shaping the decision making processes within corporates.

Managing all these technologies to help the decision makers to make informed decisions is important.  Most importantly, getting the users access the right information / data is very important, in view of plethora of data sources and tools available.  Information Access platforms should take care of parameters like Relationship Effectiveness, Operational Effectivenss, Risk Mitigation and Technology Efficiency in view to drive the Customer Centricity for business improvements. 

The challenges to the new information management are: "Findability" of relevant information, Contextual integration of fragmented information, Extraction of intelligence from unstructured data, leveraging and integrating traditional structured data. 

Search Platform is fast emerging as a viable Information Access platform: as it easily combines public domain data, unstructured and structured data.  It comes across as a Insight Driven Info Platform, by providing Guided Access, Unified Information Access, Knowledge Platform, and sometimes as a Virtual Data Warehouse. 

Posted by Narasimha Rao Kopparapu at 03:22 PM | Permalink | Comments (0) | TrackBacks (0)

I believe Manufacturing companies spend more than 50% of their costs on materials alone, hence materials become the single largest cost contributor.  Bringing down costs by adopting best strategies will bring in increased margins, and better market penetration.  I think VSA is more suitable for a manufacturing sector than any other.

VSA offers some of these best procurement strategies, as it focuses on the biggest cost area.  Looking at procurement function, we can think of multiple facets like maintaining relationship with suppliers, understand more on their performance: what quality of the products they are supplying, how punctual they are in their delivery timelines, what is their turn-around time from the purchase order to delivery etc. 

On the other hand, VSA also enables to understand the purchasing behaviour our Procurement executives, like, who are the preferred suppliers for some of the executives, how much amount of business we are doing with each supplier, are we availing the discounts and fully exploiting the payment terms, check on utilisation of freebies offered by suppliers in terms of goods or service, check on the long term vendor agreements signed & utilisation of the same, leverage big sources for more discounts etc.

VSA also acts as a good platform to integrate with the procurement processes with the sales forecasting, production planning, and raw material planning engines.

We can clearly see that VSA focusses on External, Internal and Integration Fronts.  The first one is to understand the relationship with the suppliers, second one is more on the purchase practices and the last one is alignment with the company goals (short term and long term).

Probably, we can categorise VSA into Tactical and Strategic pieces.  Tactical piece will be required to streamline the vendor relationships, vendor consolidation, discount maximisation.  The Strategic piece will be required to integrate the sales and production planning, and also in arriving at the Economic Order Quantities (EOQs). 

For a good implementation of VSA, we need some sort of Master Data Management to cleanse and maintain the vendor and product codes. 

Posted by Narasimha Rao Kopparapu at 03:19 PM | Permalink | Comments (0) | TrackBacks (0)

We are all familiar with Web 2.0 mashups from having some information e. g. office/hotel/customer locations being displayed in Google maps on the company’s website. (Funnily enough, this is the only example you’ll find in the mashup presentations given and I don’t want to embarrass my colleagues by coming up with a fresh one.  But truly there loads out there.)

So, a mashup in a web page blends “own data” with data coming from external sources to present a personalised view for the user. By doing this, we have magically created additional business value. This is expressed by the “mashup formula” of 1 + 1 = 4.

With this in mind, corporate people tend to start thinking “Why go the long way and build all these well defined services in the backend and have these painful corporate EAI programs, when we can just as easily mashup the information at the front end … for a tenth of the cost and the hassle?”.
Back at University we were seduced by the idea of using 4GL, where anybody could create software by simply using a graphical editor and “just” define the business logic behind everything. Yet, I still ask myself why are we so tool-prone? I certainly understand why the companies selling tools are mind blocked in that way! As we at Infosys sell the brain and not the tool, so I believe it’s neither the hammer nor the nail that builds a house but the intellect and the proficiency of the person using them.

But mashups certainly are an intriguing concept. It is not that they offer something otherwise impossible but that they offer a new approach that makes us think of innovative solutions for our business problems - just as Web 1.0 has.

Given time, there will be common standards for questions as essential to mashups as “what is a widget?”. In time, we will find a way to describe the APIs - or rather the services - that mashups provide, and how to access them. We will know how to relate to similar concepts like JSR168 or WSRP. And maybe after some time I won’t have to complain about my train travel 3rd party iPhone app that stopped working because the operator just changed the web interface.

So mashups will give us new ways to integrate and present corporate data more easily, timely and with far less effort. This will certainly enable us to be a lot more flexible and adaptive to challenges in a very dynamic world. That is if we “just” go and define feeds of information and types of applications (widgets) to be used. As if it were so easy! So my belief is that at the end of the day the challenge to enterprise information management is still the ability to define and model processes as well as business and information domains.

How about you? What kind of mashups do you imagine for your company? What is it that you would like to see? What sources of information (internal and external) would you like to use? Will you be able to mash the client locations along your business trips with the nearest golf courses or will someone ask “Where is the money”? Should I finally get a decent job outside software development or am I just mashup?

Posted by Christian Schacht at 11:37 AM | Permalink | Comments (0) | TrackBacks (0)