BeyeBLOGS | BeyeBLOGS Home | Get Your Own Blog

January 14, 2013

Business Intelligence - Two Key Perspectives

I had written a couple of posts in my employer's (Hexaware) blogging site that I thought of sharing with this community.

The first one was more of a technology view which I presented at one of the BI conferences in India. You can read the post titled 'Transitioning to a New World - An Analytical perspective' at this link.

The second post provides a business perspective and it was titled 'The Business Intelligence Chasm'. You can read it at this link.

Thanks for reading. Happy new year 2013.

Posted by Karthikeyan Sankaran at 11:45 AM | Comments (0)

September 8, 2012

Big Data and the Goldilocks Principle

I was inspired to write this post based on a TED talk given by David Christian titled 'The History of the World in 18 minutes' in which he narrates a complete history of the universe, from the Big Bang to the Internet, in a riveting 18 minutes. This is "Big History": an enlightening, wide-angle look at complexity, life and humanity, set against our slim share of the cosmic timeline. Check out his website and I promise you that this 'Big' has nothing to do with Big Data, as we know it. But what got me interested in his talk is his reference to the 'Goldilocks moment' – a moment so precisely right for certain thresholds to be reached to enable higher forms of complexity (life) in the universe.

That got me thinking – Is Big Data the 'Goldilocks moment' for organizations with respect to analytics helping them towards achieving better business outcomes?

I think the answer is 'Yes' and this stems from the following hypothesis – An organization can utilize analytics for better business outcomes if:

a) they have more data points to be analyzed (volume)
b) have the ability to perform sophisticated analysis on large and diverse datasets (variety)
c) and can do at a much faster rate than before (velocity)

So Big Data techniques when synthesized properly with structured transactional data can provide valuable insights helping organizations make better business decisions.

On the other hand, the exponential increase in processing power of CPUs, the steep fall in memory prices and high bandwidth availability, have enabled the practical use of Big Data techniques. From the human angle, people are creating digital data, viz. social media chatter, video sharing, blogs, mobility etc. at a rapid pace that organizations (with help of Big Data techniques, of course) can potentially solve the 'Innovators Dilemma' by providing new products and services that the consumers did not ask for simply because they couldn't figure out what they actually want.

All in all, I think we are at a precise moment in history (the Goldilocks moment) where organizations can greatly increase their ability to provide better products & services for their consumers using Big Data techniques.

Thanks for reading. Please do share your thoughts.

Posted by Karthikeyan Sankaran at 2:15 PM | Comments (0)

July 8, 2012

Business Focused Analytics - The Starting Point

Having been a Business Intelligence practitioner for the last 13 years, there has never been a more exciting time to practice this art, as organizations increasingly realize that a well implemented BI & Analytics system can provide great competitive advantage for them. This leads us to the question of - 'What is a well implemented BI system?' Let us follow the Q&A below.

Q: What is a well implemented BI system?
A: A well implemented BI system is one that is completely business focused.

Q: Well, that doesn't make it any easier. How can we have BI that is completely business focused?
A: BI & Analytics becomes completely business focused when they have 'business decisions' as the cornerstone of their implementation. The starting point to build / re-engineer a BI system is to identify the business decisions taken by business stakeholders in their sphere of operations. Business decisions can be operational in nature (taken on a daily basis) and/or strategic (taken more infrequently but they tend to have a longer term impact). To reiterate, the starting point for BI is to catalog the business decisions taken by business stakeholders and collect the artifacts that are currently used to take those decisions.

Q: The starting point is fine – What are the other pieces?
A: The next step is to identify the metrics and key performance indicators that support decision making. In other words, any metric identified should be unambiguously correlated to the decision taken with the help of that metric and by whom. Next we need to identify the core datasets in the organization. Please refer to my earlier blog post titled 'Thinking by Datasets' on this subject.

Q: What about the operational systems in the landscape? Aren’t they important?
A: Once we have documented the relationship between Business Decisions to Metrics to Datasets, we need to focus on the transactional applications. The key focus items are:
1) Inventory of all Transactional Applications
2) Identify the business process catered by these applications
3) Identify the datasets generated as part of each of business process
4) Next step is to drill-down into individual entities that make up each of the datasets
5) Once the Facts & Dimensions are identified from the entities, sketch out the classic 'Bus Matrix' which would form the basis for dimensional data modeling.

Q: All this is good if we are building a BI system from scratch – How about existing BI systems?
A: For existing BI applications, the above mentioned process could be carried out as a health-check on the BI landscape. The bottomline is that every single report / dashboard / any other analytical component should have traceability into the metrics shown which should then link to the decisions taken by business users. BI & Analytics exist to help organizations take better business decisions and that defines its purpose & role in an enterprise IT landscape.

The answers mentioned above provide the high-level view of my approach to Business Intelligence projects.

Thanks for reading. Please do share your thoughts.

Posted by Karthikeyan Sankaran at 7:45 AM | Comments (0)

May 6, 2012

Predictive Analytics - The Cure for Business Myopia

Myopia = Shortsightedness. Theodore Levitt published his landmark paper titled 'Marketing Myopia' in 1960 that led to a paradigm shift in how companies viewed their business models. Marketing Myopia refers to 'focusing on products rather than customers', and how such a short-sighted view is bound to eventually lead to business failure.

One reason that short sightedness is so common is that, organizations feel that they cannot accurately predict the future. While this is a legitimate concern, it is also possible to use a whole range of business prediction techniques currently available to estimate future circumstances as best as possible.

Some of the relevant techniques to predict future outcomes are given in this blog post. These techniques, though important in isolation, are much more powerful if they can be combined together for specific business scenarios.

Key Techniques to predict future business outcomes are:

1) Data Mining / Predictive Analytics: Data mining is the computer-assisted process of finding hidden patterns in data. Data mining tools predict behaviors and future trends, allowing businesses to make proactive, knowledge-driven decisions. Hence it is also called predictive analytics.

2) Text Mining: Text Mining is the process of deriving high quality information from unstructured text data. There are various techniques used to derive high quality information from textual data, such as computational linguistics, information retrieval, statistics, machine learning, etc. Various forms of text mining include categorization, classification, clustering, concept extraction, summarization, sentiment analysis, etc.

3) Complex Event Processing (CEP): CEP is used to discover information contained in multiple events happening in parallel and then analyze its impact from the macro level as "complex event" and then help take subsequent action in real time. Primarily an event processing concept that deals with the task of processing multiple events with the goal of identifying the meaningful events within the event cloud.

4) Statistical Simulations: Predicting the future involves building mathematical models that define the relationships between different classes of variables that are important for the organization. Different types of relationships, viz. Deterministic, Stochastic, Empirical, Heuristic are possible between the variables being modeled. Simulations allows business users and decision makers to execute the models with randomized inputs to ascertain the effect on output variables.

5) Business Process Simulations (BPS): BPS are a special case of simulations that deal with non-linearity. For example, in a scenario where advertising spend depends on revenue and revenue in turn depends on advertising spend (with a lag), there is no clear line between dependent and independent variables. Such non-linear scenarios are very much prevalent in business and can be modeled through specialized BPS tools like Powersim, Vensim, etc.

For BI practitioners, it is important to realize that synthesizing these techniques into the BI landscape is critical to deliver full value to their enterprises & customers.

Thanks for reading. Please do share your thoughts.

Posted by Karthikeyan Sankaran at 7:30 AM | Comments (0)

February 25, 2012

Assessing Your Business Intelligence Portfolio

Business Intelligence and Analytics has assumed many different dimensions and has a very wide scope in an organization. Mindmap showing the complete canvas of BI is available at this link. Adding to the complexity is the plethora of new trends taking place in the BI landscape, that it becomes difficult for enterprises to assess their current BI portfolio and plan for new initiatives in the future.

This blog post is aimed at providing a simple way for practitioners to identify the current scope of BI in their organizations at a high level, so that further drill-down into the details are feasible. In my opinion, there are 5 primary focus areas in BI as indicated below:

1) Data and Information Management - Data Management includes Data Integration, MDM, Data Architecture, Modeling, Databases (Row, Column, Hybrid), Appliances, Big Data, BI with SOA, Types of data (Social Media, Unstructured, Public data)etc.

2) Information Delivery - Reporting, Dashboarding, Visualization, Mobility, Cloud, Mashups, In-memory tools etc.

3) Industry Centered Packaged Applications - Packaged BI Applications that comes out of the box – This can be specific to a domain (Insurance, Airlines, Banking, etc.) or can be anchored on business processes applicable across industries, such as Financial Analytics, Human Resources, Supply Chain etc.

4) Advanced Analytics - Data Mining & Predictive Analytics, Text Analytics

5) Performance Management - Strategic Planning, Financial Planning & Budgeting, Financial Consolidation, Regulatory Reporting, Balanced Scorecard and Strategy Maps.

Each of the 5 focus could be in various stages of maturity and at a fundamental level can be divided into 3 categories - Establish (Initial stage), Enhance / Transform, Maintain / Augment.

5 (Five) Primary Focus with 3 (Three) Maturity Stages, provide us 15 cells. Each of the cells should indicate specific initiatives taken up within the organization. Though some overlap among the 15 cells is inevitable, this matrix provides a simple yet powerful way to assess your BI landscape.

We would like to hear your views and best practices in assessing your BI portfolio. Please do send your feedback. Thanks for reading.

Posted by Karthikeyan Sankaran at 12:45 PM | Comments (0)

January 1, 2012

The Adjacent Possible in Business Intelligence

It is that time of the year when everyone has the license to write about 2012 trends in their technology areas of focus. In that context, I was inspired by this wonderful book titled 'Where Good Ideas Come From' written by Steven Johnson. There are many fascinating nuggets in this book but I was completely intrigued by a concept called 'The Adjacent Possible' and would like to take about BI Trends for 2012 from that perspective.

First, let me clarify the concept by quoting sentences from the book itself.

"Start Quote" Scientist Stuart Kauffman coined the phrase 'The Adjacent Possible' to indicate a kind of shadow future, hovering on the edges of the present state of things, a map of all the ways in which the present can reinvent itself. Yet it is not an infinite space, or a totally open playing field. What the adjacent possible tells us is that at any moment the world is capable of extraordinary change, but only certain changes can happen. The strange and beautiful truth is that its boundaries grow as you explore those boundaries. Each new combination ushers new combinations into the adjacent possible."End Quote"

Steven Johnson goes on to illustrate many examples and concludes that almost all innovations, from pre-historic life to YouTube, are based on 'The Adjacent Possible'.

Coming to Business Intelligence & Analytics, here is my view of the key trends for 2012, or in other words, 'the adjacent possible' for BI, at this juncture. The Mindmap that describes these key trends can be viewed here - BI Trends for 2012 and they are:

1) Smorgasbord of Analytical workloads
2) Self-service BI
3) Social Media Analytics
4) Mobility in BI
5) Big Data
6) BI on the Cloud
7) Advanced Analytics
8) Data Visualization

Though the list is not something that will surprise BI practitioners, each area offers immense scope for innovation. Exploring the adjacent possible in each of these areas is bound to be very interesting and if you get it right can prove to be quite rewarding too. And in the true spirit of innovation, we should let ideas get connected with one another rather than protecting them.

Wish you all a very happy new year 2012. And thanks for reading.

Posted by Karthikeyan Sankaran at 3:45 AM | Comments (0)

November 26, 2011

Trim Tabs in Business Intelligence

What are Trim Tabs? - Trim tabs are small surfaces connected to the trailing edge of a larger control surface on a boat or aircraft, used to control the trim of the controls, i.e. to counteract hydro- or aero-dynamic forces and stabilize the boat or aircraft in a particular desired altitude without the need for the operator to constantly apply a control force. This is done by adjusting the angle of the tab relative to the larger surface.

As a metaphor, Trim Tabs are used to denote tiny components, which nevertheless have a great impact on things that they are attached to. This blog post is about the relevance of this metaphor to BI in enterprises.

Business Intelligence practitioners acknowledge the fact that BI & Analytics in any organization is a journey, an evolution over a period of time. The canvas for BI is extensive and spans the business technology continuum.

Given that BI can add value in many areas of an organization and there are many solutions possible in each of those areas, it is important to identify the BI 'Trim Tabs', i.e. those small areas that can provide the maximum value for invested money. In one of my earlier blogs, I had talked about the concept of 'analytics anchor points' - business processes that should be considered first for optimization through analytics.

Identifying the analytics anchor points in an organization is a non-trivial exercise. In my humble opinion, such a process should start with a complete understanding of the organization's 'Business Model'. Business Model is a set of assumptions about how an organization will perform by creating value for all the players, on whom it depends, including its customers. Like all good stories, a business model relies on the basics of character, motivation and plot. For a business, the plot revolves around how will it make money. Characters are the different stakeholders (internal & external) in that business and each one of them acting in their own self-interest (motivation) makes the plot plausible and meaningful. All successful businesses have a twist in their plot (think of Google, eBay, Dell etc.) that has made them fantastically successful. You can read more about business models in the book by Joan Magretta & Nan Stone titled "What Management Is".

From BI perspective, once the practitioner understands the business model of the company, many questions gets answered:
1) What drives the company's success and how BI can help?
2) Who are the stakeholders and what information are they looking for?
3) What needs to be optimized and how it can be done?
4) What is the architectural blueprint and how will it evolve?
5) How fast should information get delivered?
6) How much data needs to be collated and how far into the past should one go?
7) What are the regulatory requirements for the company?

And many more. In essence, clarity on one aspect of the problem (business model) will go a long way in selecting the analytics anchor points (BI Trim Tabs) for that particular organization.

Thanks for reading. Please do share your thoughts.

Posted by Karthikeyan Sankaran at 2:15 PM | Comments (0)

September 18, 2011

Wisdom of Crowds in Business Intelligence

This blog is in continuation to my previous post on Collaborative Business Intelligence as Wisdom of Crowds shares a close link with that topic.

The Wisdom of Crowds is a book written by James Surowiecki and the tag line for the title reads as "Why the many are smarter than the few and how collective wisdom shapes business, economies, societies and nations". In this book, the author argues that in any kind of problem solving / decision making scenarios, collective intelligence of a group of people always produces better decisions than an individual, even the expert, working in isolation. There are 4 factors that characterizes wise crowds - Diversity of Opinion, Independence, Decentralization and more importantly Aggregation.

Aggregating individual opinions and thoughts is a crucial factor in ensuring wisdom of crowds and such aggregation is enabled by software products that fall in the category of 'Collaborative Decision Making (CDM)' platforms and many of them have BI components embedded within them. Here are some of the products - Decision Lens, IBM Cognos, Lyza, Microsoft Sharepoint, Purus DecisionSurface, SAP Streamwork, Cogniti, Panorama Necto. Do check out these products as they offer something unique in the way Information has been delivered to business users.

Thanks for reading. Please do share your thoughts.

Posted by Karthikeyan Sankaran at 2:30 PM | Comments (0)

August 28, 2011

Collaborative Business Intelligence

According to this apocryphal story about the 'Tower of Babel', early humans developed the conceit that, by collaborating to work together, they could build a tower that could take them to heaven. Now, God, angered at this attempt to usurp his power, destroyed the tower, and then to ensure that it would never be rebuilt, he scattered the people by giving them different languages so that they cannot collaborate with each other.

To me, the above story illustrates the power of collaboration and reinforces the fact that 'Wisdom of the Crowds' always produces better decisions than one taken by individuals, even by the so called experts. But then this question arises - "If Business Intelligence is all about decision making, why are BI platforms / tools not conducive for collaborative decision making?"

In my mind, the core features of a Collaborative BI environment are:

1) BI environment should have a 'Facebook' like interface that greatly simplifies interaction between users.
2) Users should be in a position to reuse analytical components developed by others.
3) Users should have the ability to communicate real-time within the context of a particular report / dashboard / scorecard
4) Users should have the ability to subscribe to analysis done by other users
5) Users should have the facility to recommend and promote useful analytical components to others
6) BI platform should have the ability to aggregate the views of multiple users with respect to analytics, so that BI gets "crowd-sourced" within the organization.

I had this 'Aha Moment' in relation to Collaborative BI when I saw a demo of the new BI product developed by Panorama software, called Necto. Panorama, company more famous for their OLAP technology that was sold to Microsoft in 1996, has developed this platform keeping collaboration as the central theme of the product. Do check out the demo of Necto at this link.

Bottomline is that Collaborative BI platforms are here to stay. In my opinion, Collaborative BI along with its Actionable & Embedded BI counterparts form the triumvirate that enables Agility in Business Intelligence for organizations. Paraphrasing the above,

Agility in BI 'Equals' Actionable (A) 'plus' Collaborative (C) 'plus' Embedded (E) BI

I will write about the other two components (Actionable and Embedded BI) of the ACE paradigm in my subsequent posts.

Thanks for reading. Please do let me know your thoughts.

Posted by Karthikeyan Sankaran at 10:45 AM | Comments (0)

July 31, 2011

Business Intelligence and the God complex

This blog is inspired by the recent TED talk by Tim Harford titled 'Trial, error and the God complex'. You can view it here. In this talk, Tim argues that people, the so called experts, in the face of an incredibly complicated world, are nevertheless absolutely convinced that they understand the way the world works. And with that notion of all-knowing expertise (God complex), they propose solutions, which in many cases turn out to be wrong. Tim exhorts business leaders, politicians, etc. to abandon the god complex and turn to the problem-solving technique of 'Trial and Error', i.e. Variation and Selection process to identify the best solutions to complex problems.

The world of Business Intelligence & Analytics is complex too. I do realize that BI is nowhere as profound as some the things that Tim alludes to in his talk but having said that, I think the God complex has relevance to the world of analytics. As a BI practitioner for the last 12 years, I have numerous examples of product vendors, system integrators, consultants, etc. peddle their wares, under the guise of - "I know everything about your problems and this tool / solution will solve it".

Business Intelligence & Analytics, at a fundamental level, is all about optimizing business process in an organization. Given the complexity around business processes (even for a mid-size organization), I for one, feel that solutions have to be found only in that particular business context and not elsewhere. The practitioners have to abandon their BI 'God complex' and seek solutions only after they first understand the specific business situation at hand. Once the problem is understood, a systematic trial-and-error mechanism is to be instituted to select the best BI solution for their needs.

As a BI consultant, I try my best to work around the God complex, by following certain rules:

1) Listen intently to customers to understand their business model, strategy and specific problem / opportunity areas, before anything else.
2) Always develop 'Proof of concept' demos before narrowing down on tools / products / solutions.
3) Follow 'Wisdom of the Crowds' strategy (as opposed to the 'experts' view) by utilizing techniques like the Analytic Hierarchy Process (AHP) to drive consensus.
4) Create business process simulations that provides a dynamic view on the impact of analytics in any organization
5) Adopting Project Management techniques that allow for iterations / course change in the development cycle.

Though the specific techniques might vary, I feel that abandonment of the God complex (and the associated mental model) is the starting point for BI practitioners to seek solutions that are relevant in a particular business context rather than taking a 'templatized' view to solve problems.

Thanks for reading. Please do share your thoughts.

Posted by Karthikeyan Sankaran at 3:15 AM | Comments (0)