Friday, December 31, 2010

Social Media

Good blog on social media, analytics tools and marketing research.  It gives the Social Media fundamentals in these areas.

Analytics and In-Memory Databases are Changing Data Centers

For real-time analytics with large data volume this is a solution.  My first exposure to this technology was about 18 months and I was impressed with its capabilities. 

Another solution for realt-time analytics is the use of intelligent agents (think about the crawlers that are currently used but with some embedded analytics capabilities added-on to the recognition and identification algortihmns). I wrote an article about this in 2007:  http://atomai.blogspot.com/2007/03/application-of-data-mining-and.html

In-memory databases analytics and intelligent agents are not mutually exclusive.  If anything, these technologies are complementary in nature.

Thursday, December 30, 2010

Data Mining Awards 2010

Bringing Down the High Cost of Business Forecasting

Excellent and on point article by MIT Review about the decreasing costs of business analytics using cloud computing analytics tools.

Sunday, December 26, 2010

Robotic Surgery and Business Analytics

Robotic Surgey and Business Analytics
by Alberto Roldan
Copyright 2010 Alberto Roldan

Imagine a future where we can use the techniques, technologies, and methodologies of robotic surgery in business analytics.  For example:
  1. What if we can "see", diagnosed, and repair the health of a business using 3D visualizations, predictive algorithms, and statistical control processes?
  2. What if we could in detail the good parts of our business and compared it with the diseased or needed repair part of our business real-time?
  3. What if we can monitor the different components of the health of a company (i.e., financial, operational, marketing, CRM, and social media) in real time while we are making strategic and tactical decisions.
The surgical outcomes of robotic surgery is less pain and faster recovery.  We trust a complex organism as the human body to robotic surgery.  What is preventing us to use the same science, technology, and matehmatical methods in business analytics?  I believe that we are living in the future, because the future is now.

Click to the link above and tell me if you are interested.

Contact me at atomanalytics@gmail.com

Thursday, December 16, 2010

Trends in Business Analytics for 2011: Alberto's Predictions

There are three main trends that we can see occurring in 2011 across all industries and regardless of company size: (a) a large increase in volume of work, (b) utilization of social media as a key data source, and (c) challenges in implementation and innovation. First, the year 2011 will bring a tremendous increase in the volume of work in business analytics. Specifically, the area of predictive analytics had an increased momentum in the last half of 2010 that will carry to 2011 without any abatement in sight. My prediction is that companies will increase their 2011 appetite for predictive analytics by over 200% from 2010.


The need for more efficient ways to do work and adapt to swift changes in the economic environment will be feeding the craving for companies to use business analytics. Some of the accompanying challenges that this increased appetite for predictive analytics will include the following:

  1. Prioritization of analytics projects to align with corporate strategic and tactical objectives
  2. Analytic resources identification - the profiling of individuals that know predictive modeling technique
  3. Institutionalization of budgeting for analytics projects , or no more unplanned budget for analytics projects

Whether business analytics resides in the IT or business organization, companies will need to modify their business models to reflect the contributions and needs of an analytics organization. Companies need to decide whether they will bring analytics talent by acquisitions, hiring, or by outsourcing. Offshoring will be substantially different for IT than for analytics (rate cards, project management, and resources). Adjustments on delivery expectations for analytics projects may be needed since the analytics portion of a project can be a short-term exercise, but the enterprise implementation for the same project could be a long-term project.

Social media will become an even more important key data source. It will be used to reduce the time that it takes to predict trends affecting companies and their competitors. Companies that successfully leverage analytics in social media to detect future trends, and make changes to their strategy, will differentiate in the marketplace by using “swift insights” to quickly adapt to changing market conditions. The best example to understand the importance of swift insights is the experience of Coca-Cola vis-à-vis Gap. In 1985 Coca-Cola introduced a new coke formula, and it took them nearly three months to return to their original formula after a public outcry. On the other hand, Gap introduced its new logo, and the public backlash was so pronounced in the social media that one week later they returned to their original logo. In mathematical terms this mean that social media accelerated the identification of the need for change by eleven times (11x). In other words, if an average car ran at 50 miles in 1985, it will run at 550 miles in 2010.

The involvement of business personnel from companies will become essential in defining the social media analytics strategy, and for testing the results of any analytics project. This involves additional time commitments that must be managed both strategically and operationally. A combination of savvy, innovative, and experienced staff in consulting, technology, finance, and analytics skills will become critical in 2011 for companies to successfully integrate social media analytics into their business models. Any licensed driver can drive a car at 50 miles per hour, but only skilled experts can maneuver a car at 550 miles per hour.

The areas of innovation and implementation are connected by a combination of best practices and repeatable processes in best of breed companies. Enterprise analytics implementation requires a combination of a knowledge of statistics, analytical tools, and optimization techniques. Companies must guard against unwise investments in analytics by following best practices: references, due diligence, proof of value, pilot or POC, evaluation, budgeting, project planning, and implementation. Once the analytic problems have been defined and aligned with strategic goals, companies should look at their internal project planning process to ensure the availability of right resources, skills, and budget.

Planning a phased approach is recommended in all analytic analytics projects. This allows for the evaluation of the business lift, in addition it gives time for improvements and modifications that are department and geographically specific. The key for planning analytics capabilities within a company is to build a small but strong foundation of business, technical, and analytics skills and then move from small projects to larger projects.

The integration of technology innovations with analytics will be a crucial test for many companies in 2011 and beyond. There are three main technologies that will make an impact in the way we do business in 2011: mobile devices, visualization, and speech technologies. The delivery of predictive analytics results using mobile devices, like the iPad, tablets, and smart phones, allows executives and field personnel to have access to swift insights. Those insights will allow decision makers at all levels of a company to know the impact of their decisions in revenues, costs, and profitability.

The ability to use analytics and mobile devices to deliver filtered and ready to act information (converting data into information) will depend on innovative visualization techniques. The screen space in mobile devices is smaller than laptops and PCs, hence the need for smarter visualizations. The use of spreadsheets is not efficient in smart phones. For companies that have hundreds of products, the representation of multiple dimensions or variables (i.e., predicted revenues, profits, and commissions), the use of line or bar charts also have a limited use. The use of interactive 3D visualizations in mobile devices to represent analytics outcomes will become a new breakthrough in the business world. 3D visualizations and predictions are common practice in web-based games, and those algorithms will be integrated into the business world in 2011.

Speech technology is another innovation that will be making its mark in 2011 and beyond. In December 2010 two high school students won a price at the Siemens Competition by developing a speech recognition algorithm that can detect a speaker’s emotion better than any current technology. Imagine how many errors individuals make when they are in a hurry, or otherwise distracted. The business impact of preventable errors could be billions of dollars annually using a combination of this speech recognition technology and predictive analytics outcomes, all delivered through mobile devices.

Finally, my last prediction: in order to flourish and quickly adapt to changes in these rapidly changing economic times, we need to carefully listen to those that are our future. The examples from Siemens Competition (speech recognition technology), Gap (social media), and visualizations (web-based games) are common technologies used by the 9-to-30-year-old population. One of my main roles as an innovator is to listen to those voices and use my experience to provide guidance in implementing those new technologies and methodologies for businesses. A warning and advice to companies: listen carefully to those that represent our future. The future belongs to them, and our job is to provide guidance based on our experience.

My wish to companies for the year 2011: be the future not the past. Companies need to be open to new ideas and new ways to do business using analytics. Learn from the Blockbuster-Netflix proposed partnership in 2000. Blockbuster laughed Netflix out of their office thinking that the online subscribe service model would not be successful in the movie rental business. Now Netflix is a thriving business with 16 million members, while Blockbuster is in bankruptcy with $900 million in debt.

Friday, December 03, 2010

Smart-grid analytics to be $4 billion industry

I just recently completed a sucessful predictive analytics project in smart grid analytics for a utility with about 1 million smart meters. The methodology developed allow us to predict meter events; as well as power outages and spike in consumption at the meter level. See, http://atomai.blogspot.com/2010/11/smart-grid-business-analytics-adding.html

For additional details contact me.

Thursday, December 02, 2010

Applicability to Business Analytics of New Science Discoveries in Stem-Cell Biology and Astrophysics

In the last week, there were two major scientific discoveries: scientists “trick” cells to change identities; and there are three times more stars in the universe than previously thought. What are the lessons than we can learn and apply from these discoveries?

First, the methodology used by scientists for transforming a specialized cell into a different specialized cell is a “reverse and conversion “approach. In other words, molecular biologists reversed the process of a specialized cell development and turn that cell into a basic or stem-cell. Once this “reverse” step is completed, then the scientists guided the new basic cell to transform into a different specialized cell. In the area of retail and consumer goods/services companies are always coming with new strategies to increase sales and profitability by “converting” a competitor’s consumer into one of their own. In other words, the emphasis is in “conversion”, but there is no prior “reverse” methodology. Business analytics techniques can be used to identify actionable “reverse” methodologies that can be used with conversion methodologies to increase revenues and profitability. In other words, companies must have a methodology to determine how they can get a potential customer to reverse an established purchasing pattern, before attempting to “convert” that customer.

Second, the study that concluded that the number of stars in the universe had previously been undercounted and that the correct number is three times the previous estimate is a case study of applying good science to known facts. One of the reasons for the undercount was the assumption that the distributions of most galaxies are spiral like our own Milky Way galaxy. This recent study found out this to be an incorrect assumption. Moreover, the study found that the assumption that the ratio of dwarf stars to sun-like stars is 1,000 times greater in elliptical galaxies than in spiral galaxies. What is the application of these discoveries in the area of business analytics? There are two main lessons. One, assumptions of patterns, ratios, distribution, and correlations in large data sets based on the observation of a limited set (and non statistically valid random sample) could end up been not supported by reliable scientific methods. Lastly, it is imperative to have a methodology that deals with missing data before reaching conclusions and recommendations.

In summary, companies should be looking at methodologies and discoveries in the sciences and apply that knowledge into their business analytics strategy and tactics.

Wednesday, December 01, 2010

Business Analytics Project Prioritization - ©Copyright 2010 Alberto Roldan

As companies get involved in the area of predictive analytics, the issue of project prioritization has come more important. Companies need an analytics project prioritization framework that allows them to maximize resources while simultaneously reduced costs and improve profitability. Traditionally companies look at predictive analytics as an isolated function within an organization. As predictive analytics, data mining techniques, and other advanced methodologies are starting to permeate throughout an enterprise, the issue of prioritization has become a keystone in their analytics strategy.

Initially, companies looking for a prioritization framework look at their IT projects or IT reports processes in an attempt to adapt those processes to business analytics. Companies tend to find out that those processes are sufficient since an analytics project is more a combination of IT, reporting, and research and development (R&D) project.

Let me suggest the components of a framework to help companies prioritize their analytics projects:
1. Identify analytics projects prioritization committee composed of members from:
a. Executive sponsor
b. Business sponsor
c. IT lead
d. Analytics lead
e. Identify other interested parties – they may not vote but need to have input in the process of prioritization – for example project management lead, finance lead, human resources lead, and administrative support.

2. Develop standard operating procedure for meetings (agenda, minutes, decision making criteria) – meetings could be by teleconference and members of the committee can give their documented input by email

3. Follow the six M’s for the entry point for the prioritization of all analytic projects:
a. Must be Easy
b. Must be Friendly
c. Must be Accessible
d. Must be Efficient
e. Must be Documented
f. Must be Consistent

4. Identify how the project aligns with corporate and department-specific strategic goals

5. Prepare business case – the key is to quantify the business case: investment costs, quantifiable savings, changes and disruptions, soft benefits and ROI.

6. Prioritize analytics using a quantifiable framework approach in the following dimensions:
a. Strategic and Tactical initiatives
b. Costs
c. Business Impact
d. Disruption of other initiatives
e. Impact on organization financial health
f. Analytics maturity level of end-users
g. Resource availability
h. Budget availability
i. Technology availability
j. Leverage current IT environment
k. Leverage current business environment
l. High-level execution Plan
m. Exceptions
n. Process improvement suggestions

7. Document prioritization methodology so that the process is transparent

8. Notification of project prioritization with quantifiable scores and/or exceptions reasons
a. It should include the high-level execution plan
b. All stakeholders, interested parties and committee members should be notified.

9. Process improvement analysis and recommendations

Although the aforementioned process is not all inclusive, at least it gives a general overview of how to develop an efficient and documented business analytics project prioritization framework




Business Analytics

Business Analytics

About Me

My photo
See my resume at: https://docs.google.com/document/d/1-IonTpDtAgZyp3Pz5GqTJ5NjY0PhvCfJsYAfL1rX8KU/edit?hl=en_USid=1gr_s5GAMafHRjwGbDG_sTWpsl3zybGrvu12il5lRaEw