Everyone is talking about a data explosion:

• RFID information in Retail environments
• Social media interactions via wireless
• Behavioral event data via eCommerce
• etc

All of this is leading toward the era of “big-data”.  Most would say that the “big-data” era is already upon us.  Some would say that future data loads will dwarf current requirements just as current numbers dwarf the past 5–10 years.

However, the key to the “big-data” era will not be in the simple accumulation of data in business intelligence and data warehousing (BI/DW) environments, but the utilization of that data across the organization.

Deep Analytics on Big Data

This week The Data Warehousing Institute (TDWI) held its initial solution summit on the topic of “big-data” in San Diego: Deep Analytics for Big Data.  It was a gather of decision makers and leading vendors to discuss the topic of “big-data” and the future of analytics associated with those “big-data” BI/DW environments.

Chief among the discussion topics were how to make the correction decisions on:

• Building “big-data” environments in a greenfield environment
• Transitioning from existing BI/DW environments to support “big-data”
• Hybrids to support existing datasets and the “new”, larger requirements

Solutions.  Not just Problems.

During the solution summit, customer implementation case studies were provided by vendors that highlighted the issues with “big data” BI/DW engagements.

• Tableau showed the need to provide data visualization on “big data” in a timely manner
• Teradata demonstrated how the value of the analysis was at least, if not more, important than the size of a “big-data” environment
• Aster Data highlighted the speed at which “big-data” analytics were required.
• HP talked about the new automated action requirements from “big-data”  
• Kognitio detailed how their customers are finding new revenue streams from “big-data” analytics.

Telecom Take

Telecom organizations are on the front line of “big-data” analytics.  Wireless voice, SMS and IP-base product data are at the core of the “new” business models for both carriers and organizations looking to capitalize on new business models…

Think about it… All the information that Google uses for their targeted web-based advertising transits a telecom network at some point.  iTunes would not be possible without the networks to pass content to either the tablet or smartphone.  Carriers need to use their knowledge of the network events linked with customer information to either get a leg up on companies like Google and Apple.

Yet, carriers should exercise good judgment with all that “big-data”.  Privacy laws associated with customer information are only going to become more stringent in the future as uses like telematics and location-based services take hold.

How is your telecom organization tackling “big-data”? Reactively or with strategy?

Post your comments below or email (John.Myers@BlueBuffaloGroup.com) / twitter (@BlueBuffaloGrp) me directly.

In a recent keynote address, Marc Demarest talked about the need for increased decisioning associated with “big data”.  Whether it be complex event processing (CEP) or streaming analytics, the ability to make timely decisions on the analysis of “big data” sets is limited when you place a human element in the critical path.  Not only will bottlenecks occur, but more than likely the data will move so fast that no decisions will be made.

Rules to Live By

Being able to automate the decisions that need to be taken from “big data” analysis is going to be the key for many industries.  Business intelligence/data warehouse (BI/DW) professionals will not be able to place alarms or workflows before a human analyst or operational team.  This will come from the fact that 8am-6pm time windows will not be sufficient for the decisions that need to be made and the fact that not all human interactions will know what to do or the extent of what needs to be accomplished.

Steve Pratt of HP’s Business Intelligence Solutions group presented at the recent TDWI Deep Analytics for Big Data Solution Summit for the need to automate decisions in the healthcare industry.  His reasoning was that the complexity of healthcare decisions and the need to make the timely decisions at the point of interaction was key to improving the quality of healthcare and the reducing the cost.  Improved quality would come from leveraging standard practices and by providing the proper care “further up the food chain” and thus eliminating rework later on.

Exemplifying these concepts a common situation in pharmacies.  With each prescription that a pharmacist fills, analysis and decisions need to be made about standard medical practices (ie drug interactions) and standard business practices (ie payments, deductibles).  By automating this analysis at the point of sale, healthcare can become safer and less expensive. 

However, this does not come without cost.  The ability to institute these decisions is not as simple as implementing a database trigger.  The implementation is more than an “if-then” statement.  Often the automated analysis and related decision is more than a single action path.  Issues of this complexity take more time than other types of analysis.  Below is Pratt’s analysis of where automated decisions fit on a time to value graph.

Telecom Take

For telecom organizations, the stakes are not same as they would be in healthcare.  However, automated decisions on “big data” have similar needs.  Both telecom costs and revenues will be impacted.

Network health in the future will have greater importance as the implementation of enterprise level service level agreements (SLA) move toward consumer relationships.  Both fiber-to–the home (FTTH) and wireless connectivity may soon have up-time/connectivity obligations as more and more aspects of daily life depend on IP-based connectivity.  Imagine the issues associated with FTTH downtime on IPTV products during special events like the Super Bowl.

Product pricing and availability are already moving toward speeds that humans have hard time comprehending.  In some African nations, pre-paid wireless revenue models are moving toward a per-tower pricing structure.  Imagine attempting to control, or even worse validate, call pricing at the tower level using primarily human based decision controls.

Which decisions in your telecom environment are managed with automated decisions?

Post your comments below or email (John.Myers@BlueBuffaloGroup.com) / twitter (@BlueBuffaloGrp) me directly.

While many existing revenue streams can be augmented by “big data” business intelligence and data warehousing (BI/DW) environments; the analysis of “big data” also means that new sources of revenue and new business models can be implemented.  This can come from:

• Previous business models that were unattainable from past technologies
• New business models that unimagined before the existence of “big data”
• Hybrid models relying on analytical models fed by new “big data” sources across industry domains or industries themselves

Creating New Revenue

These new “big data” enabled business models require that both data acquisition and data analysis happen within the ability of an organization to capitalize on the opportunity. 

Analytical database engines, like Kognitio, use massively parallel processing (MPP) processing to enable these “speed of business” capitalizations.  Kognitio in particular uses the power of in-database processing and analysis to allow for the efficient loading of “big data” datasets into their WX2 platform.  This allows for BI/DW professionals to avoid some of the ETL related, “big data” issues (ie if you can’t load it effectively, you cannot analyze it in a timely fashion).

Telecom Take

New revenue streams are going to be important for telecom organizations.  Revenue pressures are coming from the declining per unit value of voice, and now data, products.  Substitute usage of lower, or zero, value voice alternatives has been ongoing for years.  Yet now, price pressures are starting to work their “magic” on data products. New providers and 3rd and 4th place carriers look to attract new customers and erode the pricing for Internet connectivity.

Being able to exploit new revenue streams provided by the analysis of the data resident in telecom networks will be one of the keys for telecom carrier operations in the short-term and long-term.  BI/DW professional in telecoms will be pressed to provide the analytical environment to enable these business models.

Which new/non-traditional telecom business models are your telecom organization linking to existing network/customer data? Which are linked to external/non-telecom data?

Post your comments below or email (John.Myers@BlueBuffaloGroup.com) / twitter (@BlueBuffaloGrp) me directly.

Everyone knows this analogy:

“Find a needle in the hack stack"

With “big data” situations in business intelligence and data warehousing (BI/DW) environments, finding the needles becomes a much more impressive issue:

“Find needles AND the like pieces of hay and how do they relate to each other in the stack”

This comes evolution of predictive analytics and the quest not just for the goal or object (ie valuable customer, influencer, churn prospect), but also for the behavioral events that lead up to those goals or objects.

Impact of Social Media

With the growth of the ability to capture the event data associated with behavior based data sources of telecom call data and social media interaction (ie twitter, facebook, etc); “big data” BI/DW environments are not going away.  They are going to expand and BI/DW professionals are going to need to find a way to perform analytics on these data sets in an effective manner. 

New analytical database providers, like Aster Data, have moved to harness power of massively parallel processing (MPP) power and MapReduce processing to attain the type of relationship processing that the explosion of behavior based data sources have spurred. 

Telecom Take

As mentioned above, telecoms are some great sources of this new behavioral event data.  Call Detail Records of all types (xDRs) will be the basis of future a telecom organization’s ability to avoid becoming a dumb pipe.  The relationships in that behavioral event data will provide either the edge that telecoms need to make addressable advertising and location based services business cases a reality; and to fight off the challenges that Over the Top (OTT) content providers like Skype, Netflix and iTunes are bringing to the traditional telecom business model. 

Are relationship analytics associated with “big data” data sources currently being used in your telecom organization?  If not, why?

Post your comments below or email (John.Myers@BlueBuffaloGroup.com) / twitter (@BlueBuffaloGrp) me directly.

Many times in the world of data warehousing and business intelligence (BI/DW) professionals a distinction is made as to the size of a BI/DW environment.  This can be an effective measure since it can be argued that the more data you have the more analysis that can be performed.  However, more often than not it can be a misleading measure.  For example, no one would doubt the power of the 500+ horsepower Corvette Z06.  Yet, if you lived in downtown London, the size of the Corvette’s engine would be wasted.  A Prius or an Oyster Card would be a much better investment.

Big Data Size vs Big Data Value

In “big data” environments, you do not have the choice of Corvette or Oyster Card.  “Big data” requires the power of the Corvette.  However, BI/DW professionals need to do more with those “big data” environments than just collect the data.  They need to provide the correct level of “big value” from those environments.

Teradata has taken just such an approach with their “petabyte club”.  Each of the members of the “petabyte club” has “big data” environments.  But the “honor” of the club resides in the ability to provide value to the “petabyte club” organizations and not just a “notch” on the size totem pole.  

Telecom Take

The Teradata “petabyte club” includes organizations like AT&T Mobility and Verizon Wireless were the value of the “big data” environment is not limited to a Teradata press release.  Each is using the environment to provide a key aspect to the business.  Verizon Wireless looks a single view of the customer. AT&T is using it for customer profitability.  Both of these represent significant portions of competitive advantage strategies.  Being able to show this contribution to corporate positioning is the key for a telecom’s BI/DW environment.

Does your telecom have a “big data” environment? Does it receive “big value” from it?

Post your comments below or email (John.Myers@BlueBuffaloGroup.com) / twitter (@BlueBuffaloGrp) me directly.