Architecture Components - Big Data Analytics: Disruptive Technologies for Changing the Game - page 39

reducing the number of situations that would lead to delays in data collection.
The most common denominators were the system logs from all of these systems.
If we know the past failures in this cascading set of transactions, can we use
Big Data Analytics to isolate the failure conditions? IBM's Big Insight includes
a query tool that allows us to study a slice of data on a chosen dimension such
as time. I can set the time for analysis to be the 24 hours preceding a failure and
look for system log error conditions, thereby helping to isolate the combination
of error conditions that happened together. If we have located a pattern of error
conditions leading to a failure, we can use the query tool to check for all the time
slices within system failure and look for any systematic failure patterns.
There are many other use cases where this analysis can be applied. For
example, a new pricing model may lead to strong negative sentiment. A new
feature release may cause disruption in consumer use. A new competitive offer-
ing may reduce interest among shoppers. As long as we have time slices with
unstructured data representing independent variables and consumer sentiment as
a dependent variable, the data can be analyzed to discover causal chains. This is
the most powerful aspect of unstructured data analytics.
4.3 Big Data and Single View of Customer/Product
In any enterprise, there are likely to be many views of customers and products.
Most of the fragmentation comes from divergent views of customers and
products. Customer and product MDM solutions are popular ways of bridging
and bringing together a single uniied view. However, over the past decade or
two, this integration has been focused primarily on intra-organization sources
of traditional “structured” data.
Automation and data collection technologies have opened up new sources
of data from the product itself, processes supporting the customers, and third
parties. For example, the web interface offers a signiicant amount of information
that can be used for additional customer insight. Tealeaf ® , a recent IBM acquisi-
tion, specializes in improving multi-channel customer experience by analyzing
customer behavior across channels and making that information centrally avail-
able. This information about customer behavior at the web interface can now be
used for a variety of purposes. It can be used by the contact centers to improve
their response to the customers, by product management to improve products,
and by IT to improve customer touch points. 23 If the customer has logged into
the website, the customer identity is known and can be used to connect customer
behavior to the rest of the customer MDM. Product usage can also be tracked and
collected, summarized, and categorized. For example, call detail records collect