Today and in the foreseeable future, huge
waves of data-intensive applications are breaking over us, with more waves to
come.
It’s not just the
data volume, often referred to as "Big Data" or "Monster
Data," which pushes opportunities in the direction of organizations. It’s
also the demand-pull of applications, processes, and journeys growing in
importance for organizations to compete. These data sources are often measured
in terabytes or petabytes, but “being large” is just the obvious, in-your-face
description of what comprises a data-intensive application. In these apps, the
data is commonly persisted in several formats and distributed in many
locations and must be cared for in various ways for organizations to flourish.
Coping mechanisms will be described in future posts, but this post identifies
the drivers of data-intensive applications.
The Demand-Pull Drivers are Data Hungry
Because of the pressures on organizations to expand their
views on the scope and impact of applications, there is a considerable demand
for more data as focused and simplistic applications transition to intelligent
large-span applications. In addition, the speed to detect emergent signals,
events, and patterns is ever-increasing, putting pressure on follow-up
decisions and appropriate actions. It’s much like a fighter plane that has to
make decisions and take steps in seconds; however, management is used to
working in days, weeks, or months.
Moving from Dashboards to Fast Boards
Today's organizations need to anticipate critical patterns
to intercept opportunities and threats at more incredible speeds to make
decisions and take appropriate actions. Some organizations are crafting
technical sentinels that sit on the edge to sense and sometimes respond if
given the freedom to do so.
Excellence with Management Cockpits
The idea of people watching many individual
dashboards/fast boards and integrating their contexts with speed is somewhat an
unrealistic expectation. Minimally, these need to be brought together into a
management cockpit to grok the intersections of the visible measures. These
measures range from KPIs to out-of-tolerance situations. Eventually, the
management cockpit would be assisted by bots/agents to notify management of
threats and opportunities. Ideally, these management cockpits could help in a
"fly-by-wire" fashion within practiced business scenarios.
Decision Management and Assistance
Besides the speed to sense, decide and respond, data has
to be available to venture into new contexts to aid the decision-making process
and play out the ramifications of any action about to be taken. Operational
adjustments may require simple tuning or kick-off other individual efforts.
Tactical moves require new versions of rules and critical adjustments of
guardrails and constraints resulting from decisions. Strategic moves require
some forms of advanced analytics and potentially gaming alternatives through a
management cockpit.
Value Chain & Supply Chain Extensions
Today, an awakening occurs that requires knocking down
organizational/skill walls to eliminate silo thinking and actions. There is a
race to kill silos in value-chain and supply-chain situations encouraged by
businesses partnering to produce products or services. There is a premium on
innovative collaboration that crosses all kinds of boundaries to create
overarching goals and results while satisfying individual organization units at
the same time. The goals, rules, policies, and constraints need to be tweaked
simultaneously during operations in the middle of changing conditions.
Supporting Journeys
There is a considerable push to define constituent
journeys, especially customer journeys, that are often integrated with
employee/support journeys. Journeys require an outside-in perspective requiring
more data to represent specific goals of the personas and individuals
interacting with an organization. The customer experience is tracked, measured,
and recorded with sentiment data, often represented by voice interactions when
live via a representative or chatbots. The data around loyalty and satisfaction
proliferate with an outside-in journey perspective.
The Data Push Drivers Overflow
Data offers opportunities in its new forms, amounts,
locations, and captured contexts. Until now, the "Big Data'' headline has
been driven mostly by the volume story. That is about to change with the new
data types and formats that are entering the organization. In addition, there
is a new generation of the distributed types and a movement that says that
views can be constructed no matter where the data resides at the moment. While
location complicates data quality and compatibility issues, there are
alternative ways to cope with new tolerances for perfection depending on the
usage described in the demand-pull section above.
Voice Data
Voice is a key new tributary to tap for organizations,
thus tempting leverage for competitive advantage, particularly in servicing
processes/applications. Voice can be leveraged to see how often competitors are
mentioned in calls. Voice can also be analyzed for emotional reactions in the
context of the servicing experience. It is often helpful in unscripted
situations, which often occur with service representatives who are skill
specialized. Now it’s not just NPS scores that count for customer satisfaction
measurements.
Image Data
Images can be helpful when brought up in context. Not only
can physical plant layouts and machinery be checked for safety purposes through
image analysis, but broken machinery can also be detected before a significant
or cascading problem occurs. Not only can out-of-bounds situations be seen, but
optimal real-time planning can be enhanced by image detection. Real-time image
help is a must for some jobs.
Video Data
Videos can be leveraged for better productivity, such as
optimizing worker movement for better quality and faster processing. Video can
be used to identify resources in action, such as people and machinery, for
various kinds of operational optimization and training opportunities. Imagine
showing an inexperienced worker a video on how to service a particular
component such as a pump after taking a video of it in operation provides
self-paced, on-demand training - without asking experienced workers for help.
Edge Computing Data
Data can be detected at the edge before it hits mainstream
processes/applications. An unexpected event can trigger the notification of an
emergent set of conditions or patterns for real-time decisions or actions in
most cases. With the proper freedom levels, goals, and guardrails, this reporting creates a bevy of data for each node at the edge, whether a sentinel
or an actor.
Distributed Meta-data
The data about the data is often called meta-data. With
the sophistication of data storage and state combinations, meta-data is
important to properly manage the distributed data sources. In addition to the
physical state of data, such as its location (on-premise or cloud storage),
its state of meaning, transformation, source, and context can be managed with
meta-data.
Net; Net:
There are so many trends contributing to data-intensive
applications that there will likely be another new one later today. New ways to
manage traditional and new coping mechanisms will emerge over the next few
years. I'm betting that both the demand-pull and the data push drivers will
only accelerate. We are in the middle of a massive revolution for managing data
differently. Be aware and get ready - data-intensive applications are coming
for your organization.
