Big-Data, IoT and IoE initiatives are mostly about analytics and real-time decision enablement
According to Gartner 72% of companies have one form or another
of a Big-data Initiative underway. According to the same source fewer than 30%
of these initiatives would deliver any true business benefits or meet business
expectations. So while investments are high the BVA (Business Value Attainment)
factor score continues to remain low. Unless you think a ROI of under 30% is
acceptable. This is concern 1
for most companies commencing their digital journey.
We started with an IoT events way back at Palo Alto, CA back in Nov 2014 and realized that we were facing an
audience that queried us on whether what we presented was just ‘smoke and
mirrors’. We followed these clarifications with another event in April 2015
with a series of IoT demo’s and suddenly the competitive positioning
opportunities started to click into place when most attendees realized that IoT
and IoE were a real paradigm shift happening all around. On April 6th,2016
we help our ‘Road-2-HANA- The digital Enterprise’
event in New York which this was a resounding success. By the way you
can still register and get access to the post event presentation slides and
videos.
In most ways, Internet of
Things (IoT) is mainly about collecting enterprise digital dust from
currently connected, semi-connected and dumb data generators. Only when we
start to collect this digital dust is when we realize that analytics are like
any other analytics.
In this digital enterprise journey the first step is digitization, the second
is analyzing the history of these data generators and the third is extracting
decision capable analytics via patterns and algorithms from this digital dust. The
Big-Data, IoT, IoE analytics mostly conform to the same rules and design
decisions as we have been using in traditional
business
intelligence (BI) and advanced
analytics. Even so, the IoT is creating unparalleled information
management and analytics challenges. This is concern 2 on
route to establishing our digital enterprise- the fact that Gartner reported in
2012 that current BI initiatives are delivering less than 40% business benefits
from thei BI projects. My validation, and interpretation, of this
statement is by validating whether this statement is true of false. ‘Move than 60% of the reports in your BI production
system are not being used, or will never be used, by your business users’. I have been asking this question to BI managers since 2011 and in
every honest reply I slowly get a ‘..probably more
than that.’
During our event in New York
we briefly touched on the evolutionary steps for a current enterprise to the launch
itself into a digital enterprise. We also discussed the difference between IoT,
which is like a digital hub and step 1, and IoE which is the end goal of
digital connectivity and digital dust generation.
Why
is IoE and IoE important?
According to Garner by 2020 50% of major competitive enterprises
will have some element of IoT installed and will be capturing essential
predictive data? The message is CLEAR – ‘DISRUPT or be DIRUPTED’ the choice now
is yours.
According to our research by 2020 30% of companies will change
their business models, i.e. from being sellers of air compressors to simply
selling Air-as-a-Service by leveraging predictive maintenance and
guaranteeing constant air supply to users without any incumbency of breakdown
or old equipment.
In order to meet these requirements business stakeholders will
need to firstly define the business needs and the business cases. Start with
the business needs so you don’t get lost in the data deluge. Once you have identified that, we do it with a
business focused ‘Design Thinking’ workshop that extracts unique business needs
that are otherwise impossible to spot, identify or extract. This becomes the baseline
for the digital transformation. This is where we sieve the coal from the
diamonds. This is but the beginning.
We simultaneously re-tool our users from Reports to Analytics
and then from Analytics to Informatics. We need to retool them from reading reports
to real-time decision enhancement. We also need to make them aware of
Analytics, Predictive Analytics and Prescriptive analytics which is already
taking out all the human errors and replacing them with pattern based
algorithms.
Concern 3: We can
no longer afford to create technocratic reports and analytics with a 30% ROI.
With such large data volumes, we need to keep our data engines surgically clean
and effective for real-time decisions. We can no longer
apply the same rules, the same architecture and decisions that have go us here.
We need to change our partners, architecture, Infrastructure, methodology to
optimize our digital experience.
So what makes the digital and IoT data
different?
As our data volumes, velocity, variety and veracity increases algebraically
we need to fine-tune our analytics engines like a formula-1 racing car. I
published this in a book in 2010 ‘BI Valuenomics’ see end of this post for
details, which is more relevant today than when it was published 4 years ago.
The changing lens of 360° Decision Enablement
Attribute
|
Legacy
|
Current - Digital
|
Data Sources
|
Inside the
Enterprise Firewall
|
Upstream, Downstream
and Midstream
|
Data Volume
|
Low
|
Very High
|
Data Extraction
|
Periodic
|
Streaming
|
Data Sources
|
Few System with
static Data
|
Data in Motion,
streaming from sensors
|
Extraction
|
Periodic ETL
|
ELT, store,
blend and manage massive data
|
Value Chain
|
Midstream
|
Downstream and
Upstream
|
Data Type
|
90% Structured
enterprise
|
Growing unstructured
streaming data
|
Information
|
Reports &
Dashboards
|
Real-time
Decision metrics and alerts
|
Complexity
|
Low
|
Medium to very
high
|
Analytics
|
Centralized
|
Distributed to
the edge
|
Transformations
|
Manual and
static
|
Pattern based
automatic algorithm’s
|
Point of
transform
|
ETL to ELT
(transform at DW)
|
TEL (Transform
at Source) then extract
|
Communication
|
One way to DW
|
Bi-Directional
in Real-Time
|
What
got us here will not get you there
Remember your current architecture and design capabilities are
running at 30% efficiency, i.e. 70% of your data is either redundant or worse
dead data. So continue with your current paradigms and your decision system will enter a dark place
we don’t want to even discuss here. So Plan your work and only then work your
plan.
What will your
Digital Enterprise require.
Better
Infrastructure Partners: Your legacy partners have been experts in the
midstream enterprise requirements, i.e. HW, servers and storage. Your future
needs are going to be driven by Big-Data, which resides at the edge, by IoT
which requires networking and communications, by IoE which will require
connected products and then interconnected devices, and most important of all a
dynamic and scalable architecture that does not lock you into long term boxed
contracts- like an appliance for SAP HANA.
Better
Connectivity: According to Michael Porter the connected
enterprise is changing the very definition of an industry and thus
your enterprise. So connections, sensors and networking are going to be the
keys to success over the next decade.
Better
Real-Time Decisions: It will not matter how many millions of
dollars you spend. It will not matter how much technology one buys. It does not
matter how big-4 your partners are. It will not matter if your project delivers
6,000 or 8,000 reports or 3,000 ODS’s and Cubes. The only thing that matters is
how much the end results impacts your day-to-day operations and the enterprise
decision capabilities.
What your need
to plan for
More
Business Participation: I stated this in 2010 and I state it today. Any
Si partner that requires your business stakeholders to be kept outside the
doors of your Big-Data, Cloud, IoT, BI or SAP HANA projects must be shown the
door very rapidly. The only path to success is to get business involved in a ‘design
Thinking’ workshop before you start your project.
More
Data: Your data is not only going to be high volume, velocity, variety
and voracity but it will get more and more unstructured. Your designers will
need to play with non-repetitive data streams tht will come streaming and the
challenge is going to be where to keep your data in motion.
More Complexity: Digital
Enterprises will no longer collect their data from enterprise, structured
applications but from across the planet in very strange and complex formats
that will be generated in devices, pass through controlled clouds and systems,
zip through servers, filtered via smart routers at the closest point fo data
generators and then placed in low cost data lakes. This data will then need to
be enterprise harmonized and then merged with structured data that users are
familiar with. We need to take a leap from analytics to informatics very
rapidly with a high focus on true business benefits.
More
Automation: With data generators that produce very large streaming
bits of data the only way to do all the above is to identify patterns and the
deploy algorithms to automate the data flow and transformation processes as
close to the source as possible. The IoT informatics system will be as strong
as its weakest link, where a stitch in time will always save nine.
More Secure: This one should actually sit at the top of these topics
but it is hard to start with security if one does not comprehend the width and
depth of the digital data flows from the edge. As the enterprise doors open to
let in data from the edge it will also leave the same doors open for hackers
and undesirables to come in via the edge streams right into the heart of
enterprise systems. So before you start your Big-Data, IoT, Cloud or SAP HANA
initiative start with security. Start with a universal security fabric that
runs from one edge to the other. Remember the cardinal rule the undesirables know
the edge devices are the least secure so they tend to reside there to gain entry
into your enterprise jewels.
…oo ––ššÅÆ››—— oo…
