As a CxO, What do you know
about your Organisation’s data?
You of
course know, that you need data to run your business. You probably already know
that you are not doing enough with the data you have. You will of course know
that information is very different from ‘raw data’. But, do you know what data
you need about your data?
“Experts often possess more
data than judgement” – Colin Powell
The data explosion
Data
volumes are exploding; it has been stated that more data has been created in
the last two years, than in the previous history of human kind. By 2020, about
1.7 megabytes of new data will be created every second. According to Google, we
perform 40,000 search queries every second on their platform alone.
“Picture a Megabyte as the
equivalent of a small novel”
The data
you possess, or will possess, will be a mixture of structured, unstructured and
semi-structured data, as well as the metadata which describes what your data is,
and what it relates to. Your unstructured data may enter your organization from
many different sources, that are often ignored, are unmanaged, and not at all
tied into the way you currently analyze and report. An example of metadata for
a document, or a scanned invoice, would be; file
size, data created, author. More complex examples are for
telecommunications, where there is a need to record; time, origin, destination and perhaps a transcript of the
conversation.
Metadata is "data that provides information about other data" - Wikipedia
Metadata is "data that provides information about other data" - Wikipedia
Metadata can be created
manually, or automatically from your source systems. Manual creation tends to
be more accurate, but as we have discussed, the volumes are very large, and
manual creation is no longer a practical option.
As you consider your current
and future business strategy, and the systems you need to support the business
you have now, and in the future, have you considered the data you will generate,
and what you need to do, to prepare yourself and your business?
From a business point of view,
there are many advantages in getting this right, and there are many
disadvantages if you ignore this issue. The disadvantages are clear:
- You have a rising cost and administrative overhead of managing this data, and no practical way of gaining any knowledge from it. What do you keep, what do you archive?
- Will you miss an emerging trend that could present itself as a missed opportunity?
- Will you miss a shift in the buying patterns of key customers?
- Do you really know how to allocate marketing and sales effort to the most promising product lines or services?
- Do you have enough historical data to enable you to model the effectiveness of your planned marketing campaign?
What you should assume, is that
your competition is doing something about this, so doing nothing, equates to
putting your company at a disadvantage.
The advantages are therefore
clear. Some examples are, that you will be able to use your data to provide
meaningful insight into your business, thereby helping you make better, and
more informed decisions. With better information you can:
- Target marketing at existing and potential customers by using predictive analytics, who will be more receptive to a product or offer, that is presented at the right time, at the right place to the right person
- Gain real-time feedback on the products or services you provide, via social media feeds, and tailor the service or product to your most valuable customers
- Better manage risk, by using your internal systems generated data, and link it to social and other external feeds, to measure and analyse risk, in almost real-time
- Increased reliability of your models, using predictive analytics on your well-ordered and valuable data sources.
When creating or refining your
IT strategy, use your data requirements to shape and define elements of the
overall IT strategy. Consider the thought and planning that will shape your IT
strategy, in terms of:
- What do you see as your differentiator, and what do you need to invest in, to create that differentiator?
- What systems do you need to generate the data you need?
- What systems do you need to upgrade or replace, to provide the outputs you require?
- What standards do you need to define and manage, to ensure compliance with, and support of your data requirements?
- What external services does your company need to consider?
- What differences are there, in my IT data requirements, and the requirements from a business or operational point of view?
When you have all types of
systems generating a variety of information, it will soon become apparent that
you cannot go back and fix any missing metadata, and justify the expense of
doing so. Just as automotive manufacturers find out, if a vital component is
missing from a vehicle as it rolls off the production line, it’s a difficult
and expensive process to recall all those vehicles and address the issue.
So a clear data strategy, to
get the data right, is an essential part of your planning, and a key influencer
on the systems or solutions you deploy.
And what examples do we have? I
(until recently) worked for a well-known IT giant that was one of the first to
make headlines using the “artificial intelligence” banner. They used a single
marketing title, for all data mining, intelligence, cognitive, and data manipulation
tools. This covered simple business intelligence, as well as the technology
that helped it win a well-known American quiz show.
While attending an internal
seminar in Nice, I talked at length to someone who worked on a trial for the
cognitive computing project, for a large and long-term client. My then employer
volunteered to invest the time to demonstrate the value of the artificial intelligence
tool against years of existing customer data. I asked my colleague how well the
trial was going, and the answer was surprisingly, “not well”. The reason was
that the data was largely unstructured (as provided by the client) and came
from a number of systems and sources, and in order to work with the cognitive
technology, it had to have the required metadata, to allow the mathematical
model to work.
“Metadata management can be defined as the end-to-end
process and governance framework for creating, controlling, enhancing,
attributing, defining and managing a metadata schema, model or other structured
aggregation system, either independently or within a repository and the
associated supporting processes” - Wikipedia
The technology was designed to help
you make informed decisions, using real-time language. Ask it “What are my most
profitable product lines”, may seem like a straightforward question, but just
think of the detail you need to answer that. Eventually, the pilot did generate
reasonable results, and the client was impressed. What the client didn’t know
though, was the manual time invested by our company to retrospectively add
metadata to make it structured. When you did the accounting, it was an
expensive demonstration for us, and the resultant business model did not add up
favourably.
Google is also entering this
arena with TensorFlow, which employs “machine learning”. But it’s the “machine
learning” that takes the time for unstructured data. The easier you make it for
the machine to learn, the faster you will start to see results.
A simple explanation of “Cognitive
computing”, is the simulation of human thought processes in a computerized
model. Cognitive computing involves self-learning systems that use data mining,
pattern recognition and natural language processing to mimic the way the human
brain works.
This sounds like the future,
but the mathematics behind this was created in the 1980’s and 90’s. The only
difference now, is that we have the computational power to cope with the workload,
and it’s now affordable and easy to procure, as and when you need it (Amazon Web
Services or Microsoft Azure, for example, will sell it to you by the hour).
So why should you care? Well, as
we discussed earlier, you are looking for a differentiator aren’t you? You are
looking to keep ahead of the competition? You do want to be the first to know
if your product or service is a hit, or has reached its twilight, and when you should
cease funding? You do want to know who your customer is, and exceed their
expectations?
In the future, when asking the
question, “What are my most profitable product lines?” the machine learning
engine will take your question, and will have assimilated a great deal of data
from internal systems, as well as social feeds, to gain insight into what is
working with your customer base, and what isn’t.
At Rolls-Royce and many Formula
1 teams, they gather huge quantities of real-time telemetry, and use this to
make immediate changes to how their systems operate, as well as feeding this
into future system developments.
When the machine starts to analyse,
starts looking for patterns, and attempts to provide insight, it will only
deliver results as good as the data you provide. Are you sure you have? Is the
data as good as you think? How do you know what you don’t know? Do you have a
process to benchmark, to catch outliers and spurious results? This is a really
exciting development, as it will find the conditions and relationships in your
data, that you had not considered, or, it could simply add weight and
validation to an instinct you had on the decision you were close to making.
Your data will feed this
disruptive technology, and in order to succeed, you will need to consider the
quality, consistency, and compliance requirements, to facilitate faster, and
better informed decisions. This technology will help you to automate a variety
of common work activities that are currently human labour intensive. It will
also help you to augment your human capabilities. Imagine your system suggesting
you extend, rather than withdraw credit from a debtor you are about to call, as
she appears to be the daughter of your largest customer! Who knew? It’s all
about the insight your data can provide!
At Searchlight Consulting we
have used the knowledge across our 300 plus associates, to create a framework
that will help you quickly get to the IT strategy and supporting plan to
deliver your digital business. You may however have started your journey, and
we can use the same framework to advise you on where we think you are, what is
missing, and help you with a plan to complete the final activities to reach
your goal. This article has attempted to explain the importance of aligning
what you do as a business, with your data and information requirements, as well
as your organisations strategic goals, and how you can use these valuable
assets, to innovate, create value and genuine differentiation over your
competitors.
As with the missing metadata
issue, the earlier you address what’s missing, the better the return on your
investment, and the faster you can get ahead of your competition!
