With big data quickly becoming a household term, as CIO you’re
soon going to be asked if this is a relevant technology for your organization
(assuming you haven’t already been asked). This can be a challenging question
to answer. Glibly dismiss big data as too expensive or irrelevant to your
organization, and you risk appearing out of touch. Jumping into a deep
technical discussion of the nuances and challenges of working with big data,
and you risk being regarded as the “plumber,” called in after the
decisions have been made to “just make it work.” Here are some
recommendations on how to steer discussions about big data, whether your
organization is large or small.
First, define “big data”
Half the challenge of discussing big data is that there is
no single product or tool that represents it. Unlike an iPhone or ERP package, big
data is more a concept than a product or clearly defined set of criteria. When
speaking with a peer outside IT, try to determine precisely what they’re
referring to when they broach the subject of big data. Are they referring
simply to rapidly analyzing large volumes of existing corporate data, or trying
to process massive sets of unstructured data? Do they want to incorporate
external or environmental data into existing reporting to provide more insight,
or do they want to throw massive data sets at PhD-level data scientists to see
what comes out?
More often than not, when people start referring to “big
data” they’re more interested in incorporating existing enterprise data
and rapidly performing ad-hoc queries across applications than wrangling
massive sets of unstructured data and performing exceptional complex analyses.
Taking the time to define “big data,” and realizing that Marketing may
have a very different definition than Accounting will ease the discussion.
Forget the “big” and focus on the data
In many organizations, data are siloed in various
departments and applications. Even a single business unit might not be able to
produce a seemingly straightforward report, since the data required are not
linked or consolidated. Much of the attraction of big data to those outside IT
is the assumption that this new technology performs some kind of technical
magic, allowing any and all data sources within a company’s four walls to be
quickly linked and analyzed.
These scattered data sources often indicate a company is not
mature” enough for true big data and could likely benefit from some
more basic initiatives like building an enterprise data warehouse, building a
basic Business Intelligence capability, or deploying a user-friendly reporting
and analysis tool.
Is “big data” code for “better reporting”?
For many, one of the major promises of big data is improved
reporting. Big data demonstrations are rife with fancy online reporting and
easily digested dashboards. In some cases, requests for big data may really be
veiled requests for better reporting capabilities. Even organizations with
extensive data warehousing and BI capabilities sometimes skimp on reporting
tools, either assuming dashboards and reporting will be done “later”
or providing tools but skipping educating people on how to actually use them.
If much of the infatuation with big data seems to be around the reporting, it
might be time to dust off that old spreadsheet with reports that were tagged
for a “future phase,” or launching a few mini-initiatives and quality
training to improve reporting capabilities.
The good news about all of the hubbub around big data is
that it creates an opportunity to talk about, and get funding for, data-related
initiatives. Sharing the excitement around big data while steering the
conversation toward corporate data in general lets the CIO temper some of the
enthusiasm with realism, and often retarget a big data initiative toward
something more realistic and beneficial in the near term. Even if you’re not
technically correct in referring to a basic data cleansing and consolidation
exercise as “big data,” the growing interest around the subject just
might let you build some increased capabilities that accelerate your “data
maturity” far more than an ill-conceived and unwarranted initiative.