Fresh Data Blog
Fresh Data Archive
Making Sense of Big Data
Date: December, 2012 --
Big Data – It’s the latest buzz word in direct and
data-driven marketing and has largely replaced social media as the headline
topic at industry events, webinars, expert articles and case studies. Big Data
is even playing a part in changing direct marketing nomenclature as evidenced
down under with the Australian Direct Marketing Association’s recent rebrand as
the Association for Data-Driven Marketing & Advertising (ADMA).
Big Data means lots and lots of data, from which
companies attempt to draw ideas, conclusions, and above all predictions of
future human behavior. It now has meaning because data capture tools, computing
capability and analytical programs have reached such an elevated level of
sophistication that we are now able to perform increasingly accurate predictive
exercises on more and more massive quantities of data. That data can be
anything, and often is. Climate Corporation takes data about soil, weather, seeds
and many other subjects to help insurance companies set rates for crop
insurance in the US.
Predictions using Big Data are increasingly accurate, and
deployable in nearly every aspect of life, and, of course, the privacy
implications are massive. Add Facebook, LinkedIn, general browsing behavior and
similar online behavioral data to increasingly robust personal information data
of the legacy kind, mix with powerful analytical tools, and you have a huge Big
Data versus personal privacy clash.
Lots of data, smart analysts, and cutting edge technology
together have applications in a wide diversity of fields, from national
security to predicting heart attacks, to finding love. For an interesting
exploration of some of those fields, the recent best seller The Numerati, by Stephen Baker, is an
Closest to our discipline of marketing, we learn that
supermarkets have developed the ability to target their weekly offers by many
different variables, such as religious observance of households, to “bargain
hunter shoppers”. As one might guess, they endeavor to avoid sending circulars
to customers who only purchase items on sale. The maker of a barbecue cleaning
product sold in Walmart focuses its mobile and online spend within 5 miles of
the stores, in neighborhoods in which rain is not expected over the weekend,
and on the two days before the weekend.
Big Data is also all the data about us that has been collected
over time, or is being collected in real time. In the latter case predominantly
but not exclusively on the internet. The scope of data, especially but not
exclusively in the US market, is mind-boggling: purchase history, of course,
property (real estate, second home, car, investments), education, religion,
children, marital status, employment record, etc. And real time data is what we
express about ourselves through our on-line behavior.
During and after Hurricane Sandy, we were not at all
surprised to find that following our searches for “generators” and “chain saws”
on Monday morning before the storm, that our other shopping searches and
weather site browsing brought us new display ads from a multitude of generator
and chain saw vendors through the afternoon, some of them localized to stores
close to us. This lasted until about a week following the storm. Big Data
analysis using the new technologies and algorithms available enable us to make
offers to customers before they even know they need the offers. Lowe’s
conveniently reminded us we would need a gas can with our chain saw.
Amongst online retailers, we’ve seen Big Data play a role
in customizing the online experience to fit a given user based on data elements
attached to that user. As an example a retailer might target the images, copy
and products showcased on its homepage based on the users IP address, e.g.,
coats and mittens for a visitor in Buffalo and shorts and t-shirts for another
in Phoenix or something as simple as ensuring prices are listed in the visitor’s
Social Media data can also be leveraged here. At a recent
industry event we saw how the luxury watch retailer Rolex is able to utilize
social media data from those who “like” their brand on Facebook, which gives
the company access to all the profile information for those users, to customize
web experiences based on this data. For example if, due to a user liking Rolex
on Facebook, Rolex, who are a major sponsor of many athletes and sporting
events, knows a user likes “golf”, the Rolex homepage will prominently feature
their involvement with/sponsorship of golfers, sponsorship of golf events, etc.
and feature watch models in sport categories.
Whether the tools are indeed that sophisticated; whether
the data available to you is robust enough; and whether you, or your customers,
are comfortable with the process, are all the subjects you will need to
consider and will, at least for the time being, likely vary from one vertical
to the next.
In many respects, “Big Data” is merely the point at which
companies and professionals in the direct/data-driven and interactive marketing
world, and indeed in the broader business community, recognize that an
inflection point in the sophistication of the subject has been reached.
No longer do we need to wait for orders on the catalogue
mailing to peak five weeks out to know whether we need to restock. In theory,
provided we have a set of stable data, such as an identified and trackable
customer base, we’ll know much earlier whether to reorder based on some
collection of other data categories which are informative. Perhaps those points
will be what is purchased, or where purchasers are coming from, or maybe even
the colors ordered, or event what the weather is like. Like any other form of
knowledge, the more data we have, the more we can plan.
Do beer sales in South Boston go up when the Boston Red
Sox win? Do sales of high end golf clubs increase after the Masters? Do calls
to golf courses to make tee-time reservations go up when Tiger Woods sinks a
50-footer on TV? Does Amtrak experience more ticket cancellations on Monday
trips if it rains on the week-end? Can they fill up the seats with discount
offers? (Some of the airlines are able
to do price changes “on the fly” literally up to the moment the plane leaves
Does your clicking on this article on the Data Services
website indicate you are more likely to need hygiene services on your database
than other subscribers?
In closing, we also must face the reality that is the
corporate culture and skill set. Using Big Data to market in real time requires
a different mindset and visualization capability through the entire
organization. Decisions now must be made in minutes, not weeks. Risks will be
more numerous, mistakes more immediately visible. Much of our marketing
exercises will be “real time experiments”! Everyone will have to adapt.
There are also
pitfalls that can come with too much data that can lead to inefficiency and
gridlock. Therefore marketers should always be smart about the data they gather
and only gather data for which they have an actionable purpose in mind. Data
Services helps marketers better leverage their customer data through advanced
online segmentation, analytics and intelligence tools within our MarketView database
platform. Contact your Data Services, Inc. representative today to view an