The Digital Transformation Playbook: Rethink Your Business for the Digital Age

Big Data Is Really Unstructured Data
Traditionally, a firm’s data processes were based on analyzing structured 
data—the kind of data sets that fill a database with neatly organized rows 
and columns (e.g., with addresses of customers, inventories of products, or 
expenses and debits of various financial accounts).
But the big-data era has been marked by the profusion of new types 
of unstructured data—information that is recorded but doesn’t fit easily 
into neat forms. A business may have access to the ungrammatical text 
posts of social media, the flood of smartphone-generated images, real-time 
mapping and location signals, or the data from sensors rapidly spreading 
over our bodies and our entire world; all these types of data are rich in 
meaning—but difficult to parse by familiar tools like spreadsheets.
One of the biggest sources of unstructured data is social media. As over 
a billion users worldwide participate in networks like Facebook, Twitter, 
and Weibo, they are constantly producing vast amounts of data in the form 
of their posts, comments, and updates. This social data is attitudinal (what 
people are saying can capture their opinions, likes, and dislikes) and can be 
used to measure affinity (whom they friend, follow, or link to reflects social 
ties and allows businesses to infer relationships between them and oth-
ers in their network). And this data is real-time and continuous, allowing 
businesses to analyze shifts in opinion, sentiment, and conversation with 
precise longitudinal detail. Because of this, numerous organizations have 
sought to gain insight from the analysis of social data. Brands monitor their 
reputation over time based on what customers are saying, the Centers for 
Disease Control uses social media to help track the spread of flu and influ-
enza, Hollywood predicts the opening weekend performance of new mov-
ies based on the social “chatter” after opening night, and economists have 
even used social media to effectively predict stock market performance.
Another new kind of unstructured data is location data. The data being 
generated by mobile devices like smartphones comes with geolocation 
markers, which provide a continuous record of where we are and where 
we’re going in real time. The inclusion of location data with other kinds of 
behavioral data adds tremendous additional context. Increasingly, search 
engine results are shaped not just by the words we are using in our search 
but also by where we are when we search. (If we Google the word pizza, we 
are likely to be shown the closest establishments, with links to their phone 
numbers and addresses, instead of pizza history or recipes.) Research by 

---

T U R N D A T A I N T O A S S E T S

99
my colleague Miklos Sarvary has shown that the patterns of where we go at 
various times of the week (as measured by our phones) reveal a great deal 
about who we are. By analyzing these “co-location” patterns, Sarvary and 
his coauthors were able to show that customers with similar location “foot-
prints” were likely to buy similar products and could be effectively targeted 
for marketing based on that data alone.
6
The biggest emerging source of unstructured data is the sensors that are 
becoming embedded in everything around us as we shift to a world of truly 
ubiquitous networks. By 2020, Cisco expects that over 50 billion devices 
will be connected and sharing information over the Internet—and the 
vast majority of these devices will not be computers, smartphones, or Web 
servers. This phenomenon, known as the Internet of Things, encompasses 
smart automobiles, factories and product supply chains, and lightbulbs and 
home appliances as well as sensors embedded in the watches and clothing 
we wear and in the medicines we ingest. Together, all of these applications 
will soon result in billions of devices transmitting and generating new sets 
of data that can be put to business use. For example, GE has installed sen-
sors on its jet engines that allow the engines to continuously post updates 
on their status and operating details. (GE calls the system “Facebook for jet 
engines.”) This real-time data lets airline mechanics monitor the status of 
critical aircraft equipment so they can make repairs when they actually are 
needed rather than on a schedule of estimated need. This makes fleet main-
tenance more efficient and makes air travel cheaper and more convenient.