Big Data and Smart Data - what's the difference?

Gustavo Fidel Uy
Gustavo Fidel Uy


Big data is evolving. Many are realising that being inundated with information may be counterproductive. This, in turn, has led to a new trend called smart data. 

Quick Links
Big data, done haphazardly, can be counterproductive 
The Concept of Big Data is Evolving
What are the tenets of smart data?
Big Data is in a state of flux, now is the time to adapt

Big data is probably the buzzword of the decade, dominating the world of enterprise technology simply because of how big of a game-changer it is. According to IDC, worldwide revenues for big data and business analytics will soar to US$274.3 billion in 2020, up from a projected US$189 billion by the end of 2019.  

The surge in revenue is hardly surprising. For many companies, big data holds the keys to the kingdom: information and insights they normally wouldn’t have access to. This information, in turn, enables organisations to make data-driven decisions, with applications in government, healthcare, manufacturing, sales and marketing, financial services—the possibilities are virtually endless. 

Traditionally, the definition of big data has always involved big data sets. The assumption is that the more data you gather, whether from inside or outside the organisation, the more likely you can uncover trends, patterns, and correlations, especially when it comes to desirable actions and human behaviours. However, more people are realising that bigger data warehouses/lakes don’t necessarily produce better, more actionable insights.  

In other words, it’s not just about the volume of data, it’s also about quality and accuracyThis shift in focus has given rise to what’s known as Smart Data. But what exactly makes it any different from big data?

Big data, done haphazardly, can be counterproductive 

There is a growing misconception among the thousands of organisations eagerly jumping on big data offers and data infrastructure that big data is about acquiring as much information as possible. However, not as much attention is given to whether that data is actually relevant or correct.  

To understand this problem, analysts from Deloitte conducted a survey of big-data findings from third-party sources (firms that provide raw data to marketers) and found that more than two-thirds (71 percent) of respondents said that the data about them was only 0 to 50 percent correct. A third of respondents said the data about them was only 0 to 25 percent correct.  

In contrast, smart data ensures that data is free of errors by sourcing information through smart sensors or raw data that’s been sanitised and optimised for analytics at the highest quality and speed. It’s the extra step of processing that provides context and relevance to your data, allowing you to be more confident when making decisions based on that information.  

For example, ‘regular’ big data as we know it would simply amass a list of numbers referring to weekly sales. Smart data, on the other hand, seeks to make sense of those numbers by using, for instance, algorithms to identify peaks and valleys in sales. Again, it’s that extra layer of business intelligence that adds value to raw data, making it ‘smart.’ 

The concept of big data is evolving  

Big data is composed of six Vs: 

  • Data volume 
  • Data velocity 
  • Data veracity 
  • Data value 
  • Data variety  
  • Data variability 

Smart data’s emphasis is on reducing data volume, which, in turn, results in an increase in the other Vs. In particular, the insight-oriented approach to data means that value, velocity, and variety should all increase alongside a reduction in data volume. Data variety and variability can also decrease, but this depends on how the data is sanitized.  

While much has been said about tech companies like Amazon, Netflix, and Spotify leveraging big data to understand customer behaviour, more traditional enterprises are also using big data effectively to create memorable customer experiences. 

Take Starbucks, for example.  

According to data collected by Harvard Business School researchers, the coffee chain records over 90 million transactions every week across 25,000 branches around the world. The amount of data the company is collecting, of course, is staggering, but it’s also compartmentalizing these data sets to make it easier to bring their focus back to the customer. 

One way Starbucks does this is by using local real estate data, including road congestion, food traffic, local demographic data, and customer data to determine the potential success of a new location before it even opens. Note how this data becomes ‘smarter’ simply by adding a filter (i.e. real estate potential of a branch).  

What are the tenets of smart data?

Apart from cutting data volume down to size, smart data is about knowing what to look for and where to look for it. Gone are the days of believing that there’s gold in zettabytes of data. From the onset, your organisation needs to have defined reasons for using the information collected. This will make data gathering more insights-oriented which businesses can be confident about.  

Smart data also needs to be fast—think information as-it-happens, which enables real-time decision making. For example, healthcare providers leverage data in order to make better decisions for their patients. However, most of their data reports are backward-looking—they simply aggregate information and let practitioners know what happened last month or last quarter. Smart data allows healthcare providers to get straight to the critical information so they can make impactful, real-time decisions for their patients. 

UOB, a bank based in Singapore, is an example of a brand that uses these principles effectively. In 2018, the financial institution partnered with Intel to test how data analytics can enhance its cross-border anti-money laundering (AML) efforts.

The joint project set out to combine advanced data analytics with leading technology (i.e. the system of filtering data) to provide greater clarity on the extent of transactions made by a single client across multiple countries and entities. UOB and Intel’s partnership resulted in a federated analytics model that gave them an expanded view of a client’s cross-border banking activities without infringing on their privacy and local data protection laws.  

The federated analytics model allows for the sharing of algorithms across different geographical data sites without actually having to share data, which protects its sovereignty. The algorithms are designed to sift through the data and draw specific insights, patterns, and indicators to identify money laundering activities.  

Big data is in a state of flux, now is the time to adapt 

Big data was the new kid on the block that got companies excited about gathering huge volumes of information about their customers/stakeholders, supply chains, manufacturing processes, sales, and marketing performance among many other things. Things have begun to take root and enterprises are quickly learning that time is the enemy of data-driven organisations due to the lag between data gathering and data preparation in traditional analysis. 

Smart data is knowing where and how your data aligns with your business objectives and being able to craft a story from the numbers collected. Again, the key is to reduce data volume by filtering out unimportant data, identify what you wish to do with it, and collect data as it happens.  

If you need help leveraging smart data and other data analytics trends, talk to the experts at DataVLT. Our easy-to-use analytics platform filters and interprets your data so you make smarter, more effective business decisions quickly and effortlessly. Contact our team today to learn more




The Forefront of Data Analytics — AI Builders of the Future 
DataVLT provides artificial intelligence as an outsource solution. We exist to help enterprises accelerate their growth through digital transformation by way of data analytics. 

Learn more at 


Sign Up For Our Newsletter