Small Data, Big Impact!

In recent years, the discourse on big data has echoed across various industries. The transformative impact of Big Data on the technology landscape is undeniable, unravelling unprecedented insights from diverse datasets.
Small Data, Big Impact!

Small Data, Big Impact!

In recent years, the discourse on big data has echoed across various industries. The transformative impact of Big Data on the technology landscape is undeniable, unravelling unprecedented insights from diverse datasets. The evolution of sophisticated tools designed to extract valuable insights from colossal data sets has been a notable development. Beyond the surface, a relentless pursuit of the “why” has propelled data scientists to delve into data’s depths, enhancing our understanding. This intrinsic quest for understanding, answered by the subtleties of “small data,” underpins the advancements shaping our lives and making them more meaningful.

In the realm of data, big data emphasizes the ‘3V’s: Volume, Velocity and Veracity. Conversely, small data focuses on manually collected information through avenues like group discussions and interviews, devoid of tools or shortcuts. The arduous, time-consuming nature of small data collection is unparalleled, yet its impact is ineffable, capable of dethroning even industry giants.

Small Data, Big Impact!

Nokia's Tale: A Lesson in Small Data's Might

Take Nokia, once deemed too colossal to falter in the mobile industry. Its rapid downfall within a year shattered this perception, and much credit is attributed to “Small Data.” Tricia Wang, a tech ethnographer, witnessed this demise firsthand, collecting ‘small data’ or ‘thick data’ that foretold Nokia’s fate before its executives. By observing everyday people and their mobile phone usage, she unveiled a societal shift towards smartphones, a trend overlooked by Nokia’s leadership engrossed in conventional data like sales figures and brand value. This narrative underscores the organisational significance of small data, even for industry behemoths like Nokia.

Defining Small Data: Human-Centric and Labor-Intensive

The preceding example vividly illustrates how small data can wield substantial influence, emphasising the imperative not to underestimate its power. Now, let’s delve into a comprehensive understanding of what exactly constitutes small data. While diverse sources offer varied definitions, there isn’t a singular dictionary definition that encapsulates its essence. Put simply, small data can be characterised as “data in the size and format of human comprehension.” Crucially, it lacks tools or shortcuts for collection, remaining a labour-intensive, time-consuming, manual endeavour at the human level. The versatility of small data applications transcends industries, domains, cultures, and borders, as data, by nature, does not discriminate.

Small Data's Historical Roots

Operating discreetly, small data often remains in the shadows, allowing its sibling, big data, to bask in the limelight. Despite its recent surge in recognition, small data is far from a novel concept; its origins trace back to the 1800s, predating the advent of big data. During that era, the technology required for generating, storing, and processing extensive data volumes was nonexistent. Consequently, people heavily relied on observations, intuitions, and common sense to craft innovations and guide humanity towards the future.

Martin Lindstrom's Framework: The 7Cs

Renowned author Martin Lindstrom, honoured among Time magazine’s 100 most influential, extols the virtues of small data in unravelling significant marketing trends in his book, “Small Data: The Tiny Clues that Uncover Huge Trends.” Lindstrom introduces a comprehensive framework called the ‘7Cs’ for small data practitioners—Collecting, Clues, Connecting, Causation, Correlation, Compensation, and Concept. This framework underscores the multitude of factors directly or indirectly shaping consumer emotions and decisions. Lindstrom advocates small data’s potential in crafting detailed customer profiles, delving beyond surface-level data to include choices and moods at specific moments. This holistic approach facilitates behavioural studies, considering both individual and environmental influences on decision-making. Small data insights prove invaluable for personalised marketing campaigns, especially in scenarios with limited data collection or when faced with novel challenges, where acquiring relevant data may prove challenging.

Small Data in AI: The Chatbot Revolution

In the contemporary landscape, a groundbreaking wave of artificial intelligence has emerged through entities like ChatGPT, representing a new frontier in technology. This innovative paradigm relies on a specialised form of knowledge known as ‘Prompt Engineering.’ Simply put, Prompt Engineering involves meticulously crafting queries, or ‘prompts,’ for chatbots like ChatGPT to elicit the most optimal responses. This practice ensures that the prompts are not only relevant and aligned with objectives but also adept at handling diverse scenarios. In the realm of prompt engineering, small data plays a pivotal role, extensively aiding large language models (LLMs) like ChatGPT. Its contributions extend to customising prompts, fine-tuning models, evaluating responses, iterative refinement, and, notably, addressing contextual prompts and navigating edge cases. The undeniable value of small data in this contemporary era is evident, more crucial than ever in enhancing the efficacy of cutting-edge technologies.

Big Data + Small Data: The Complete Picture

In conclusion, small data excels in addressing the pivotal question of “why.” Recognising that big data alone falls short of providing comprehensive answers, small data emerges as its complementary counterpart. Embracing the maxim, ‘Big Data plus Small Data equals Complete Data,’ underscores the synergistic power of integrating these two approaches. So, the next time you delve into big data insights and find yourself pondering the reasons behind the data’s configuration, remember that the nuanced influence of small data, its elder sibling, might be at play.

P Deepak Harish
P Deepak Harish


Author Bio

Deepak Harish, a seasoned professional with 5.5 years of experience, holds a B.E. in Mechanical & Production Engineering. Currently contributing to L&T IDPL for 2.5 years, his expertise lies in the domains of Machine Learning and Data Science. Committed to innovation, Deepak aims to stay at the forefront of technological advancements, applying his knowledge to create impactful solutions.

Disclaimer: The information, statements and opinions contained in this content are of a general nature only and do not take into account your individual circumstances including any laws, policies, procedures or practices you or your employer or businesses may have or be subject to. Although the statements of fact on this page have been obtained from and are based upon sources that L&T EduTech believes to be reliable, it does not guarantee their accuracy or completeness.