Big Data vs Small Data

Jump starting your people analytics capability
“Great things are done by a series of small things brought together.”
— Vincent van Gogh

Technology provides a competitive edge when it comes to analyzing the value of talent within an organization. Although there is a lot of talk about Big Data, Artificial Intelligence, and Machine Learning, “small data” can offer a significant amount of insight if used correctly.

When it comes to people analytics, there just aren’t sufficient data points to collect about employees that provide a large enough data sample to apply “Big Data” techniques. This is a quality, not a quantity, issue. Identifying the right data points, applying existing theoretical models of psychology, leadership and organizational effectiveness allows you to create great value from just a few simple data points. The most important aspect of collecting and managing small data is data quality, you can’t rely on the usual error correction techniques of big data to remove noise or compensate for inaccuracy.

Martin Sutherland, Global Director at PeopleTree Group, believes that the process of people analytics can use small data. Small data can facilitate meaningful conversations between line managers and HR. There is obviously a role for data science in people analytics, but the best way to jump-start your people analytics capability is to start with small data and robust conversations.

We live in a data economy, and that means data is “currency”. But there is a difference between having money and creating wealth. Creating wealth is the process of increasing value. Even an organization that is data “rich” (they have the “currency”), may lack the ability to turn that currency into wealth that enables growth.


Creating wealth from data is nicely illustrated with the help of a data pyramid.

DATA: Data forms the foundation, and data quality is essential to ensure your “wealth” is not built on a shaky foundation. As an example of people data, let’s consider tenure (how long someone has worked for a company) and age (an indicator of retirement risk).

INFORMATION: Drucker defined information as data endowed with relevance. The easiest way to do this is to combine two data points together and create a 2 x 2 matrix. Using the two data points mentioned above, we could build a matrix of high and low tenure, and high and low retirement risk. Someone who has high tenure and is a high retirement risk means your organization is about to lose significant institutional knowledge and memory. That is relevant.

KNOWLEDGE: Knowledge is information endowed with utility or usefulness. If we were to take what we now know about the lose of institutional knowledge we could create a contingency or replacement planning process. That would be useful.

WISDOM: Wisdom comes from the application of knowledge. By identifying people with significant institutional memory, and creating process to ensure the continuity of that memory in our organization, we have become a smarter company. That is clever.


The illustration above shows the process of moving from raw data to true value. If you want to build people analytics capability within your company, you need to know how to manage this process.

An HR function that is ‘data-poor’, i.e. lacking the basic building blocks cannot participate, therefore they cannot manage the process of creating value, and could not possibly have an impact on the business. On the other hand, a ‘data-rich’ HR function, has the means to participate, can design and implement the processes need to create value, and therefor is capable of having a significant impact on the business.

Analytics continuum

When it comes to people analytics, HR should focus on acquiring three key capabilities: data literacy, technology literacy, and visual literacy. Ideally you want to build these over time, but if you want to jumpstart your people analytics capability, it’s better to “buy” this capability initially.

Your role as an HR Practitioner is to do the job external vendors can’t do, and that is use the tools that are available to have the right type of conversations with your business partners. Turn the analytics into action.

There are two sides to the analytics continuum. The left side focuses on creating actionable insights (AI). The right side focuses on customized research (CR) and model building. They both have their place, but there are some key differences that are useful to understand before you decide how to start building your people analytics capability:

  1. Data Source: (AI) Focuses on easily accessible data; what can we do with what we already have. (CR) Focuses on difficult to acquire data; how do we design data requirements and go and collect it.
  2. Concepts: (AI) Works with relatively simple concepts and models. (CR) Works with more complex concepts and statistics.
  3. Primary Skill Set: (AI) Primarily uses HR skills. (CR) Primarily uses data science skills.
  4. Area of Focus: (AI) Broader HR processes and concepts. (CR) Narrow focused research, limits the noise of multiple variables.
  5. Presentation: (AI) Interactive visualizations to promote conversation. (CR) Mathematical models to crunch data and predict outcomes.
  6. Validity: (AI) Consensus-driven understanding of the data. (CR) Validated statistical models with a high correlation coefficients.
  7. Purpose: (AI) Make decisions and take action. (CR) Build models and predict events.

Extracting insights from Small Data

There are several sources of data that can be accessed to create the foundation of your people analytics platform. The most obvious is the standard HRIS dataset. This consists of several parameters giving the background and basic details of employees. It is important to not only understand the value of each data point, but also the “half-life” of the data, i.e. its period of validity. The second is data from the talent processes within you company and to understand how to tap into each of those processes to secure good quality, relevant data. This includes feedback from managers, performance scores, etc. The more you can design your talent processes to provide you with good quality data, the more valuable those processes become as a source of data.

In summary, a sustainable source of good quality small data is an important foundation, but creating processes that establish connections between data, and having conversations that extract value from that data is how a company converts data from currency to wealth.

To hear more from corporate thought leaders, education visionaries, and ecosystem builders, you can visit our YouTube channel and connect with us for future events.

Get Boarded

This channel provides you career tools, tips and recommendations on opportunities. We talk about trends in tech and…


Peopletree Group Talent Analytics

PeopleTree provides scalable, cost effective talent and career management software solutions to companies who...