In today’s world, as we all know, there is a huge groundswell of change that is taking place in the form of digital technology revolutions. Digital technologies are impacting all aspects of our lives and more so in ways businesses are run. A very significant aspect of using digital technologies is that any activity conducted leaves large data trails that are amenable to analysis. This capability in turn allows us to derive insights on what occurred, spot trends and use this to anticipate future patterns that are likely to emerge. This possibly creates immense pressure on organizations to invest in Analytics with the hope that it would give their businesses a competitive edge.

In the HR domain, we typically use Analytics for quantifying life events of individual employees, identifying common patterns among the different segments and analyzing past career experiences and use that to predict their performance, potential, engagement, productivity, managerial effectiveness, leadership quality, decision making competencies and emotional intelligence. We also use Analytics for quantifying the value derived out of various people practices and programs in an organization. We can identify different early warning signals among employee groups through linkage studies combining internal and external insights that can influence business continuity and sustainability. This is used in prioritizing various people initiatives and analyzing their impact on the organizational culture in the long run. Another area where Analytics is used is for identifying gaps in our current workforce that may lead to business continuity risks in near or long term horizon, forecasting manpower demand to manage fluctuating business demands or production cycles and more importantly optimizing cost of labor.

Meaningful Analytics, however, requires a critical mass of data points to be available in the first place. Usually historical information is not in a form that is readily usable in most organizations. This has a lot to do with how information technology has been deployed in the first place. More often than not, investments in technology, especially for HR, are made through stand-alone applications which are patched together over time which results in data that has questionable reliability, is inconsistent and incomplete. Added to this data definitions and formats across different applications tend to be out of sync, making things more difficult for getting started with Analytics.

If this scenario has an uncanny resemblance to the situation in your organization and you are under pressure to implement Analytics what do you do? For one, you have no choice but to take steps to fix your broken processes and systems if you want to have sustainable, meaningful and value adding Analytics. This however does not mean you cannot get to the Analytics part while you are taking stock of the situation and rethinking your IT options for implementing digitized processes. As a matter of fact, you could actually use Analytics tools in this exercise of rebuilding your base.

The steps you need to take to add the Analytics layer when you have some IT systems in place are:

Revisit the thought process and reasons that led to your situation with your IT infrastructure

When organizations implement IT systems in HR, the tendency is to choose an application first, usually payroll and the other processes are built around this. Overall this approach would not be wrong, were it a phased implementation of a clearly articulated larger design of an HR operating model. However this is rarely the case as mostly systems when implemented in a piecemeal manner, due consideration to evolutionary paths of processes, systems and data is not factored in. The interrelationships between process, systems and data have a significant role in determining the quality of data. So draw lessons from the past to change how you implement digitized processes in future. The first step in course correction is to revisit your HR operating model.

Redefining your HR operating model

Redefining your HR operating model will help you articulate how you want to deliver the “HR experience” to employees and will also help you to redefine your HR processes and establish a clear direction for your technology investments. This will also help you establish your basic requirements from Analytics. As mentioned earlier HR Analytics covers a large spectrum, so pick basic descriptive Analytics like demographic insight reporting and more specifically diagnostic studies or insights on employee profile data.

Run an assessment of quality of data you already have

Even if you are lucky or have been prudent to have in place a good integrated IT platform and have a well-defined HR operating model, you still need to assess the quality of data you have. This can be done by assessing your individual employee data. Use the data you already have to derive basic insights on demographics by using any basic analytical tools on data extracted from your HR systems. Results of your studies will give you a fair idea of completeness, correctness and consistency of data. Typically in process steps where certain pieces of data are not mandatory, that data point will obviously not be available for the entire population. This data point will be incomplete and any Analytics requiring this will therefore have limited value. However, it will definitely give you pointers on what needs to be fixed at the process level and the system level. Similarly correctness of data i.e. duplication, format issues, etc., and conflicting data will show up. These results if analyzed patiently will give you a wealth of information on what happened at the transaction level and what assumptions went into the process and system design.

Define data standards that would make your data “Fit for Purpose”

Once you have an assessment of your data quality, it will not be difficult for you to get down to laying down the standards for your data. The purpose of defining data standards is to provide a consistent meaning to data shared among different information systems of HR that are in use. Also remember there will be certain HR master data that will be shared with other domain systems or perhaps 

even imported from other domain. The most common data that is shared throughout the organization is the organization structure, personnel structure and enterprise structures apart from some basic employee data. Therefore standard definition should include format and structure, description, definition of terms and management of data. 

Another important activity here will be to determine what data needs to be captured so that your Analytics requirements are fulfilled. Quite often this is overlooked, which results in limiting the value of Analytics. For data input needs to be fulfilled, you will have to ensure that there are processes in place that generate the required data.

Understand how your digitized processes generate and use data

Having established what you need and in what format, you have to pay attention to how this data will be generated. To do this you will need to understand where and how key data is generated and who uses them and for what purpose. A review of your HR processes at this stage will be useful. This review will help you identify reengineering and process improvement opportunities aligned to your Analytics objectives.

Re-engineer your processes such that bad data generation is neutralized or controlled

Having identified the key improvement areas in your HR processes, the next obvious step is to launch into a reengineering and data quality management program. However, depending on the state of your current HR systems, this could be a massive reengineering or transformation exercise or if you are lucky, just a process improvement program. In any case, a data quality management program will be required. This is the foundation to implementing Analytics.

Lay the foundation for building an Analytics layer

To state the obvious, it is very clear that the foundation to any meaningful Analytics implementation is the availability of clean, reliable, “fit for purpose” data. Generating clean data is influenced by your processes which in turn have to deliver to the objectives implicit in your HR operating model. The earlier mentioned steps lay the foundations for a digitized environment, which is the starting point for adding an Analytics layer to your infrastructure. Summarizing, the foundation activities are:

1. Preparing data available for use. This involvescleansing, standardizing and enriching the data.

2. Defining an HR operating model and begin aligning process to deliver to the objectives of the operating model.

3. Building a data governance framework to maintain and sustain the quality of data and ensure it is “Fit for purpose”.

4. Establishing your Analytics roadmap.

The implementation roadmap will have three streams that should be aligned

1. A process evolution path,

2. A systems and technology evolution path, and

3. An Analytics evolution path

Each influences the other and are not necessarily sequential. For instance, Analytics may throw up ideas for interventions which require process changes but technology may be lagging. Or technology evolution may throw up capabilities that could be leveraged to deliver radically new experiences but process and culture changes may become the limiting factor. Analytics capability itself will evolve from strategic reporting, to descriptive Analytics and then to predictive and prescriptive Analytics in different areas of HR, as the operating model will determine the priorities. So on this journey there will be a constant iteration of ideas and solutions between the streams.

So is it possible to add an Analytics layer to your current IT landscape without waiting for a full technology and business transformation exercise? The answer is a firm “Yes”.