People Analytics: Here’s what you need to do to ensure data quality
People Analytics is the latest buzzword in the field of HR. As HR specialists, we have seen a lot of phases coming and going. While some were mere fads, others got woven into the essence of the human resource function. So where does people analytics stand? It is still early days to answer the question. But the reason I am seeking to find the answer to my title is much more fundamental in nature.
Do not get me wrong- I am a firm believer in the power of data and its analysis. I am an amateur in this thriving field of study (surprise! surprise!- for me at least, the field of people analytics has matured quite a bit while I was living under a rock). Even then I can visualize the tremendous impact that an organizational network analysis can have on an organization when used effectively.
Then what makes me start this article on such a pessimistic note? Perhaps the hesitation in implementing people analytics in Indian organizations stems from my experience within the HR teams till now. I hate to generalize but my experience made me realize that my colleagues in human resources do not understand (or sometimes lack time for) the importance of collating data. More so, their orientation is not towards collecting data in a manner that it can be utilized in an effective and useful manner.
Let me illustrate this with a few examples
A stitch in time saves nine!!!
- Store Resumes digitally – Storing resumes in clearly labelled folders in soft copies assists the search functionality in the system along with Text Analytics. More importantly, feedback should be captured in a standardized format against the role for which the individual has been screened. This sheet will go a long way in churning the data as and when the need arises.
- Not using spreadsheets to collate HR data, be it related to recruitment, L&D or Performance Management- word is the go-to tool. Word format is good for proposals and presentation of an idea. However, its utility finishes there. It has no search and filter capabilities and hence its utility, in the long run, is very limited. Hence spreadsheets should be used to capture data as much as possible.
- Not using standardized status in spreadsheets - Sometimes the right information is not captured in the right column as well. HR fraternity also loves to colour code and each status row is coloured in different colours. Someone at Microsoft Excel was kind enough to add colour filters and hence this is no longer a roadblock! Not using standardized status in a spreadsheet is criminal. The comments column can always be added to enter subjective data. However, the rest of the columns should capture the data in a standardized format as much as possible. This means not turning date & number columns to text format, capturing date, numbers & email ids in pre-defined formats, and not adding additional characters in standard responses (adding an extra space or dash, wrong spelling). Validation rules can be added to avoid these errors if required.
I do understand that not everyone can be a pro at data analysis and organize it in a structured, disciplined manner. Nor do we want everyone to be clones of each other. Some of us have the numerical aptitude and are naturally inclined towards numbers while others can be more creatively inclined (not that both are mutually exclusive). However, all of us as a team is the source of collective data that HR has in its repository. Hence, it is important that all of us follow some basic principles regarding data collection in a structured manner so that this data is available to be utilized when required.
Best Practices for Data Management
It will be unfair to talk about only individual related problems while discussing the challenges faced in getting started with people analytics. A lot of times the organizations start small. Over a span of time, the data lakes slowly turn to data swamps and we are not able to keep up with the requirements. This may be time for us to wake up and take a critical look at our data management. Some of the basics that we need to follow are mentioned below:
- By definition, data management refers to the practice of collating data that is uniform, accurate, consistent and complete. You may want to invest in a data governance program where it is clearly defined who is accountable for various elements of data.
- Regular audits to ensure this also need to be built in the governance model- either random checks or better still, checkpoints within the process. A lot of times people who are required to complete the data are under stress and/ or do not understand the importance of accurate, uniform, complete data. Hence ensuring that this is done at the grass root level is very important
- Align the definition of each of the data points. This is perhaps one of the most important aspects. Each of the business stakeholders maybe defining the components differently and hence even though the data headings are same, the data points are completely different. Merging these two data points is a recipe for disaster.
- A lot of times data is collated in a system and multiple teams (within or outside HR) use different systems that work in silos. Getting all of them to talk to each other, share and enrich data is a key challenge to get a comprehensive master data. If possible, you may want to invest in a comprehensive HRIS system. If it is not possible to create a greenfield, integrating all the systems to create one source of truth is the time & effort well spent.
How each data can be utilized is something for a separate discussion. We are sitting at a point where we are engaging new software to dig into the data collated by our earlier software and help us to use those (all the recruitment data collected over time is now sifted though by deep learning & artificial intelligence software to help us re-find the candidates already in our database!). However, the stepping stone to get us ready for the power of data is a collation of data with integrity, diligence and intelligence. The work well begun is half done!!!