It is time HR starts focusing on Data

It is time HR starts focusing on Data

There is an old saying, “if you want something to get done, you need to start measuring it”, Unfortunately, for the Human Resource Function measurement has eluded most aspects of it ever since the Function came into existence. While the name of the Function has undergone several changes over the last few decades from being called Personnel to Human Resource Department to being called the Human Capital Function, what has remained unchanged is the lack of universally accepted measurement practices.

There are essentially six fundamental issues with the use of data for Human Resources Function. First is the lack of data on a uniform basis across organizations and across time to make it universally comparable. Unlike the financial metrics which have a common definition and is universally understood, metrics in HR function do not have universal definition and hence comparison across organizations and across time within an organization often becomes irrelevant and hence is not used. The second issue is that the metrics that are used, more often than not, the metrics are only descriptive and not predictive in nature. In other words, they only describe events in the past and does very little to throw any light as far as the future is concerned. For instance, attrition rate in the year gone by is rarely any indication of the attrition in the coming year. In a world where the present and the future were a linear extension of the past, this kind of data may have been useful but in a world where everything changes so rapidly that can hardly be correct. The third issue is with lack of interpretation of the data that is captured. Even if the data is available unless the same is interpreted and inference drawn from it, the data is of very little relevance. For instance, if the attrition percentage is 10% in 2013, 12% in 2014, 8% in 2015 and 7% in 2016, one needs to interpret the same and draw an inference out of it. Otherwise these are merely numbers and describe an event over the last few years. The fourth is the absence of effective benchmarking. Data per se is nothing but numbers unless it is seen in relation to others. Again, the attrition numbers mentioned above have very different meaning if the comparable organizations have double that attrition rate or they have half that attrition rate. Unless one is able to benchmark and compare much of the data may be quite meaningless. Fifth is the classification of data between Lead and Lag Indicators. In order to understand the reporting and use of data in a holistic sense, we need to distinguish between what are Lead Indicators and which are Lag Indicators. The former typically help in predicting what is about to come whereas the latter describe what has already happened. Any holistic approach to data would require the use of both. Some of the examples of Lag Indicators would be, say, Cost Per Hire. This describes the average cost that the organization has incurred for every hire. However, this data is describing the event after it has occurred and does not help in predicting anything for the future. Although this data is very valuable but one also needs to look at a Lead Indicator to predict what this metric would look like in the next Quarter or next year. The Lead Indicator for that would be the proportion of hires being done through on-line job portals versus being done through head-hunters. Typically, the latter is likely to increase the cost per hire as compared to sourcing candidates through on-line job portals. In this example, unless an organization is able to increase its dependence on job portals as compared to head-hunters, it is unlikely to be able to make any significant difference on cost per hire.  Sixth and the last point pertains to our ability to measure the impact of people on business. Are we actually measuring the return on investment on people? If not, which I am afraid is the case more often than not, then it is very difficult to justify increase or decrease in investment on people related initiatives. It would be left to the discretion of individuals within the organization and the general financial condition of the organization. When the going is good, there would be a higher learning & development budget and when the going gets tough, the first casualty would be on investment on people.

In order for data on human resources to be used universally and authentically, first and foremost is the need to have universally accepted metrics across organizations and over time. It is time that there is a universally accepted norms on reporting human resources data. For instance, attrition data which is captured by virtually every organization, should be calculated using a standard norm and should be reported as part of the annual reporting. It should not be left to individual organizations and, even worse, to individual HR Functions to decide how they would calculate attrition percentage. While there is a universally accepted norm for calculating and reporting Net Profit After Tax, there is no such norm for attrition percentage, for instance. More than anything else, this would require the discipline of capturing and reporting data on human resources in a rigorous and transparent manner by the Human Resources Function.

The second point is that Human Resource Function should engage in predictive data analysis going way beyond just describing a data. This requires thorough analysis and understanding of the parameters which influence a particular metric and how those parameters are likely to change over the coming period and therefore the influence those parameters would have on the metric concerned. For instance, for attrition, one of the parameters may be the GDP growth of the country which has an impact on external job market. The critical issue would be to identify the factor as having an influence on attrition and then basis that predicting what the attrition percentage is likely to be in the foreseeable future. Of course, there would be multiple parameters which would have an impact on any of the metric and hence one has to be able to identify the influence that each of the parameters has on the metric concerned and thereby able to predict basis the likely change in each of those parameters. For example, while GDP growth may be one of the factors influencing attrition, the market position of the company with respect to comparable other organizations on the compensation front may be another parameter. Now if the organization plans to increase its relative position on the compensation front vis a vis other organizations then even if there is an increase in GDP there may still not be higher attrition.

The third point is the ability to interpret the data, even for the past period. It is one thing to merely report the numbers and quite another to be able to draw inference from the same. Most HR Functions only report numbers without interpreting the same. This needs to change. It is one thing to state that attrition percentage is x and quite another to be able to draw an inference from there. Inference is always based on trends and it is vital that we start looking at trends to be able to interpret the data. Interpretation of data is what would make the data relevant and useful.

The fourth aspect is the ability to benchmark with relevant others. I deliberately use the “relevant others” instead of industry or region. It is time we shift focus from industry or region benchmarking to identifying the relevant other organizations for the same. There are few organizations which lose talent within the same industry or hire talent from the same industry. Now if talent is being lost to and hired from different industries or regions, why restrict benchmarking only within the industry or region? Organizations need to identify the companies where there is talent flow (both in and out) and mark them as “relevant others” for all kinds of benchmarking. Reason benchmarking is so critical because everything is relative. For instance, an Organization Commitment/ Engagement Score of 65% in an organization health survey is irrelevant unless it is seen in the context of what similar scores are for relevant other organizations. Of course, there is a challenge in making a like for like comparison given the plethora of surveys in this particular example; but by & large every survey worth its salt measures more or less the same dimensions and hence one should not shy away from comparison just because the survey is not exactly identical across organizations.

Fifth point on Lead & Lag indicators is very critical because we need to be able to distinguish between the two and use the data accordingly. The Organization Health Survey scores, for instance, is a lead indicator for retention or attrition in the organization. If the score is poor or has deteriorated over the previous time period, it usually reflects a higher dis-satisfaction amongst the employees and is therefore a signal that attrition is likely to increase in the foreseeable future. Or if the scores on Manager Feedback Survey has improved over the previous period it usually means that employees are more satisfied with their managers now than they were in the past and hence the attrition is likely to be lower.

The last point is on our ability to measure the return on investment on people in an organization. While there are methods prescribed which are available in public domain, there are very few organizations that I know of, which are using the same. The simple formula that I find very useful is the ratio of Net Revenue plus cost of salaries and benefits to the employee costs. This provides the return on every dollar spent on employees. If the ratio is greater than one, it means that every dollar spent on employees is yielding more than a dollar. Higher the ratio, the more is the return on investment on employee cost. In this case employee costs should include all costs from cost of hiring to cost of attrition including the cost of salaries and benefits. There may be better methods of capturing the return on investment on employee costs, and if there are, we must use them but the fundamental point is to shift focus in measuring it. We cannot not measure it just because as of now there are no universally accepted method of measuring it or it is not mandated by law of the land to report the same.

It is time that HR as a Function starts focusing on data in a much more holistic manner instead of merely reporting numbers. It is important because that would accurately measure the impact of the Function on business and would go a long way in understanding the relative impact of various people-related initiatives on the business per se. That would be beyond the mere gut-feel of which initiatives are having an impact on business performance and which are not.


Tej Mehta

Senior Executive | Airline, Aerospace, MRO

7y

Very well written. The article captures very well some of the challenges with widespread adoption of HR analytics in industry...

Like
Reply

I especially appreciate your calling out the distinction between leading and lagging indicators. There's utility in describing what's happened before with lagging indicators, and in many business contexts, it just so happens that the best predictors of what will happen next is what has happened in the past. However, it's particularly those situations where the underlying factors may be changing - where the signals associated with lagging versus leading indicators may lead to different conclusions - where we get blindsided.

Akhil Agrawal

Founder and MD Gajaraj Hyundai, Past Chairman Automobile Subcouncil Eastern Region, ASSOCHAM

7y

Cannot agree more!!

Like
Reply

To view or add a comment, sign in

Insights from the community

Explore topics