Hawthorne Revisited! (People Analytics For Beginners- II)

           The five people elements that have the potential to change organizational performance are Headcount, Attrition (Workforce Turnover), Competence, Engagement and Energy (Employee Health & Safety). HR policies and processes are normally linked to these elements. These 5 elements define the workforce efforts and their interplay with the other organizational systems and variables. Naturally, the correlation and predictions relating to any or all of these 5 with the other key organizational performance indicators are of interest to data scientists. Some of these are customer satisfaction, profitability, productivity (sales, cost and labour), safety (incidents), quality (defects) etc.

            Education teaches us to find out the reasons for everything we see, hear and do. Sometimes we forget that and continue doing things because those were being done earlier. This is true to organizations as much as to individuals and true to our personal lives as much as in organizational lives. Sometimes we run programs because those were run in the past and sometimes we accept events as unalterable. This post discusses few such practices which are continued because of legacy in many organizations. Two interesting studies that were conducted around cost, sales productivity and attrition are described here. The first one was to understand the impact on productivity and attrition because of engagement programs. The second one was to find the zone in the service period of an employee where retention is critical from a financial perspective. Both studies taken together also provide a method to calculate the ROI of the dollars spent on employee engagement.

CASE ONE: Hawthorne Revisited!

             This experiment was conducted over a period of one year in one of the mid-sized Call Centers, part of a large global IT-ITES company. The Call Center process that was chosen was a late stage credit card collection process. The experiment was conducted over a period of one year. The interventions were made in an experimental group and the results were compared with the control group.  The hypotheses to be validated were (a) Increased dollars spent on engagement activities (measured in event-hours) results in increased sales productivity (measured in increased dollars collected), and, (b) Increased dollars spent on engagement activities (measured in event-hours) results in decrease in workforce attrition rate. The independent variable considered was event-hours and the dependent variable was dollars collected and attrition rate. The independent variable was changed over a period of one year. The changes in both the dependent variables were recorded. The outcomes of the experiment are as given below:

1. There is a very high correlation between event-hours and productivity.

2. There is a correlation between event-hours and attrition and the correlation becomes very high if a time lag of 60 to 90 days is taken into consideration.

              The two outcomes mentioned above justify the investment in employee engagement. But the right amount would depend on the gains made from productivity changes and retention. Extra dollars made because of productivity gains were very much measurable in this experiment. The following study provides a way to measure the retention gains.

CASE TWO: Adam Smith Resurfaces!

              In one of the largest IT-ITES companies we conducted a study, covering over 25000 employees, to understand when and how much of attrition is good, for a certain section of employees. This study which was conducted over a year provided many quantitative support to some assumptions that we make normally while managing the workforce. The key hypotheses to be validated were (a) Higher attrition in some vintage (number of days in service) buckets are financially more critical than the others, and, (b) Agent Incentive Program improves productivity of the process. The study was conducted for all processes where Agent Headcount (HC) was more than 50. The processes were divided into 4 categories for better analysis (a) Off-shore Voice, (b) Off-shore Non-voice, (c) On-shore Voice, & (d) On-shore Non-voice. The voice processes had only inbound customer support calls.

                Month-on-month HC and attrition data, for first half of the first year (H1Y1), were collated. Cost parameters were identified to arrive at overall cost of an individual Agent for each process. Process-wise revenue and other expense data were collected. Similarly, attrition rates in each vintage (length of service) bucket were calculated. Adjustments for productivity increase in steps with the learning curve was made. The costs and earnings were mapped for each process and the cross-over points were identified as the Operational Breakeven (OB) points. A simplistic chart is provided below for easy understanding.

              The net cumulative earnings were similarly tracked and the point where those became positive were tracked. The point where the crossover to positive zone occurs is the Cumulative Breakeven (CB) which means the organization’s investment on an Agent till that point gets recovered and the organization starts benefiting from the work of the Agent. The chart given below indicates how the net cumulative earning moves when plotted against time.

              Based on the above, predictions were made on the actual dollar loss that the attrition rate was going to make which would directly impact the bottom line. Efforts were made to change the trend in a positive manner. New programs were introduced to reduce the voluntary attrition rates before OB and CB. All incentive programs were moved to post-OB period so that at least the OB point does not get delayed. Because of the actions taken based on the projections made, there was substantial dollar saving. 

               A sample calculation sheet is given below. Vintage in days and attrition percentage for a process are shown here. Cost of Attrition (CoA) for the same is calculated for different vintage buckets. As you can see after the action taken based on H1Y1 data, there were substantial reduction in dollar losses caused by attrition in H1Y2. H1Y2 represents the same period in the subsequent year. In total the reduction in attrition in OB and CB zones saved USD 1.2 mi from all processes by reducing 13.5% annualized attrition rate.

Note: Red Zone- OB || Red + Grey Zone- CB || White Zone- Profit Period

                The outcome of the study can be summarized as (a) For every Agent in every process there is an OB and a CB. Every agent who leaves before reaching these points causes financial loss to the company, (b) Where there are periodic wage changes for an Agent, depending on the process, the OB and/or CB might get delayed which has to be accounted for (c) Incentive programs unless introduced at the right point will delay the OB and/or CB which in a high attrition process may increase the financial losses, and, (d) Where there are periodic wage changes for an Agent and a fixed earning rate for the process, the cost will overtake the earning at a certain point. The Agent who reaches this stage has to be moved out of the process through career interventions for improving process profitability.

                 Profitability can no longer be achieved only by cutting costs, adding people and/or raising prices of products/services. Companies have to look deeper at data related to “Customers”, “Machines/Technologies”, “Material” and “People” as well as their interplay, to ultimately derive the “Financials” that are desired. In this new scheme of things People Analytics will play a vital role in the days to come. The most important thing to note is that to start working on People Analytics one need not have high end applications. Many of the studies that we have conducted have not used them. This includes studies in companies having well known analytics products. It is all about what you need!

 Related Posts:

1. In Search Of (Data) Excellence (https://www.dhirubhai.net/pulse/search-data-excellence-people-analytics-beginners-v-debasish-das?published=u)
2. Absconding People, Vanishing Dollars (https://www.dhirubhai.net/pulse/people-analytics-beginners-iv-debasish-das?trk=prof-post)
3. The World Is Flat (https://www.dhirubhai.net/pulse/people-analytics-beginners-iii-debasish-das?trk=prof-post)
4. Hawthorne Revisited (https://www.dhirubhai.net/pulse/people-analytics-beginners-ii-debasish-das?trk=mp-reader-card)
5. Actionable People Analytics Model (https://www.dhirubhai.net/pulse/people-analytics-i-debasish-das?trk=mp-reader-card)