Applying Case Study Methodology to Your Shared Services Problems

Taking The Best (and Worst) From Others Via Case Studies Editor's note: Not many of today's SSO leaders will have taken a degree in statistics, so Anupam Diwan may stand apart. But it does not take a degree to understand that you need to measure what's important in order to make decisions based on real data. In this column, she points out that any and every problem can be solved by taking a "case study" approach. For example, to identify predictors of high performance and quality, by plotting skills against their positive impact on quality. This example highlights the importance of analytics in making crucial HR decisions and guiding business strategy. It's a great approach that you can apply to any number of problems. We learn by observing the actions of others. Whether it’s in real-time – you might have a second thought on trying the grilled squid if your tablemate makes a face after biting into it – or after the fact during a review of a project or initiative that has just concluded, our observations help us understand where we go right and where we go wrong. The case study approach allows us to take a deep dive into observing, retrospectively, the successes and mistakes others make. The world’s top business schools all adopt a case study analysis approach, which gives students the opportunity to dig into the decisions, the data behind those decisions and evaluate other outcomes had things gone differently. You can also apply case studies to your own work. In my previous columns, I’ve talked about dashboards, scorecards, Six Sigma methodologies, making a business case, HR analytics, etc. To apply these strategies, you will need a tool to understand a problem or unravel a specific challenge, and then develop a method to tackle it. Enter the case study. Let’s look at it another way. Have you ever wondered how an orchard owner picks out their best apples for premium pricing? One option is to randomly select, say, every 4th apple and consider it to be a premium varietal. Or, smell every apple and select the ones with a particularly ‘appetizing’ aroma. Or, select them by size, shape, color, nutritional value, the cost of growing a particular type, or by any other favorite attribute. Perhaps a smart business mind will first want to collect data on these various attributes from all varietals in their inventory and perform a detailed analysis. The varietal that scores the highest on all or most of the desirable attributes could then be designated as ‘premium’. Such a descriptive exploration of a problem, often combined with supportive statistical analyses, constitutes a case study. It lays the foundation for developing a strategy that employs the tools discussed earlier, to work towards a solution. The goal of a well done case study should be to confer a strategic advantage beyond that obtained by flipping a coin. You can refer to such a compelling case study on The Ultimate HR Challenge: How to Evaluate In- vs Outsourcing and Implement a New Model – in Just 3 Months. Let me illustrate with another example: A prominent organization’s HR department was facing a conundrum. This manufacturing company decided to send 30 workers for advanced training, to gain a strategic advantage. While everyone in a particular group would benefit from such training, the budget available limited the number of people that could be sent. To guide the selection of workers for the training program, it was important to identify skills that should be accorded the highest priority for inclusion. When ideas fly around the room, tempers can rise, but thankfully, the HR team came to the rescue with an analytical approach to this problem. Actions: HR reviewed the rating system used to evaluate workers’ proficiency in each skill area. The objective was to identify predictors of high performance and quality. The ratings used a seven-point scale ranging from 1 (no skill) to 7 (mastery of skill). The quality assurance specialists provided data on the output from each worker in the form of an index of average quality ranging from 1 (low quality) to 25 (high quality). Approach: A multiple regression analysis was performed with the production quality index as the dependent variable and the ratings for each of the skills as independent variables (say A to M), in order to determine the contribution of each skill to an individual worker’s quality index. This regression model explained the variance in the dependent variable (worker’s productivity) and provided the regression coefficients for independent variables (different skill-sets). This provided a good initial indication of whether the skills make contributions to the quality of production. Variable E and F had the highest correlation coefficients with the lowest P values. This suggested that skills E & F predict quality and therefore might be helpful in identifying individuals that will stand to gain the most from the advanced training program. Therefore, a hypothesis was formulated that employees with higher skill level in E & F will show higher quality and performance, other things being constant; and it was tested using the available data. The regression analysis was re-run using only two factors (skills E and F) that may be assigned the highest priority in the training program. The R2 changed from 0.53 initially to 0.59 when the model was limited to these two independent variables. Given that an R2 of 0.5 or more is considered a decent fit of the model, this confirmed the strength of regression relationship. This two-step approach weeded out variables with small effects by allowing less standard error, resulting in increased adjusted value of R2.Importantly, as the value of R is positive, this relationship indicates that there is a positive correlation; i.e., as the independent variable increases, the dependent variable follows in the same direction. Stated simply, these data suggest that ~59% of the variation in the quality (the dependent variable) could be explained by the variation in E & F, the independent variables. Another aspect of this data worth examining is the ‘P’ value. This is the probability that the null hypothesis is ‘true’, i.e. that the observed relationship between the independent and dependent variables is by chance. In case of variables E and F, the P value was very small (<0.001) suggesting that the chance that this variation was the result of a freak association were less than 1 in 1000, i.e. tolerably small. The next step was to address the second part of the problem, which was to employ the data gleaned from skill-sets E and F to select the workers for the training program. These workers would be the top 30 on the list of those whose observed value of quality is greater than the expected value of quality, based on these skill sets. The data for observed value for each worker was provided by the quality assurance specialists. Expected value for each worker was calculated based on the formula: Y (expected value) = Intercept + [coefficient1* independent var. 1] + [coefficient 2* independent var. 2] Based on this step-wise analysis, 30 workers were identified and sent abroad for advanced training. This example, and those that I have presented previously, highlight the importance of analytics in making crucial HR decisions and guiding business strategy. The case study puts all of the other tools in place and allows us to make better decisions and drive organizations toward better outcomes. Who knows? Perhaps your own case study will enlighten future business students. [inlinead]

Applying Case Study Methodology to Your Shared Services Problems

To continue reading this story get free access

Upcoming Events

From Cost Centre to Control Tower: How Platform-Driven GBS is Redefining Enterprise Value in the Age of AI

Why Your AI Costs More Than Your Workforce and How AI Employees Change the Equation

From Transactions to Trust: Transforming AP, AR, and Communications into Customer Experience Drivers

Through Trust and Governance: How AI Agents are Redefining Accounts Payable

Why Shared Services and GBS Transformation Stalls and How to Overcome Change Fatigue

How AI-Native Accounts Payable Is Changing the Finance Function

Follow Us

Latest Webinars

From Visibility to Action: How Finance Risk Intelligence is Transforming Modern Finance Operations

The AP Automation Reality Check: What True Intelligent Automation Looks Like Inside D365

Accelerating O2C Transformation: How AI Is Reshaping the Cash Conversion Cycle

Recommended

Human‑Centric Leadership in a Skills-Short Environment

The Future of HR Shared Services: 5 Lessons for Leaders

Career Pathways and Capability Uplift in Shared Services

Employee Engagement as a Strategic Lever in the War for Talent

Crafting Teams That Deliver Excellence: Nick Fury’s Framework from Avengers

8 Steps for Integrating DEI into the GBS DNA

The AI Edge: From Experiments to Strategic Advantage in Talent Acquisition

Lookin’ for Love in all the Wrong Places: Hiring Managers Miss the Right GBS Tal...

NEAR, NOW, NEXT. Why Latin America is ready to supercharge your North American b...

Five GBS Leadership Actions to Deliver High-Impact AI

The Better Agents We Need: Reclaiming Human Agency in the Age of AI

The Symbolic Shift: Technology, Meaning, and Leadership in Shared Services

FIND CONTENT BY TYPE

SSON COMMUNITY

ADVERTISE WITH US

JOIN THE SSON COMMUNITY

Applying Case Study Methodology to Your Shared Services Problems

To continue reading this story get free access

Upcoming Events

Follow Us

Subscribe to our Free Newsletter

Latest Webinars

Recommended

FIND CONTENT BY TYPE

SSON COMMUNITY

ADVERTISE WITH US

JOIN THE SSON COMMUNITY