How To Retain Star Performers Without Breaking the Bank: Data-Driven HR Strategy Series, Part 3

Higher salaries don’t always produce more winners.PAY FOR Performance

Consider Andrea Bargnani of the New Yorks Knicks as an example: Bargnani is paid an $11.8 million salary ($6.65 million higher than the average NBA player), but has produced wins of -1.31, which according to this Forbes article, makes him one of the most overpaid and underperforming players in the league.

No organization wants to be like the NBA, overpaying for team members who are underperforming. And everyone wants to find the up-and-comer who will bat above their pay-grade.

But underpay your employees, and your best and brightest will go running to competitors.

So how do you determine whether you are paying star performers what they are truly worth without busting the budget? It all starts with asking the right questions. In the words of the late statistician, John Tukey: “An approximate answer to the right question is worth a great deal more than a precise answer to the wrong question.” With advanced information processing technologies, big data analytics offer organizations the chance to ask questions that yield multidimensional answers that enable HR to dig into the “why” of situations — instead of simply looking at “how many” or “how much.”

If your organization is serious about retaining the right talent at the right price, you can’t afford to ask the wrong questions. Here are the three critical compensation questions made possible by big data analytics that will put you on the right track:

Question #1: How are we compensating our high-performers in comparison to our mid-level-performers?

Why you need to ask it: This question yields a performance-based compensation differential metric. The value in this metric is two fold: It helps you quickly track whether or not your pay for performance goals have been achieved, and  to build and test models to determine the point at which the level of differentiation impacts retention and/or productivity. This knowledge and insight then translates back into your compensation philosophy so that you can optimize your spend to generate the best return. In other words, it allows you to achieve the dream of every sports GM: winning the most games for the least amount of money spent on the players.

How big data technologies help answer the question: Answering this question involves three separate math operations: summing, comparing the difference and then converting the difference to a single output. Even running a SQL query to automate this process to perform these types of calculations could take many hours, let alone trying to do this manually. With the invention of “in-memory” analytics, which is 250,000 times faster than a a traditional database query, creating these types of metrics for analysis and insight has become a matter of seconds, rather than days, so the information is ready at hand when you need it. (To learn more about in-memory and how it works, check out our Big Data cheat sheet.)

Red flags to act on: The amount of differentiation in compensation required to generate results will vary dramatically by role and organization. For example, in some sales roles, it is common for the top performers to earn two or three times what the mid-level performers earn. While in other types of roles, a difference of 10% could be sufficient to generate a positive impact. Hence, it can be hard to generalize, but as a rule of thumb, if there is less than a 5% variation between the compensation of your mid level and your top performers, you will have trouble holding on to your best people.

Question #2: How are we paying people relative to the market rate?

Why you need to ask it:  Whereas compa ratio is a measure of internal fairness, market compensation ratio is a measure of pay relative to the “market rate” for that job. Putting these two measures together enables you to tell a powerful story with numbers. For example, with these two numbers combined, you can give your CEO a clear picture of where your company stands relative to the competition, and give an overview such as: “Our market compensation ratio is 100%, on average. This means we are competitive with peer organizations. Also, our performance based compensation differential is at 1.5, meaning our high performers earn 50% more than their mid-level peers, who are already earning well compared to the market.”

Having this information on-hand also puts you in a better position when negotiating salaries with staff and new recruits. Before sites like and appeared, organizations generally had the upper hand when determining salaries. But now that salary information is more widely available, employees and potential candidates come to the table armed with market rate information. If you know an employee’s  market compensation ratio, you can track — and communicate — how their compensation compares, not just to others in the organization, but also to all those organizations trying to lure away your best employees.

How big data technologies help answer the question: Calculating a market compensation ratio requires a lot of calculations to be conducted at the same time. Without in-memory, it takes hours to pull the data, cluster it, find the dependencies and see how one variable compares to the others.  Every time you run comparisons against different variables — such as different populations related to different market ranges — data moves and gets reorganized. That introduces latency or delay, which is one of the problems in-memory analytics is designed to overcome.

Red flags to act on: A market compensation ratio of 100% means that an employee is paid exactly at the midpoint for the role in your organization’s market. Making the comparison of pay to market common, or normalized, allows you to average this across your organization or different business units and  geographies. If you fall behind the market rate by 30% or so, you will need to make sure your intangible rewards are enough to retain staff. On the other hand,  if you are 50% ahead of market across your organization it is likely that you are paying too much for some of the roles that are less critical to your success.

Question #3: Are salary increases motivating the right people?

Why you need to ask it: This third and final question will yield a metric that deals with employee motivation and change. While most finance people are concerned about “how much” and “how much more” the organization is spending, HR needs to take a more nuanced approach and consider how the overall increase is being distributed amongst individuals. Calculating the direct compensation change per employee helps you determine whether the increased spend is bringing returns in terms of motivating the right people or retaining the best performers.

One of the less frequently discussed topics is the notion of “fairness” and how this impacts employee motivation and retention. There are positive impacts on both employee motivation and retention where employees believe that their contribution is suitably recognized and fairly compensated compared to their peers. It is not possible to get this right for everyone, however getting this right for enough people is a goal worth pursuing.

How big data technologies help answer the question:  In order to calculate the direct compensation change per employee, you need to take a time frame, look at what the individual’s direct compensation was at the start and end of your time frame, and then calculate the difference. This then needs to be compared against performance and employee retention data. With analytical processing and data visualization technologies, HR and other business users can explore the relationships between data sets in a more spontaneous way (enabling non-technical people to change the time frame parameter on the fly, for example) without IT’s involvement.

Red flags to act on: If there is no positive relationship between your high performing and critical talent and their rate of pay increases, compared to their peers, then you may need to revisit how raises are decided and approved. The additional money you are investing in your existing workforce is not being used to deliver the best results. Much as the CFO applies judgement to determine how much more the organization can afford to spend on labor to achieve the organizational goals, HR can show the CFO and the CEO that the money has been deployed wisely, against an effective compensation approach, to generate beneficial outcomes for the organization in terms of productivity (performance) and / or retention.

Want to see the forest and the trees? Ask the right questions

When looking at compensation data, many organizations run the risk of asking over-simplified, two-dimensional questions that will only get people stuck in data rabbit holes. With compensation analytics, you can ask more complex questions, that generate fewer numbers and clearer answers. By focusing on the above three areas (internal compensation differentials, market compensation ratios, and the link between pay increases and employee motivation), you can effectively monitor, understand and optimise the benefits that come from paying for performance.

Ian Cook |
Curious about the differences between gaussian and pareto distribution? Ask Ian. Want to know what it’s like to kite ski North of the Arctic Circle? Ask Ian. Not only is he an expert in statistical analysis and HR metrics, he’s also an avid cyclist, skier and runner. At Visier, Ian helps customers drive organizational change through linking workforce analysis to business outcomes. He is responsible for the workforce domain expertise within the Visier solutions.