Regional differences, generalized data complicate compensation benchmarking
As an HR manager working in the small, southwestern Ontario town of Mitchell, Jennifer Van Den Broek often runs into problems finding the right salary surveys for her company.
The surveys she finds are typically skewed, most often by too many big city employers in the sample, said Van Den Broek, who works at Durisol, a manufacturing firm employing about 140 people.
What’s more, the category of manufacturing is often too broad for her needs, said Van Den Broek, who is interested in data for the pre-cast concrete industry. What she finds more helpful are the get-togethers with other HR practitioners in the area. It’s at these types of networking events that she and her peers compare information to at least determine what a range might be for a particular position.
Phil Wallace, senior vice-president of Aon Consulting, understands Van Den Broek’s difficulties. Finding the right sets of data is “an extremely difficult thing to get right,” he said. It’s also tough to know how to analyze the data in a survey and spot potential biases.
If companies have difficulty finding data for the jobs they need that are specific to their locales, Wallace suggested looking for local surveys, even if they don’t have the same jobs the company is looking for.
“By matching up some of those jobs to the general industry data that you get in the larger survey, you can see the differential between those jobs and the ones you can find in the local survey. That same differential might apply to a wide range of jobs,” said Wallace.
But he added a cautionary note. The local surveys tend to be smaller surveys, so they are more easily biased. There may be a strong presence of one particular industry in the local market. Or out of a handful of companies in a given industry in a geographical area, one may be a global player that dwarfs the others in competitiveness and, hence, pay.
Sorting out such potential biases in the data is indeed tricky, and that’s why companies often turn to compensation consultants for help, said Wallace.
Another type of mistake companies tend to make is to look at the spread of wages for a given job category, then decide they’ll pay at the 75th percentile, which means they pay higher than 74 per cent of the companies in that sample.
“Not usually a very good idea,” said Wallace. “If you’ve got a very tight survey, one that represents your industry alone, it might be okay. But if it’s a broad industry sample, the 25th percentile may represent some low-paying industries like textiles and retail, industries where the base pay is low, and the mining and petroleum and high-tech industries may be in the 75th.”
What Wallace recommends instead is to peg the salary structure at the median level, and add to that a premium of the company’s choice.
“One more thing. You’ve got to have a good reason to pay a premium. What most companies say they want a premium for is to get the best employees, but to get the best employees it’s not all about paying a premium. It’s about first having the search techniques and methods of hiring that get you the best employees. Then you have to attract them with the right pay,” Wallace.
There are a some good arguments for paying at the 75th percentile, and many are negative.
“One reason is if your company has a bad reputation in the market. Or you’re in an industry that’s in decline, because nobody wants to join a company in an industry in decline. Another will be the criticality of the job.”
When it comes to executive compensation, the issues are the same — how to benchmark to the right peer group and how to put the results into context, said Lisa Slipp, a worldwide partner at the Toronto office of Mercer Human Resources Consulting.
At the executive level, a peer group may consist of only 15 or 20 companies, and that’s because they have to share certain characteristics such as where they’re based and what markets they operate in.
With such small peer groups, anomalies can easily occur. “For example, if there’s a change in the incumbents in the group,” said Slipp. That’s why it would be a mistake to rely too much on benchmarking results. “(It’s) part science and part art.”
The surveys she finds are typically skewed, most often by too many big city employers in the sample, said Van Den Broek, who works at Durisol, a manufacturing firm employing about 140 people.
What’s more, the category of manufacturing is often too broad for her needs, said Van Den Broek, who is interested in data for the pre-cast concrete industry. What she finds more helpful are the get-togethers with other HR practitioners in the area. It’s at these types of networking events that she and her peers compare information to at least determine what a range might be for a particular position.
Phil Wallace, senior vice-president of Aon Consulting, understands Van Den Broek’s difficulties. Finding the right sets of data is “an extremely difficult thing to get right,” he said. It’s also tough to know how to analyze the data in a survey and spot potential biases.
If companies have difficulty finding data for the jobs they need that are specific to their locales, Wallace suggested looking for local surveys, even if they don’t have the same jobs the company is looking for.
“By matching up some of those jobs to the general industry data that you get in the larger survey, you can see the differential between those jobs and the ones you can find in the local survey. That same differential might apply to a wide range of jobs,” said Wallace.
But he added a cautionary note. The local surveys tend to be smaller surveys, so they are more easily biased. There may be a strong presence of one particular industry in the local market. Or out of a handful of companies in a given industry in a geographical area, one may be a global player that dwarfs the others in competitiveness and, hence, pay.
Sorting out such potential biases in the data is indeed tricky, and that’s why companies often turn to compensation consultants for help, said Wallace.
Another type of mistake companies tend to make is to look at the spread of wages for a given job category, then decide they’ll pay at the 75th percentile, which means they pay higher than 74 per cent of the companies in that sample.
“Not usually a very good idea,” said Wallace. “If you’ve got a very tight survey, one that represents your industry alone, it might be okay. But if it’s a broad industry sample, the 25th percentile may represent some low-paying industries like textiles and retail, industries where the base pay is low, and the mining and petroleum and high-tech industries may be in the 75th.”
What Wallace recommends instead is to peg the salary structure at the median level, and add to that a premium of the company’s choice.
“One more thing. You’ve got to have a good reason to pay a premium. What most companies say they want a premium for is to get the best employees, but to get the best employees it’s not all about paying a premium. It’s about first having the search techniques and methods of hiring that get you the best employees. Then you have to attract them with the right pay,” Wallace.
There are a some good arguments for paying at the 75th percentile, and many are negative.
“One reason is if your company has a bad reputation in the market. Or you’re in an industry that’s in decline, because nobody wants to join a company in an industry in decline. Another will be the criticality of the job.”
When it comes to executive compensation, the issues are the same — how to benchmark to the right peer group and how to put the results into context, said Lisa Slipp, a worldwide partner at the Toronto office of Mercer Human Resources Consulting.
At the executive level, a peer group may consist of only 15 or 20 companies, and that’s because they have to share certain characteristics such as where they’re based and what markets they operate in.
With such small peer groups, anomalies can easily occur. “For example, if there’s a change in the incumbents in the group,” said Slipp. That’s why it would be a mistake to rely too much on benchmarking results. “(It’s) part science and part art.”
