Comparing salary ranges? Really?

Comparing salary ranges

I was recently approached by a well-meaning Organisation Development specialist working for a governmental organisation. This gentleman emailed me the following :

“I have been requested to benchmark our salary grades with other (semi-)government entities. Therefore we are approaching some organizations to see what they can share with us. We are mainly interested in the salary policy scales (not actual salaries).”

Well, I work for a semi-governmental organisation and we have a duty to support our “cousins” so I said I would cooperate if needed – sending out our salary ranges.

But honestly ? Asking an OD guy to perform a salary comparison ? And on top of that, to compare salary ranges ?

This is a good example of what bothers me about the state of Compensation & Benefits knowledge and understanding in the GCC.

First, this request assigned to the wrong person.

It is important to delegate the work to the person most able to deliver it so that, if anything goes out of plan, they can take corrective action to get to the best outcome possible. Compensation questions ? Give them to C&B specialists !

A Compensation specialist (in-house or external) would hopefully have the right arguments to influence the request, and be able to convince management that in order to produce the results they aim for, the survey needs to be done in a different manner.

Yes, influence management… because this specific request is likely going to produce irrelevant results. Why ?

The first reason why comparing salary ranges is not going to produce relevant results

All government and semi-government entities do not have the same grading structure. So the salary ranges are not created to represent the same levels of responsibilities. Thereforecomparing the salary ranges is notcomparing apples to apples.

The second reason for the irrelevance of comparing salary ranges

Maybe the government departments all use the same grading and target position to market ie they all want to pay at the third quartile, I honestly don’t know. But for sure, the semi-governmental organisations don’t. Each has their own compensation philosophy. So, even if we ignore the fact that the grades/levels are different and therefore difficult to compare, we have the difficulty of not being able to compare midpoint to midpoint across organisations.

The third reason why you shouldn’t compare salary grades

Beyond the midpoint, all organisations don’t create their salary ranges in the same way. Some will go plus or minus 20% around the midpoint. Some will create wider salary bands. Some may have a mix, with narrow salary bands for junior levels and wider bands for senior grades. Again, just getting the ranges will not guide the person doing the survey into understanding the data.

The fourth reason why salary bands comparison does not produce relevant salary information

Salary structures are just that – structures. When you look at a salary range, you have absolutely no notion whatsoever of how much employees are actually paid in the organisation. They may be spread across the whole range, massively represented at the bottom or top of the range, there could be significant numbers of employees paid below the minimum of the range or above the maximum. So again, looking at salary ranges does not provide information about how the organisation actually pays.

The fifth reason why it is not relevant to compare salary ranges

Let’s put aside all previous points and imagine that the company has managed to get salary ranges from multiple organisations. What do they do with them ? How do you compile the information ? Averaging midpoints does not represent anything. Maybe looking at the lowest minimum and highest maximum can bring some element of information, but as I said in the third point, if the range spreads are not the same across the surveyed organisations, then even that is not going to be meaningful.

All these reasons mean that even if they are able to collect data, this organisation’s HR department would not be in a position to move it from its data form to value-adding information. Yet information is what helps inform future decisions. So please, if you ever receive this kind of request, try to influence top management to reconsider what they are trying to achieve and get them to allocate the appropriate resources to perform the appropriate study !


Liked this post ? Why don’t you share it on LinkedIn or Facebook ? Simply click on the link below and help your friends and connections benefit from it too !

Related posts :

Print Friendly, PDF & Email


  1. Sandrine,

    I read your post about salary range comparisons and felt compelled to share some slightly different views.

    I totally agree with the point you make about assigning a compensation and benefits study to a trained specialist. Further, I agree if you collect salary range data, and nothing else, it is fairly useless and won’t yield what is needed for a robust salary comparison.

    However, in a proper salary survey there is a key role for salary structure information. You see, the structure captures the range of compensation levels that the employer uses – they don’t pay less than the minimum or more than the maximum (if they do, then the range being shared isn’t accurate). All of the incumbent data should fall within the range.

    Now, if you capture range data in a proper survey, which includes a disciplined approach to job matching, you solve the issue of varying range definitions, since the job matching allows the survey provider to capture the contribution level for the job, and therefore, the contribution level for the range as well. Once this benchmarking is in place, you can compare ranges, and in fact, we think such comparisons work better than incumbent data comparisons alone.

    What? Am I crazy? Sadly, no. Traditional survey lore suggests that incumbent data drawn from a highly granular, job-based salary survey will yield an accurate picture of the market. But incumbent data is personal data. It tells you how much Johnny, Mary and Sue are making, and it highly depends on their individual situations. In fact, we believe that there are many influencers for incumbent data movement, but the job isn’t one of them!

    In Birches Group surveys we capture incumbent data and salary range data. We use the ranges to “bookend” the incumbent data and express each percentile as a range. This reflects what is really happening in the market since you can determine the market range for a job or level, and the extent to which the incumbents have penetrated the range.

    We call this the “Market Footprint” and believe it is simply a better way to look at market data. We invite your readers to learn more about our approach by reviewing these articles from our website (

    Do Jobs Really Matter in Salary Surveys?
    Salary Survey Percentiles Exposed!
    Rethinking Market Comparisons: A New Definition of Compa-Ratio
    Rethinking Survey Comparators – Your Sector Is Not An Island!

    I agree with your position that the scenario described of collecting just salary scales is a pointless exercise. But structured properly with professional job matching, a salary survey that uses both incumbent data and salary ranges offers much more context and more stable data than incumbent data alone.

    Thanks for your excellent work to help inform and educate reward and HR professionals.

    • Dear Warren,

      Interesting points. I agree that seeing the ranges can give a useful complement of information, especially in markets where information is less readily available.

      But, just like any other data point, it needs to be taken with a pinch of salt :
      * not all companies target the same position to market
      * not all companies, even within the same industry, define market in the same way
      * not all companies have the same spread to their salary structure
      * and not all companies actually update their salary ranges on a regular basis to reflect the reality of pay in their organisation.

      A “salary range” around the pricing for a specific job is another thing and can indeed be very useful. It can be a nice complement to the percentile information, in that it helps define what companies consider to be acceptable pay for the job.

      Good food for thought ! Thanks for your comment 🙂

Speak Your Mind