Kris Dunn’s recent post about online salary data for The HR Capitalist reminded me that HR often has a hard time responding to online salary data.  What used to be ‘secret’ info is now out in the open for anyone with a web browser to see!  How should HR address this issue?

The usual scenario goes something like this.  The employee brings a printout to your office showing she’s underpaid by at least $10,000.  Forget the fact that she found the information using your PC, your software, and your printer, how should you address this issue?  After all, if the info is on the web, it’s good stuff, right?  Not so fast!  You need to know a little about surveys before accepting or rejecting survey data.

Many of our clients have already heard this lecture and I’ve presented it more than once.  However, it bears repeating for those who may still be in the dark.  In short, you should think about developing a policy for dealing with online salary data.  Accepting or rejecting survey data should be based on the  quality of the data and little else.  Think about using the following quality standards before including any survey in your market-pricing process.

ONE:  Who provided the data?  Was it provided by HR (or comp) professionals or is it a compilation of whatever anyone wants to fill in on the survey website?  Generally speaking, the data provided by the HR pro will be more reliable than that provided by random individuals.

TWO:  What is the geographic coverage of the data?  Is the information for your particular area or for the entire US?  Here’s a good rule of thumb to follow:  the data should reflect your recruiting area for a given job.  If you recruit locally for a given job, use local data.  If you recruit regionally, use regional data.  If you recruit … well, you get the idea.

THREE:  What is the effective date of the data?  In other words, when responding to the survey, what payroll date was to be used?  If you can’t find the effective date, don’t use the data – it could be old stuff.  Further, if the effective date is more than a year old, be careful about using it for highly-competitive positions without at least aging the data.  If the data is more than 2 years old, find something more current.

FOUR:  How many companies are represented in the data?  How many employees?  The real question here is how comfortable do you feel in setting a pay rate based on what you see from the survey?  If the data is from a handful of companies, you may want to find an additional survey with more participants.

FIVE:  Finally, how does the survey data fit with other sources?  I prefer to use at least two and preferably three surveys when market pricing a job.  Why?  To minimize the ‘bias’ of the surveys.  In other words, one survey may be comprised of primarily manufacturing companies while another may be primarily white collar.  Averaging the results will help minimize the type of companies represented in the survey.

In summary, consider developing a brief policy on the use of online survey data.  Not all online data is ‘bad,’ but there’s much of it that I would not use in setting pay rates.  If employees and supervisors clearly understand the company’s quality requirements for survey data used in determining pay rates, you’ll have fewer meetings about online survey data.