The Value of Compensation Survey Data in Today’s World of AI and of Publicly Available Salary Data from Job Postings

With recent legislation requiring pay transparency in job postings, and increased use of AI technology to aggregate compensation data published on the web, we’ve seen a surge in apps that now offer “data scraping” from public posted job offerings. This aggregated public data may appear more “definitive” than the self-reported data previously available on the web (on sites such as Glassdoor.com). It also may be viewed as timelier than data from annual compensation surveys. While public data provides validation, it does not replace the value of annual compensation survey data to provide a full and accurate picture for setting pay ranges and pricing jobs.

Aggregated Public Data Has Limitations

These new data sources can be helpful. However, it is important to know their strengths and limitations. The challenges of working with these sources, as opposed to using robust compensation survey data, is that they provide only part of the compensation picture and may not reflect information that aligns with your company’s overall approach to setting compensation.

The challenges of working with aggregated public compensation data include:

  • The data are for individual open jobs when posted, rather than for a full suite of jobs across a job family.
  • When aggregated, algorithms look at job titles. The same title does not necessarily represent the same job at different companies or a common benchmark. Although postings will have a job description, it does not appear that AI compares the core responsibilities and requirements to match “like” job descriptions and group them.
  • Hiring ranges are not necessarily the full ranges of pay for roles. At Croner, we often advise clients to publish up to or just below midpoint for hiring ranges; the full range often is not published.
  • Alternatively, some companies do post the full range for a role, without revealing where the actual pay level is. Some of these ranges are broad bands, making it impossible to know the “real” range of pay for a specific role.
  • The data provide only a view into base salaries, and salary ranges differ based on each company’s approach to total rewards (e.g., lower salaries, higher bonuses, percentile positioning etc.). Data will not be available for total direct compensation such as bonus and target value of long-term incentive plan awards.
  • The companies picked by internal departments for the web-based data may not reflect your company’s peer set. For instance, if you are seeking competitive data on software engineers, it is very easy to look at MAANG (i.e., Meta, Apple, Amazon, Netflix, Google) ranges and use this information as the “market” for your company. Ideally, all internal employees should know their company’s philosophy and competitive framework, and it is likely not all companies’ approved compensation philosophies state that they set their pay levels to be competitive with MAANG.
  • Departments within a company should be on the same page with the data they use, and that can be a problem when relying on data aggregated from public information disclosed by individual companies.

Using public information disclosed by individual companies does have benefits. The aggregated data are helpful for cross-checking a company’s current salary range. It’s always good to validate with a variety of sources. And looking case by case gives a feel for at least part of the range for a selected job at a selected company – again a good way to validate. The data are “real time” in that they reflect current postings. In addition, the data are easily available to internal departments, such as Talent Acquisition and Finance, for compiling.

Annual Compensation Surveys Continue to Play a Critical Role

Compensation surveys continue to play a critical role in providing finely tuned and accurate compensation data, as well as providing the full picture of pay. Our survey participants and consulting clients use compensation surveys as their primary source of external compensation data. Publicly reported hiring ranges are recognized as supportive or complementary material to published compensation survey resources.

Compensation survey participation provides access to robust data along with other benefits:

  • Compensation survey publishers closely review the data and follow up with participants about their submissions at each step of the survey cycle. This ensures accurate, consistent and reliable compensation data elements across companies with common jobs in relevant markets.
  • Surveys provide a depth of industry-specific information not found in publicly aggregated data. At Croner, for example, we conduct careful research and writing of job descriptions, thorough follow-up and matching review of each submission with a focus on accurate job matching, extensive data cleaning, report writing with year-to-year analysis, data observations and “titles matched” resources.
  • Compensation surveys collect actual pay levels for actual incumbents in matched roles. This is not possible to discern from posted ranges in job listings.
  • Published surveys report on total cash compensation, total direct compensation and other total rewards elements that cannot be gathered systematically from the web. They provide a comprehensive picture for guidance and decision-making as opposed to selected data available from limited job posting data.
  • Participation in total rewards aspects of compensation surveys and related meetings (e.g., planning meetings, webinars, roundtables) gives participants and total rewards professionals important insights into industry practices and trends, establishes important industry relationships and ensures that the industry has a trusted source of reliable compensation data on a regular basis.
  • Compensation surveys follow the Safe Harbor guidelines, protecting major organizations from antitrust concerns. To summarize, these guidelines currently require:
    1. Management by a third party
    2. Data more than three months old
    3. At least five participants (companies) to provide data
    4. No individual participant’s data representing more than 25% on a weighted basis of that statistic
    5. Any information disseminated is sufficiently aggregated, so participants are not able to identify the compensation paid by any company
  • Results are offered in such a way as to review special cuts in the data as guided by each company’s compensation philosophy. This ensures that survey participants have regular insights into a broader base of data sets as well as special cuts.
  • Compensation surveys are cyclical, to allow time for participants to thoroughly submit accurate data, with most being conducted annually and some updated quarterly. While the data may be perceived as “old,” these cyclical surveys provide regular insights into external pay movement, allowing companies to age the data, bringing it forward to current levels.

We value data and believe that each type of data source can be helpful. As long as we know the strengths and limits of each, we can use them to ensure a company’s compensation remains competitive and equitable. Data from sources such as scraped public job postings can give a measure of support and validation. Compensation surveys provide the full picture of compensation across multiple roles, with the added benefit of ongoing education and insights.

For more information about The Croner Company’s compensation surveys, please contact Laurie Krashanoff, Manager, Account Relations, at +(415) 485-5521 or laurie@croner.com.

Challenging Times Lead to a Renewed Focus on Performance Management