Real-Time Market Compensation Data Is Not A Panacea

Written by Enrique Esclusa, Co-Founder & Co-CEO, Assemble

How do you decide what to pay an employee?

This is a simple and important question with a complex answer. With a volatile labor market, global inflation, and the continued rise of distributed work, it has never been more difficult to answer this question.

Traditionally, HR teams have primarily relied upon two sources of data to inform decisions: salary datasets, compiled and distributed by third-party consultants such as Radford and Mercer; and candidate data collected by recruiters asking prospective employees about their salary expectations.

Now, advances in legislation and technology are changing the way companies think about and handle compensation. In many states such as California, it is now illegal to ask candidates about their salary. Additionally, the rise of pay transparency laws have open-sourced new insights about what organizations are paying for various roles. Finally, technological advances have enabled web-based datasets with lower data latency, even enabling some pay data to be delivered real-time.

While access to real-time market data is growing, companies must remember that data alone does not decide what to pay an employee. In other words, real-time market data is not a panacea. And if misused, real-time market data can actually lead to costly and regrettable mistakes.

Data as an advisor, not a dictator

Companies must view market data for what it is: a factor that informs a broader decision-making process, not the sole determining factor that dictates decisions. Companies must consider the context when making decisions, such as: company industry, stage, culture, budget, and internal pay equity, as well as job requirements and employee performance.

Market data providers generally report data in an aggregate format that does not provide underlying context explaining why employees are paid what they’re paid. And because datasets are standardized for use by all customers, they do not adjust to a company’s compensation philosophy or program.

What does it mean when a market dataset says the 25th, 50th, and 75th percentiles are $90,000, $100,000, $120,000? How should a company use this information to decide what to pay an employee?

Relying on data alone, real-time or not, is insufficient and inappropriate. It can lead to missing important considerations and oversights when making compensation decisions that can be expensive and difficult to unwind.

No market dataset is perfect

At a high level, market datasets collect data from companies about their employee compensation, then aggregate and structure the data into a usable format that companies can use to inform their compensation programs. However, current aggregation methods offer many challenges.

Market datasets are often incomplete, especially as searches or data “cuts” become more specific and narrow. Sample sizes matter. It is not uncommon to find datasets with sufficient data points to populate only part of a career ladder, but not all of it, or for one location like New York, but not another like Toronto, Canada. This results in data gaps that surface skepticism around the trustworthiness of the dataset.

Then, there’s the question of matching employees to the jobs. Job matching is a difficult task for participating companies, and a key pressure point for creating high-quality data. On the one hand, companies generally lack the in-house expertise and readily available data to match employees against an external job structure. (Companies often lack the bandwidth or patience to do this at all, let alone do it well, and often find themselves rushing through the process.) On the other hand, market data aggregators lack important context about each employee (which may be available nowhere but on a manager’s mind) to accurately match employees against the job structure, at least without heavy support from each company. Proper and well-vetted job matching is critical to create a reliable dataset. Doing this at scale and real-time makes it a much more difficult task, resulting in datasets that are lower-latency but less accurate. As the saying goes: “garbage in, garbage out”.

The risks of using real-time data

There are important risks with market data being real-time that companies must be aware of. Relying on ever-changing datasets that can fluctuate up and down almost daily can lead to more confusion, lack of trust in how decisions are made, and pay inequities within a company. This can result in biased, inaccurate, and inconsistent compensation decisions.

For example, a company may hire an employee based on today’s real-time market rate, paying a competitive wage. Fast forward a few weeks, and the same dataset now shows that this employee is “underpaid”. What happens when the market rate increases significantly again in a few weeks or months? What happens if it decreases? Should the company make real-time adjustments in response to real-time data? Continually changing an employee’s compensation to match the market rate is extremely problematic operationally, and for little gain.

Setting initial pay decisions based on a market rate that is changing real-time may also inadvertently create pay inequities that are difficult and costly to unwind. Imagine a company hires Jane and Amanda at today’s market rate of $100,000. Shortly thereafter, the company hires Andrew and Juan at the new real-time market rate of $110,000. The company now finds itself paying these two male employees more than its female employees. Closing the gap would immediately cost the company an annualized $20,000, while doing nothing potentially exposes the company to litigation risk and employee frustration.

Last but not least, truly real-time datasets are a violation of U.S. safe harbor laws which seek to prohibit collusion among companies to engage in wage fixing. As of this writing, these safe harbor laws render participation in real-time datasets illegal. Instead, companies are only legally allowed to participate in market datasets organized by a third party, with at least five data points per job surveyed, and with data that is at least 90 days stale. Real-time market datasets are in clear violation of these laws, opening any participants of these surveys up to legal risk.

Pay transparency laws could complicate things further

The United States and Europe are both experiencing a rise in pay transparency laws that in some cases, like in California, will require companies to disclose compensation bands to employees.

Companies relying exclusively on market data will still need to comply. If basing their compensation bands on an ever-changing market rate, the operational complexity and administrative overhead to maintain these compensation bands will be extremely intensive. The work itself would be complex, but more importantly, it will be challenging to communicate information to employees with clarity and in a way that builds trust. Surely, employees will be confused with continually changing compensation bands, and will be more likely to ask the company uncomfortable questions. If these are not answered satisfactorily, distrust in the company may grow, which could lead to reduced productivity and employee attrition.

How to properly use real-time data

To be clear with my message: market data is important and should not be dismissed. Emerging real-time datasets offer a key benefit: low data latency, which can be helpful to companies, especially in competitive labor markets and inflationary environments like the one we find ourselves in. However, companies must be thoughtful in how they leverage these low-latency datasets.

First, companies should be careful to remain compliant with U.S. safe harbor laws, and only use data that is at least 90 days old. Second, low-latency data should be treated as all other data: as an input into a broader decision-making framework. It is here where companies should be very intentional and thoughtful. Not only should companies have a well defined compensation philosophy that incorporates how compensation decisions are made, but it should also capture how and when market data will be used.

Market data is a critical component to make compensation decisions, but it alone is not the solution. Companies must still have a framework for how market data will be used to inform — not dictate — how employees are paid.

Companies must understand the second and third order effects of using real-time market data, navigating the limitations and downstream impacts of a continually changing market rate. As pay transparency laws continue to become more commonplace, the need for a decision-making framework becomes even more important, especially if these laws result in more frequent changes to market rates. Lower latency market data is an emerging and important evolution in how companies make compensation decisions, but be clear in your understanding: real-time market data is not a panacea.