If you've been in compensation long enough, you've heard it before: Skills-based pay is the future. Organizations should pay people based on what they can do, not the job title they hold. Makes sense, right? The problem is, we've been saying this for roughly 50 years, and it still isn't normal practice.
It's not a lack of desire. Over the past decade, large organizations have invested heavily in building skills taxonomies. Hiring teams have become more intentional about calling out required skills in job postings. And compensation professionals have been genuinely trying to make this work. Yet the vast majority of organizations still rely on job-based pay structures.
I call it the reward unicorn. It's this thing everyone talks about that just never seems to materialize at scale. And there are solid reasons why.
The infrastructure was never built for skills
Here's the core problem: the entire infrastructure of compensation market data was built to price jobs, not skills. Survey vendors standardize job definitions. Benchmarking tools map titles across companies. HR systems organize pay by role and level. That machinery is why, when one company calls a role a "Product Manager" and another calls it a "Senior Product Lead," we can map them together and arrive at a reliable market rate. As an industry, we built this system. It works. The challenge is that skills never had the equivalent investment. Skills have lived inside job postings, resumes, and the instincts of hiring managers — bundled inside job titles, priced as part of the package, but never measured or benchmarked on their own.
Skills are genuinely harder than jobs to apply market value to. Some of them — leadership, judgment, communication — are nearly impossible to define consistently across organizations, let alone price. That's a real challenge, and it hasn't been solved.
But there's a meaningful category of skills where the definitions are clear, and the signal is strong: technical capabilities, tool expertise, and domain-specific certifications (including security clearances). These are increasingly named explicitly in job postings and where real pricing intelligence is now possible. That's the foothold. And from a foothold, we will build a flywheel: as more organizations track and report skills consistently, the data accumulates. As the data accumulates, pricing gets sharper. As pricing gets sharper, more organizations adopt it and the system gets more reliable. The same flywheel that built decades of reliable job benchmarking is just beginning to spin for skills.
Skills data tells a more complicated story
We can't wait for the flywheel to fully spin up. The challenge of compensating for key skills needs to be solved right now. The labor market is moving faster than job definitions can follow and traditional comp systems are struggling to keep up. AI skills are exploding in demand. AI-related skills are appearing more frequently in job postings, and according to our Compensation Best Practices Report, 61% of organizations have updated existing roles to include AI-related skills or competencies. Organizations are racing to hire people who understand multi-agent systems, machine learning, large language models, and data engineering.
But here's what we're not seeing: consistent pay premiums for those skills.
One organization is paying a significant premium for AI skills. Another isn't. In some cases, companies don't even know how much to pay for these skills because there's no market precedent. This creates two problems:
First, it's inefficient for organizations. You're making compensation decisions in the dark — which means you may be struggling to hire and retain talent in roles where skills are commanding a premium, while simultaneously overpaying in areas where the market has normalized. Without a clear view of what individual skills are worth right now, you can't tell which problem you have.
Second, it's unfair to employees. Every time you make a one-off exception about pay to bring in the right talent or to retain a key employee, you run the risk of introducing pay compression or pay inequity in your compensation program. By incorporating skills into your compensation strategy, you introduce the flexibility that you need to hire and retain that key talent without making each decision an exception. This leads to greater fairness and transparency and can serve to incentivize your current employees to develop the skills that are most important to your business.
Why technology (and time) have held us back
The reason organizations haven't solved this at scale is partly technical and partly organizational. On the technical side, the way we collected compensation data (surveys, slow cycles, manual updates) made it impossible to do this work at speed or scale. Collecting data takes months. Normalizing it takes more time. Building pricing models around skills added yet another layer of complexity. By the time your data was ready, the market had moved on.
But there's also a structural problem: the people defining skills — hiring managers writing job descriptions, recruiters filtering candidates — operate largely outside of the compensation function. Different teams and different playbooks with no shared language. Comp professionals were trying to price skills that other parts of the organization were defining inconsistently.
The pace of change makes this even harder. Consider prompt engineering. In early 2023, it was a genuinely rare and valuable skill, and some organizations were paying significant premiums to hire people who could effectively work with large language models. By mid-2024, it had shifted from "hot skill" to "expected baseline" for anyone in an AI-adjacent role. That's a complete lifecycle in under two years.
Now think about how traditional compensation surveys work: vendors define a job and then start collecting participation data 12 to 18 months before you're using the data to price your roles. That means by the time a "prompt engineering premium" shows up in your survey data, the market has already moved on. The skill you're pricing is no longer the skill people are hiring for. The slower your compensation system, the further behind you fall.
But this is finally changing. The technology has advanced — and so have the data sources, and even the law. Here's what's different now:
- Pay transparency legislation has changed what's visible in the market. Companies are now required to post salary ranges alongside detailed job requirements. For the first time, you can observe skills and pay together, at scale, in the open.
- Large Language Models are being used to parse job postings with enough precision to identify required skills and key job attributes, such as experience level, education requirements, and industry context. While great for a text-based process, there are caveats to using general-purpose LLMs for benchmarking jobs.
- However, combining structured data from job postings with HR-reported salary information and compensation-specific machine learning enables us to determine how much each skill and certification impacts pay for a given role.
- And it can happen continuously as job definitions shift. Not waiting 18 months for jobs/skills to be reporting in a survey. Even if you're not ready to adjust your ranges today, you can keep a continuous pulse on what the market is actually valuing.
Check out the Compensation Pro’s Guide to AI
What forward-thinking organizations are doing now
If you're ready to move beyond job-based pay (even if you're doing it gradually), here are two concrete things to start with:
1. Map skills to your job architecture with specificity.
Start by identifying which skills are required for each role — not a generic list, but the specific capabilities that drive performance in that job. This becomes the foundation for everything that follows: clearer hiring criteria, more defensible evaluations, and a way for employees to understand what they need to develop to advance.
2. Look at your internal compensation through a skills lens.
Run an analysis: Are there critical skills your business needs that you've consistently struggled to hire or retain? That pattern is often a compensation signal; it suggests you may be paying below market for exactly the capabilities that matter most to your business.
These two suggestions aren't about completely overhauling your pay structure overnight. They're about building a foundation for skills-based thinking. Most organizations haven't done this yet.
A new kind of solution is here
This is where it gets exciting. Over the past couple of years, the breakthroughs in how we understand and model skills have been genuinely meaningful. Here at Payscale, we’re building a skills differential engine that can help you understand exactly how much to pay for a skill or a combination of skills within a specific role. It takes all of this complexity and turns it into something actionable: What should I actually pay for this?
We're launching it this summer, and the early feedback from compensation experts has been exactly what you'd hope: That's the number. That's what it actually takes to hire for this skill in this role.
This is part of our broader vision to give compensation professionals what they need when they're pricing jobs: not just the market price for a job, but a 360-degree view of the intelligence around it. See how in-demand that role is. Understand the skills that matter. Look at combinations that create premium value. Then make an informed decision about where you want to land on pay.
The reward unicorn might finally be becoming real.
Looking ahead: What's changing in compensation
Skills-based pay was always the right idea. It makes sense from a fairness perspective (pay for what people actually contribute). It makes sense from a business perspective (pay for what the market values). And it makes sense from a talent perspective (give people clarity on how skills map to compensation).
We've had earlier versions of this work. What's changed is the quality of the inputs. HR-reported salary data and skill requirements extracted directly from job postings give us a more trustworthy base.
And with AI accelerating workforce change (making some skills more valuable, retiring others, creating entirely new categories of work), organizations that can understand and price skills dynamically will have a real competitive advantage.
The technology is here. The data is here. The frameworks are coming.
Here's what you need to do next
Listen to the full Comp and Coffee podcast conversation with Ruth Thomas, Payscale's Chief Compensation Strategist, as we dig into the 50-year history of skills-based pay, why the disconnect between skill demand and pay is growing, and what the future actually looks like. Ruth and I also discuss what we're building at Payscale and the breakthroughs that are making this possible.
But that's just the beginning. Over the next month, we're hosting live events where you can dig into this with our team. Get your specific questions answered, see the skills differential engine in action, and learn exactly what you can do right now to move toward skills-based compensation, even if you're years away from a full overhaul.
Stay tuned for those announcements.
In the meantime, start with those two things: map your skills clearly, and audit your internal pay through a skills lens. The organizations that start building this foundation now will be the ones ahead of the curve when the pace of change accelerates.
