Compensation Benchmarking: What to Pay, How to Decide, and Why Most Companies Get It Wrong
You’re hiring a software engineer. You don’t know what to pay them. So you guess. You offer 120k. They accept. Three months later, you’re hiring another engineer. Same role. Same level. You offer 140k. They’re excited. Now you have two people doing the same work, same level, different pay. The first engineer finds out. They’re furious. They leave. You’ve just created a retention problem by not doing compensation benchmarking. This is a common story. Companies don’t know what to pay. So they guess. They overpay. They underpay. They create internal inequity. They damage retention and morale. Compensation benchmarking isn’t complicated. But most companies don’t do it. And that costs them money and people. Here’s what to pay, how to decide, and why most companies get it wrong. Why Most Companies Get Compensation Wrong Most companies make compensation decisions based on: Reason 1: Gut Feel “I think we should pay them 110k.” Why 110k? Because that’s what feels right. Or because that’s what their last company paid. Or because that’s what the CEO paid themselves once. Gut feel is not data. Gut feel is bias. Gut feel leads to bad decisions. Reason 2: What They Asked For “They asked for 130k. Let’s offer 120k and negotiate.” But what should they actually be paid? Nobody knows. So you negotiate based on what they want, not what the market says. This creates inequity. The person who negotiates well gets more than someone equally qualified who doesn’t negotiate. Reason 3: What You Paid the Last Person “We paid the last engineer 125k. Let’s pay this one 128k.” But was 125k right? Did market conditions change? Is this person more or less experienced? You’re just copying the past without asking if the past was right. Reason 4: What Your Competitors Pay (Sort Of) “Google pays engineers 180k. We should pay 150k to be competitive.” But you’re not Google. Google has way more money. Google’s costs are different. Google’s market is different. You’ve compared yourself to a company that’s not comparable. Reason 5: Arbitrary Tiers “Junior engineers: 90-110k. Mid-level: 110-150k. Senior: 150-200k.” But where did these numbers come from? Are they based on market data? Or did someone just make them up? If they’re made up, they’re probably wrong. Reason 6: Infrequent Review You set compensation in 2021. Now it’s 2026. Market has changed. Costs have changed. People now make 30% more for the same role. But you haven’t changed compensation because you haven’t reviewed it. Your compensation is now significantly below market. Your people know it. They leave. All of these lead to the same problem: Compensation that’s not based on market reality. What Compensation Benchmarking Actually Is Compensation benchmarking means finding out what the market actually pays for a role. Not what you think the market pays. What it actually pays. This requires research. Good data. Regular updates. Compensation benchmarking answers these questions: What’s the market rate for this role in this location? What’s the range? (10th percentile to 90th percentile) What percentile should we target? (Are we a 50th percentile company or 75th?) How does our current pay compare? What should we pay? How Market Rates Actually Work Market rates are not fixed. They vary based on: Location A software engineer in San Francisco makes more than a software engineer in Columbus. Not because they’re better. Because cost of living is different. Demand is different. Competition is different. San Francisco: 180-250k Columbus: 130-180k Rural area: 100-140k Your location matters. Role and Level Senior engineer makes more than junior engineer. That’s obvious. But the distinction matters. Is this person: Junior (0-2 years experience)? Mid-level (2-5 years)? Senior (5+ years)? Staff/Principal (senior+ with scope)? Each level has a different market rate. Industry Software engineers make more in fintech than in nonprofits. Not because fintech engineers are better. Because fintech has more money. Fintech: 200-280k Tech: 150-220k Nonprofit: 80-120k Your industry matters. Company Size Software engineers at early-stage startups make less than at big tech companies. Early-stage startup: 120-160k Scale-up (50-500 people): 140-190k Large tech company: 180-250k Your company size matters. Experience and Specialization A Python engineer with 10 years experience makes more than a Python engineer with 2 years. A DevOps engineer specializing in Kubernetes makes more than a generalist. Experience and specialization matter. The Right Way to Do Compensation Benchmarking If you want to get compensation right, here’s how. Step 1: Define the Role Clearly “Senior Software Engineer” is too vague. Be specific: What do they actually do? What’s their scope of responsibility? What level of seniority? What specialization (if any)? Example: “Senior Backend Engineer, 5+ years experience, leading backend architecture for a 500-person SaaS company, location: US remote with occasional SF travel” Step 2: Research Market Data Use multiple sources: Free sources: Levels.fyi (ask people what they make) Blind (anonymous salary data) Glassdoor (salary reports) LinkedIn (salary insights) Payscale (self-reported data) Paid sources: Radford (expensive, comprehensive) Mercer (expensive, comprehensive) Salary.com (affordable, reasonable data) PayScale Enterprise (subscription) Equifax (subscription) Primary research: Ask your network (what do you pay for this role?) Talk to recruiters (what are candidates asking for?) Look at job postings (what are competitors offering?) Don’t rely on one source. Use multiple sources. Look for patterns. Step 3: Calculate the Range From your research, you should see a range. Example for a Senior Software Engineer in San Francisco: 10th percentile: 160k 25th percentile: 180k 50th percentile (median): 210k 75th percentile: 240k 90th percentile: 270k The range is 160k-270k. Most people fall in the 180k-240k range. Step 4: Define Your Positioning Where do you want to be in that range? Are you a startup competing for talent? Pay 75th percentile (240k). Are you profitable and focused on efficiency? Pay 50th percentile (210k). Are you early-stage with limited budget? Pay 50th percentile for critical roles, 25th percentile for others. But make a deliberate choice. Not a default. Not a guess. Step 5: Set Bands Set salary bands for each role. Example band for Senior Software Engineer: Minimum: 200k Midpoint: 230k Maximum: 260k This


