Compensation that doesn't leak

Compensation leaks.

The numbers leak too, but that's the smaller problem. What leaks is the belief that the system is fair, and it leaks through a hundred small holes, every one of which looked perfectly reasonable at the moment you made it.

The engineer who negotiated hard on the way in and is now paid 20% more than somebody better. The counter-offer you matched in a panic in March, now the highest salary on the team. The person who joined in the cheap year and never got corrected.

None of them were bad decisions in isolation.

Together they're a system nobody would have designed, and everybody can feel it long before anybody can prove it.

Say your market position out loud

Pick a position and state it.

We pay at the 60th percentile of the market for this role in this location. Or the 75th. Or below market with more equity.

Any of those is defensible. Not knowing is the only indefensible answer.

Say it internally, because without a stated position everybody assumes one, and they assume it from the single data point they happen to hold.

Somebody's friend at a bigger company earns 40% more, so this place underpays. Somebody else read a survey saying the opposite.

Both are now managing a private grievance built on nothing, and neither will ever raise it with you.

And decide it once, instead of per candidate.

A company without a position pays whatever each individual negotiation produces, which means it pays for confidence instead of contribution.

That effect lands unevenly across the people you employ. It also compounds every year.

Bands, visible internally

Every role has a band, visible to everyone inside the company. What each level pays, the range within it, what moves somebody through it.

Not published externally. Not a spreadsheet of individual salaries.

That's a middle position and it's a deliberate one.

Internal visibility does the work that matters.

It kills the private grievance, because a person can see whether they sit inside the band for their level, and if they do, the conversation becomes about level instead of about whether they're liked.

It disciplines managers, because a manager who wants to pay outside the band has to explain why, to somebody.

And it removes the negotiation premium at the offer stage. That's where most of the unfairness enters the building.

Full external publication does something different, and what it buys is mostly recruiting brand plus a strong signal about your own confidence.

What it costs is optionality. Published numbers are hard to move down, hard to vary by market, and they start an argument with every candidate before they've even met you.

Companies that do it well are usually companies whose entire hiring pitch includes it. If that isn't your pitch, internal is where I'd stop.

The version I'd avoid entirely is bands that exist and stay secret.

Worst of both. All the constraint, none of the trust, and everybody learns the bands exist eventually anyway.

The counter-offer problem

Somebody comes to you with a competing offer. It's 25% higher. They're good, you can't lose them this quarter, and you have a band.

Match it and you've taught the whole team the mechanism for a raise. They will use it.

Refuse on principle and you lose somebody you needed.

Both are real, and the choice is genuinely uncomfortable.

What I'd do, and what I'd tell a founder to do, is decide the answer before the conversation happens.

If this person is currently paid correctly for their level, then the competing offer is either a different level of job or a market you've chosen not to compete in, and the honest answer is that you can't match it and you'd rather they went with your blessing than stayed resentful.

If they're paid incorrectly, fix it immediately, tell them it was wrong, and understand that you've just learned your correction pass isn't working.

What you can't do is match it quietly and hope.

It won't stay quiet. And the second person through that door will have watched exactly what happened to the first.

Equity, explained properly

Most engineers hold equity they don't understand. Most companies prefer it that way without ever having decided to.

The minimum honest explanation, in writing, at offer time.

How many shares, out of how many total. The denominator, not just the percentage, because the percentage moves and people don't expect it to. The strike price. The vesting schedule. The exercise window after leaving. The tax exposure exercising creates.

Then the current preferred price, and a plain statement that it can go to zero.

That last part is the one founders resist, and it's the one buying you credibility.

An engineer told plainly that the equity is a lottery ticket, who takes the job anyway, is a completely different person from one who was allowed to believe it was a bonus.

The specific cruelty to avoid is a ninety-day exercise window presented as standard, with no explanation attached.

Somebody leaves after four years, and to keep what they earned they have to find a large amount of cash within three months and pay tax on a gain they can't realise.

Extending the window is a real, cheap kindness. If you can't, say so clearly at offer time. Never at exit.

The annual correction pass

Once a year, before any performance or promotion conversation, look at everyone's comp against the band and against each other, with no names attached to the first pass.

You're looking for three things.

People below band for their level, who get fixed on the spot without being asked to negotiate for it.

Clusters where the same level is being paid meaningfully differently, which is where the negotiation premium accumulates.

And any pattern in either that tracks something it shouldn't. When somebody joined. Who hired them. Which manager they've had.

That last check takes thirty minutes, and it's the one keeping a small problem from turning into a legal one.

Fixing an inherited inequity costs money, and it isn't optional.

The alternative is that the person discovers it themselves, usually in a conversation with a colleague. At that point you've stopped correcting a number and started rebuilding trust with somebody who has just learned they were underpaid for two years while nobody said a word.

How the correction actually gets made

The part nobody writes down. It's why a first-time manager trying to fix a pay gap by asking their own boss discovers the other parties one at a time, badly.

It takes acknowledgement from everybody involved before it takes money. HR, finance, the department.

Then the adjustment has to be baked into a budget, and how much elasticity that budget has this quarter against next is genuinely specific to your company.

A collaborative move. Treating it as a unilateral one is how it stalls.

Which leads to the thing that surprises people. Movement matters more than arrival.

The instinct says a partial correction is an insult, so either you fix it fully today or you've done nothing worth doing.

The opposite is true. Somebody watching the gap close on a stated schedule understands that budgets are real and that a person is genuinely working on it. Somebody who sees nothing at all concludes they weren't worth the trouble.

Which only works if you tell them. So tell them. The intention, the plan, and the pressure. All three.

The pressure is the one people leave out, and it's the one making the other two credible.

Nobody in a company believes things happen magically, and being handed the constraint you're working against is what lets a person reason about your decision instead of merely receiving it.

The tell

You'll know the system is leaking when a manager starts a comp conversation with "let me see what I can do."

That sentence means the outcome depends on the manager's advocacy instead of the person's level.

Once that's true, the best-paid people in your company are the ones with the most persuasive manager.