Why your best and average performers earn the same, and how to fix it affordably

Every year a founder decides the increment. Everyone gets roughly the same percentage, give or take a rounding. It feels even-handed, it avoids arguments, and it takes an afternoon.

What it actually does is tell your strongest performer that their year was indistinguishable from the year of the person who did the minimum. They rarely say anything. They start answering recruiters.

Why businesses end up flat

Almost never because the founder believes everyone performed equally. Three other reasons, all practical.

• It cannot be defended. If two people get different increments, both will eventually know, and there is no written basis for the difference. Flat removes the conversation.
• The ratings mean nothing. Where nearly everyone is rated good, the rating cannot drive a distribution — so the distribution falls back to uniform.
• The budget feels too small to split. A modest pool divided twenty ways looks like nothing, so differentiating it feels pointless. This is the reasoning that most needs challenging.

The same budget, distributed differently

Differentiation is a distribution question, not a budget question. The pool stays what you can afford; what changes is who gets what share of it.

FLAT (what most businesses do)
  20 people, everyone the same share of the pool.
  Your strongest performer and your weakest receive
  an identical outcome.

DIFFERENTIATED (same pool)
  Top performers      ~1.5x the average share
  Solid performers    ~1.0x
  Below expectation   ~0.5x or nil this cycle

  The pool does not change. The signal does.

Rule of thumb: if the top and bottom outcomes are within a
few percentage points of each other, the exercise has not
communicated anything and your best people have noticed.

The multiplier matters more than the absolute amount. A difference small enough to be mistaken for rounding is not differentiation; it is flat with extra arithmetic.

Fix defensibility before distribution

The reason flat is chosen is that difference cannot be explained. So the sequence is: make it explainable, then differentiate.

1. 1

   Make the rating mean something

   If four out of five people are rated at the top, the rating carries no information. It does not require a forced curve — it requires managers to be asked what evidence supports the rating, and to be willing to record a middle rating without treating it as criticism.

2. 2

   Write the distribution rule before the ratings

   Decide the shares first, in the abstract, while nobody has a name attached. Deciding afterwards means arguing about individuals and conceding one by one.

3. 3

   Have managers recommend, not decide

   Recommendations from managers, moderated across departments so one generous manager does not consume the pool. This calibration step is what makes the outcome defensible across teams rather than only within one.

4. 4

   Say it out loud in the conversation

   "You are in the top group, which is why your increase is higher than the average" is a sentence most managers never say. Unsaid, the differentiation is invisible and buys you nothing.

Things that differentiate without touching fixed pay

Where the pool genuinely is very small, fixed pay is not the only lever, and some of the alternatives are better signals.

• A one-off payment rather than a permanent increase. Recognises the year without committing the business to it forever — useful where performance was exceptional but the role has not grown.
• Title or scope change, if the role genuinely changed. Meaningless if it does not, and people can tell the difference.
• First claim on development spend — the training, the certification, the conference. Cheap, and read as investment rather than consolation.
• Visible responsibility — running an initiative, mentoring, representing the team. Free, and frequently valued more than the money.
• Off-cycle correction for someone visibly below their band. Often the highest-return money you will spend, because it removes a specific grievance.

None of these substitutes for pay indefinitely. Used for a cycle while a structure is built, they buy time honestly.

The structural version

The durable fix is a grade and band structure, because it converts "why do I earn this?" from a negotiation into a position in a range. A strong performer moves up through their band each cycle; an average one moves slowly; someone at the top of their band gets a progression conversation rather than a pay one. That is the substance of compensation and rewards strategy work, and it usually costs nothing to establish beyond the effort of doing it.

It also protects you from the other failure mode. Without bands, the person who threatens to leave gets matched, word travels by lunchtime, and next year several people try the same. A structure gives you something to say other than yes or no.

What to expect the first time

Discomfort, mostly among managers who have never had to justify a rating. Expect two or three to recommend everyone at the top; that is the moment calibration earns its place.

Expect also that one or two strong people will be surprised to learn they were seen as solid rather than exceptional. That conversation is uncomfortable and far better had now than at a resignation. If performance has never been discussed honestly during the year, the increment conversation is the wrong place to start — which is why this work usually has to run alongside performance management rather than ahead of it.

And expect the change to be noticed. The first differentiated cycle is the one your best people have been waiting for without saying so, and the signal it sends is worth considerably more than the rupees involved — which is the whole argument for treating compensation and rewards strategy as a retention instrument rather than an annual administrative task.

Start where it costs least

If the whole thing feels like too much for one cycle, do the narrowest version: identify the three or four people you would genuinely be damaged to lose, and make sure their increment is visibly different from the average. Leave the rest of the distribution flat this year.

That is not a complete system and it is considerably better than nothing. It also demonstrates the principle internally before you have to defend a full structure, which makes the following year's conversation easier.

The manager who rates everyone the same

Differentiated pay depends on differentiated assessment, and the most common obstacle is a manager who will not distinguish between their people.

The reasons are usually decent. They believe the team succeeded together, they are protecting people they like, or they know that a lower rating requires a conversation they have not prepared for. The effect is that their strongest performer is paid the same as their weakest, which is a decision the manager has made without experiencing it as one.

It is worth naming to them in exactly those terms. Not as a rating policy but as a consequence: if everyone is rated identically, the person carrying the most gets nothing extra, and they will eventually notice. Most managers have not made that connection, and it changes the conversation from a form-filling obligation into something they care about.

What they need after that is evidence and practice rather than instruction. A manager who has kept no notes across the year genuinely cannot differentiate, because recall favours whoever was most visible recently. Asking for three specific observations per person before the cycle, gathered over a quarter rather than the week before, is what makes differentiation possible at all.

What it does not do is fix the underlying problem, which is that the business cannot yet explain why anyone is paid what they are paid. Sooner or later that needs grades and bands built from your own payroll, and the sooner the structure exists the less it costs to correct — which is the case compensation and rewards strategy work makes, and why it usually runs alongside role design and job descriptions rather than after it.