Off-cycle adjustments: a decision framework

Off-cycle raises are where consistency quietly disappears. A simple framework keeps urgency from becoming precedent.

Off-cycle adjustments exist because annual cycles are too slow for some situations: a genuine retention risk, a role that expanded materially mid-year, a correction to a placement that was wrong from the start. They are also where compensation consistency most often erodes, because each one is handled under time pressure and outside the process where comparisons naturally happen.

Sort the request first

Most off-cycle requests fall into one of four categories, and they deserve different treatment.

• Correction: the person is paid below where your own structure says they should be. This is not really an exception; it is fixing an error, and it should be the easiest to approve.
• Scope change: the job materially grew without a promotion. Legitimate, but ask whether the honest answer is a level change rather than a raise.
• Retention: a competing offer, or credible risk of one. The most dangerous category, because it rewards the act of looking.
• Market movement: the band itself has fallen behind. This is a structure problem wearing an individual's name.

Classifying the request before evaluating it prevents the most common mistake: treating a structural problem as an individual accommodation. If three retention requests arrive from the same function in a quarter, you do not have three retention cases. You have a band that is out of market.

The checks that should run every time

Whatever the category, the same context should be visible before a decision.

Where does the proposed new number sit in the band, and where did the current number sit? Who are the internal peers, and what does this do to the relationships with them? Does it compress against the level above? What happened the last three times a similar request came from a similar situation? And what is the aggregate effect if this becomes the standard answer for this category?

The question is never only "is this fair to this person?" It is also "is this the answer we will give the next four people?"

On counteroffers specifically

Counteroffers deserve a stated position because ad hoc handling is corrosive. Whichever way you lean, the important thing is consistency: if one person's outside offer produces a 15% increase and another's produces a polite decline, the difference will be attributed to favouritism unless the reasoning is documented and defensible.

A reasonable default is to decline to match on principle, but to treat the offer as information: if a credible outside offer is well above your band, that is market data, and the right response may be to review the band for everyone rather than to move one person.

Recording the decision

Every off-cycle decision, approved or declined, should leave a record: what was requested, what context was reviewed, what was decided, and the reasoning in a sentence or two. Declines matter as much as approvals — a declined request that leaves no trace means the next similar request gets evaluated with no knowledge that the question has been answered before.

Over a year, that record becomes the most useful compensation artefact a lean team has. It answers "what do we normally do here?" with evidence instead of recollection, and it makes new members of the team effective far faster than a policy document ever will.