Which retention interventions actually move the numbers

A mid-cycle comp correction can be approved and announced. A retention bonus shows up in a spreadsheet. An engagement platform ships with a dashboard and a quarterly report. So that's where the money goes: toward interventions you can point at in a review. It's also, on our read, largely the wrong place.

When we sorted the levers our clients reach for against who actually stayed, the correlation with spend was weak. The interventions that showed up again and again in the people who remained were quieter ones. A real scope. A manager worth working for. A next role the person could believe in. None of those fits neatly on a slide, and none of them generates a line item, which is exactly why they get starved of the attention they need.

The split isn't academic. A GCC that spends against the legible levers can run its retention numbers flat for a year and never understand why the money isn't moving them. The figure it's trying to move is being decided somewhere the spending never reaches.

Comp corrections and retention bonuses do real work in one narrow case: when someone was genuinely, provably underpaid against the market. Applied there, they hold, and they should be paid without hesitation. Applied as a reflex to every resignation, they teach the organisation that the way to get paid is to threaten to leave, and they do nothing at all for the person whose problem was never the number.

The larger part of what walks out the door is walking for reasons money can't reach. A scope that quietly shrank. A manager who never gave the work air. A promotion that kept arriving next cycle. Spending cash against those problems is expensive and it doesn't move the figure, because the figure was never about cash. The read is to diagnose first, then spend on the lever that matches the diagnosis.

The visible levers also carry a hidden cost: they buy only time. A retention bonus converts an exit into a delay. The person stays through the vesting date and leaves the week after, and the number that looked saved reappears 2 quarters later with a training bill attached. A lever that only defers the loss is a financing decision dressed as a retention lever.

The levers that consistently kept senior people were structural. Widening a scope that had been quietly trimmed. Moving a strong contributor out from under a weak manager. Making the next step real, with a date and a mandate. These cost management attention, not budget, which is precisely why they're under-used. Attention is the scarcer currency, and it doesn't show up on the retention line.

None of this is an argument against paying people fairly. Fair pay is table stakes, and paying below market is its own resignation letter, written in advance. It's an argument against treating fair pay as a retention strategy. The number keeps people from leaving for the wrong reason. It has never, on its own, given anyone a reason to stay.

There's a diagnostic discipline underneath all of this. Before any spend, the useful question is “what is this person actually leaving.” The answer is usually available in the exit conversation, or earlier, in the stay conversation nobody had. When the leaving reason is money, pay it. When it's scope, manager, or trajectory, no cheque closes the gap, and the organisations that keep reaching for the cheque are the ones whose retention numbers never move.

Every outside hire is a bet on incomplete information. You're guessing at fit; they're guessing at the reality of the seat. The boomerang removes half the guesswork. They worked the place, they left, they tested the alternative, and they came back with a clear read of what they were missing and what they weren't. That's a hire with the illusions already burned off, and it belongs in the retention conversation as much as the hiring one.

It shows in the ramp and in the tenure. There's no relearning the systems, the politics, or the parent's quality bar, and the returner has already priced the alternative and rejected it. The common outcome we see is a longer, steadier tenure than a fresh external join at the same level. A senior who left and returned joins with eyes open, and eyes-open is what stays.

Most GCCs leave this lever on the floor. A good exit is read as rejection, the alumni link goes cold, and 18 months later, when the person is quietly ready to come back, there's no open door and no one keeping the thread warm. The play works; it just dies of neglect. The centres that use it well run departures as relationships that merely paused: they part cleanly, stay in light touch, and make it genuinely easy to return without loss of face. Done properly, it retains someone who already voted with their feet, twice.

For senior roles, the clean 2-week departure is largely gone. Notice periods have crept toward 3 months and, at the top, past it: a defensive design by employers who learned that a long goodbye is a long chance to reverse the decision. The mechanics of leaving now take a full quarter, and a quarter is enough time for a lot to change. That runway isn't neutral. It hands the incumbent employer a window to work the counter, reassign the scope, and let second thoughts do their quiet work.

Read as a retention lever, the long notice is a gift, but only to the side that engages first. The employer who starts the stay conversation the week the resignation lands has 3 months to make a genuine case. The one who waits until the last fortnight is just haggling. The window rewards whoever moves early, and mostly that's the current employer, not the new one. This is why so many resignations that looked firm on day one soften somewhere inside a 90-day notice, and why the counter-offer window widens exactly as the runway lengthens. (It's also why the counter itself, once it lands, mostly still fails to hold, a separate pattern we cover in counter-offers are biting again, they still don't hold.)

For the hiring side, the discipline flips. Keeping an accepted candidate genuinely warm across the wait becomes part of the job, done through connection, so the seat they said yes to stays real in their mind for 90-plus days. Treat the gap between offer and start as the most fragile part of the process, because with notice periods this long, it is. A retention strategy that stops at the people already inside the building misses the ones whose decision is still, technically, reversible.

The centres that hold their senior people spend their budget differently. They diagnose the leaving reason before they reach for a lever, and they pay cash only where the problem is genuinely cash. They protect scope and move people out from under weak managers before those become resignation reasons. They run exits as paused relationships, so the alumni list stays a live re-hire pool. And they treat the long notice period as a retention window they open on day one, on both sides of the move.

None of that shows up well in a quarterly deck, which is the honest reason it stays under-used. The levers that move the number are cheaper than the ones that don't, and harder, and quieter, and they ask for management attention rather than a budget line. A retention programme that keeps buying the legible levers will keep reporting activity and keep losing the people it needed most. The work is to spend against the diagnosis, and to be patient enough to let the quiet levers do what the loud ones never could.