Contamination: feedback vs reward

I stood in front of a room in New York, then Hong Kong, then Bangalore, explaining how the new system worked. And in one of those rooms, somebody put their hand up and asked a question I still think about now.

“Can I agree to be a three?”

I asked what he meant. He said: you’ve just told me that seventy per cent of the organisation is going to be a three, and twenty per cent will be a one or a two, and only ten per cent will need help. So can I sign something at the start of the year that says even if I’m brilliant, I don’t want to be a one or a two? I’ll just be a three, and I don’t have to do all of this.

My first instinct was to say no. But underneath the cynicism, he was doing a straightforward cost-benefit calculation. He knew that seventy per cent of the time he’d land on a three regardless. He was quite happy to offset the twenty per cent chance of being rated a one or a two, in exchange for disengaging from the process entirely. He’d do his job, and do it well. If anyone genuinely thought he was slipping towards a four or five, he’d re-engage. Until then, he wanted to be left alone.

That conversation was symptomatic of something much wider. People don’t engage with performance management because they don’t see any value in it. They just want to be a three.

Skill was never the problem

I’ve spent twenty years being asked the same question in different forms. Can we do feedback? Of course. Do we teach a model, whichever one an organisation has chosen? Yes. Do we train people to give and receive feedback well? Yes. Do we redesign the forms, restructure the conversation, insist every discussion follows a set structure?

I’ve done all of it. None of it moves the needle on its own.

A lack of skill isn’t the problem. It’s a necessary condition, but not a sufficient one. People do need to understand feedback, and they do need to know how to have a good conversation. But that alone doesn’t change behaviour, because the conversation itself is carrying a weight it was never designed to bear.

Think about what a single one-to-one is expected to do. Give feedback. Receive feedback. Talk about development. Talk about reward. Talk about promotion. Cover compliance. Provide evidence. Manage engagement. Improve motivation. Challenge people. Give them a voice. All of that, in one thirty-minute conversation. No wonder it doesn’t work. The conversation is overburdened.

I’ve boiled this down to three structural reasons these conversations don’t change behaviour. The first, and the one I want to focus on here, is contamination.

What contamination actually is

Feedback exists to land one of two messages. Either this behaviour is effective and should continue, or this behaviour isn’t effective and needs to change. That’s it. Feedback is information about behaviour and its impact.

The problem is that we contaminate that process with something else: reward.

For good organisational reasons, most functions have decided that feedback conversations and reward conversations should happen in the same meeting, or close enough together that they blur into one. As a manager, I’m being asked to talk about somebody’s behaviour, and in the same breath, or the same cycle, I’m being asked to have a conversation that feeds directly into their reward, their bonus, their promotion. Those are two fundamentally different conversations, and we try to run them as one.

I’d wager that for most people reading this, at least one meeting a year does exactly this. And that’s a problem, because we’re asking a conversation about behaviour to also carry a conversation about finance.

This has two visible consequences.

First, managers start reducing the developmental feedback they give to their high performers. If I’m telling the organisation that someone deserves a significant pay rise and a promotion, but the form I’m filling in lists a string of development areas for that same person, it sends a contradictory message. So I dial down the developmental feedback, to keep the message about reward clean.

Second, the people receiving feedback become more defensive, for exactly the same reason. If a meeting covers both my reward prospects and areas I haven’t done well, I’m going to push back on the feedback, because I know it’s connected to what I get paid. I’ll find reasons it shouldn’t count. Not because I’m being difficult, but because the two things have been fused into one conversation and one outcome.

Contamination made visible

Go back to that forced distribution model. Seventy per cent of people sit in the “three” band. That band spans the twenty-first percentile of performance all the way to the eighty-ninth percentile. Genuinely different levels of contribution, paid the same, rated the same.

That’s contamination made visible. People look at that and think, again, can I agree to be a three?

And once people see that, discretionary effort disappears. If working hard enough to be at the sixtieth percentile gets you the same outcome as working hard enough to be at the twenty-fifth, there’s a great deal of effort that simply isn’t worth giving. Nothing about the system makes it worthwhile.

What actually helps

This is genuinely simple to describe, and genuinely hard to do. Separate the documentation that feeds reward from the documentation that feeds development. Have separate conversations for each.

To be clear, performance across the year should still be one of the things considered when reward is discussed. That’s reasonable. But the two sets of conversations need to be distinct, so people know where they stand in each, and so they can engage with developmental feedback honestly, without immediately calculating what it means for their pay packet.

When that separation is missing, people don’t disengage because they’re lazy or cynical by nature. They disengage because, somewhere along the way, somebody worked out that the honest answer to “what do I get from fully engaging with this process” was: not much, and quite possibly a smaller bonus. Once that calculation has been made, no amount of training in giving better feedback will undo it.

That’s the first of three structural issues. The next is value: whether people can actually name what they personally get from their objectives, beyond the fact of being paid. More on that shortly.