Performance Is a Signal. Not a Diagnosis.
A rating or number drops. What can it tell you, what can't it, and how do you prepare for the conversation before you decide anything?
The rating is real. The explanation is a guess.
Imagine a customer-support lead whose quarterly rating has slipped from "exceeds" to "meets". Her manager has a review in two days and a theory already forming: she's lost her edge.
Nothing has been checked yet. The rating is real. The explanation is a guess.
That's the gap this article is about. A performance number tells you that something is different. It doesn't tell you why.
The short version: a fall in a number or rating is a signal that something changed. It isn't a diagnosis of the person. Before you decide anything, check what the signal measures, look at what changed around it, and find out what the person sees. Then open with a question, not a conclusion.
What a performance signal can tell you
A signal does three useful things. It says that something moved. It says when. And it says how far, against whatever the measure compares to.
That's worth having. It's a reason to look.
Keep the wording precise. "Her rating fell from exceeds to meets this quarter" is an observation. "She's lost her edge" is an interpretation. Our guide to Behaviour Intelligence explains why that difference matters before any people decision.
What it can't tell you
It can't tell you why. A rating is the output of many things: the person, the work, the targets, the tools, the team, and whoever did the rating. It squeezes all of them into one figure and drops the detail.
It can't tell you whether the change is new or has been building. One quarter is a moment. In Beyond the CV, we make the same point about every stage of understanding a person: each captures a moment, and the reading ages.
And a rating carries the rater. Scullen, Mount and Goff studied developmental ratings of two large groups of managers (2,350 and 2,142). Each manager was rated on three performance dimensions by seven raters: two bosses, two peers, two subordinates and themselves. The authors estimated what drove the ratings.
Idiosyncratic rater effects, meaning each rater's own rating tendencies, accounted for 62% and 53% of the variance in the two groups. The ratee's own performance, general and dimension-specific combined, accounted for 21% and 25% (Scullen, Mount & Goff, 2000).
These were developmental ratings of managers, not annual reviews, so we wouldn't assume the same split in yours. But it's a fair reason to ask how a rating was formed, as well as what it says.
Causes worth checking
A fall can come from several places at once. These are the main groups to check.
| Where to look | What to check |
|---|---|
| The measure | Did the metric, scale, target or rater change? Is the comparison group the same? A new rater or a stricter target can move a rating on its own. |
| The work | Did scope, volume, priorities or tools change? Did a target move while the deadline stayed put? |
| Dependencies and team | Did inputs arrive late? Did colleagues leave, move or change roles? Did the reporting line change? |
| Role and expectations | Do you and they agree what good looks like now? Has the role grown without anyone saying so? |
| Support and resources | Did coaching, handover, access or time change? Did a source of help disappear? |
| Skills and motivation | Do new demands outrun existing skills? Has interest shifted? Both are possible. |
| Life outside work | Not yours to probe. Make room for them to tell you, if they choose. |
The table isn't a ranking, and a signal can have more than one cause. Skills and motivation are on the list because sometimes they are the cause. They come near the end because they're the hardest to check from outside and the easiest to assume.
Prepare before you open the conversation
You don't need much. A page of notes will do.
Write the observation. One or two sentences that someone who disagrees with you could accept. Leave the explanation out.
Draw the timeline. Place the change against the person's own earlier pattern, not only against the team. Mark anything dated that happened nearby: a reorganisation, new targets, new tools, a new manager or client. Look for what else moved at the same time.
List two or three explanations. Include one that involves you, such as an expectation you set or a priority you changed. It's easier to ask about it if you've already named it.
Check what you can without asking. How the measure is defined. What the workload looked like. Who the person depends on. Some of the answer is already in the records.
Plan your first question. The first question sets the tone. Write it down, and read it as the person would.
Open the conversation
The order matters. Say what you've seen. Say that you don't yet know why. Then ask.
Here is one way to put it: "Your resolution times and your rating both moved this quarter. I don't know what's behind that yet, and I'd like to hear how the last few months have looked from your side."
That opening makes no claim about the person. It names the signal, admits the gap, and hands over the floor.
Then listen. Take notes. Test what you hear against your timeline, and ask what they'd change if they could. You may hear something that fits one of your explanations. You may hear something you hadn't thought of. Both are useful.
Resist offering a conclusion before they've finished. We'd rather hear an account than a defence, and a conclusion invites the second.
Then decide
Now the decision. You'll know which parts of your reasoning are evidence and which are still hypothesis.
Sometimes the answer is training, because the conversation showed a real skills gap. Sometimes it's a clearer target, a lighter workload, or a different kind of support. Sometimes it's to wait and look again at the pattern. You'll be choosing among them knowing why.
Some situations can't wait, such as safety or legal concerns. For those, act first and understand alongside. Most performance signals leave room for a few days of understanding first.
If the conversation points to something around the person rather than in them, the next question is what to change. We take that up in What If the Behaviour Isn't the Problem?
Opptimise helps managers understand what sits behind a change in performance before they decide what to do.
A performance signal is where understanding starts. Better understanding comes before better action.
Sources
- Scullen, S. E., Mount, M. K. & Goff, M. (2000). Understanding the latent structure of job performance ratings. Journal of Applied Psychology, 85(6), 956-970. doi:10.1037/0021-9010.85.6.956. Abstract reviewed.
- Behaviour Intelligence: Understanding Behaviour Before Deciding What to Do.
- Beyond the CV: What If Understanding People Became Continuous?.