Most managers think their direct reports avoid feedback. They don't. They avoid one specific thing: feedback that arrives without warning, about something they thought they were doing fine.
Walk into any 1:1 in the average mid-sized company and the same dynamic shows up. The employee is calm and engaged when the conversation is about work. They go quiet, defensive, or preemptively apologetic the moment they sense the manager is about to deliver feedback they didn't expect. The body language has nothing to do with how harsh the feedback is. It's about whether they saw it coming.
The thing employees actually dread is unpredictability. Their relationship with their manager has been operating on hidden information, and now the hidden information is about to arrive in concentrated form, in a meeting they can't reschedule, with consequences they can't fully anticipate. Surprise feedback violates the most basic working assumption people have about their jobs: that they have a roughly accurate model of how they're being evaluated.
This post is about why surprise is the actual failure mode in performance feedback — not the feedback itself — and what consistent communication does to fix it.
Psychological safety and the cost of unpredictability
Amy Edmondson's research on psychological safety, conducted at Harvard Business School over the past three decades, established a counterintuitive finding: high-performing teams are not the teams with the fewest interpersonal risks. They're the teams where people feel safe enough to take those risks visibly — to disagree, to ask questions, to admit they don't know, to flag problems early.
The mechanism isn't comfort. It's predictability. People take risks when they have a calibrated sense of what will happen if the risk goes wrong. When the social environment is predictable — when feedback follows from work in legible ways, when the manager's reactions are recognizable in advance — the cost of speaking up is bounded. People speak up.
Surprise feedback systematically breaks predictability. When an employee learns in November that something they did in March was a problem — the compounding cost we covered in detail — the implicit message isn't "this thing was a problem." It's "the relationship has been operating on signals you couldn't see, and now you don't know what other signals are missing." The next eight months become a search for the hidden second shoe, which manifests as second-guessing, defensiveness, and a noticeable drop in the willingness to take the productive risks that defined the team's best work.
This is not a soft cost. It shows up in slower decision-making, fewer questions asked in meetings, more political behavior, and a measurable increase in the time managers spend mediating instead of building. The team gets quieter and worse at the same time, and the manager often blames the team for changing — when what changed was the social contract.
Why annual reviews create anxiety, not signal
Annual performance reviews concentrate twelve months of unspoken information into one meeting. By design, the process is the maximum-surprise format.
The structure ensures it. The review cycle starts with goals set in January. The manager and the employee don't formally compare notes against those goals until November. Whatever the employee did in February, March, and April that the manager noticed but never named gets stored, sorted, and delivered nine months later — without the employee having had any way to course-correct in the intervening time.
The result is a meeting whose entire information value is concentrated in surprises. Even when the surprises are positive ("we're promoting you"), the process trains employees to dread the meeting, because the dominant feature isn't the content — it's the distribution. Something might be coming. The employee can't tell what. They can't tell from how the manager has been acting in the months prior, because the manager has been holding back the information that would have given them a useful signal.
This is the anxiety the annual review process was always going to produce. It's not the rating that does it; it's the asymmetric information up to the rating. When the system is designed to surface twelve months of feedback in one sitting, every employee learns to assume that their manager's quiet weeks contain hidden information they're not seeing.
The annual review is a surprise factory. The anxiety isn't a side effect of the format; it's the defining product.
How consistent communication builds trust
The opposite of surprise isn't praise. It's predictability.
When a manager talks to a direct report about their work three times a week — in passing in Slack, briefly in a stand-up, structured in a 1:1 — the report develops an accurate model of how their manager sees their work. They know what's noticed, what's valued, what's flagged. The map is current.
Consistent communication does three specific things to the working relationship.
It removes the hidden information problem. When the manager has been narrating what they observe as it happens, the employee has the same information about their performance that the manager does. There's no quarterly reckoning in which the employee learns about a pattern that was forming all along. There's no September feedback about something that happened in March. The information is always at parity.
It separates noticing from grading. Most performance anxiety isn't about the substance of feedback — it's about the conflation between observation and judgment. When feedback only happens at review time, every observation feels load-bearing because it's tied to a rating. When feedback is constant, observations get untangled from grades. A manager noting that the engineer chose well in a sprint planning meeting isn't an evaluation; it's a fact, and over time the facts compose into a fair picture rather than landing all at once as a verdict.
It calibrates the manager's tone. Managers who only deliver feedback during reviews end up sounding stern by default — their few opportunities have to carry a lot of weight. Managers who give feedback continuously can use a normal voice. The conversation becomes ordinary instead of ceremonial. The employee responds in kind: they discuss their work like adults, not like defendants.
The compound effect is what trust actually means. The manager's view of the work is no longer a hidden variable; the employee operates with it visible. Risk-taking goes up because the calibration of consequences is clearer. The team's output rises because people are spending less energy guessing what their manager thinks.
Small coaching moments versus major corrections
Most experienced managers will tell you that the hardest conversation is the one they delayed for six weeks. Most newer managers will tell you that they don't know what counts as a small enough issue to bring up in the moment.
The line between a coaching moment and a major correction isn't fixed; it shifts based on how recent the underlying behavior was. The same observation — "you missed Tuesday's standup again" — is a coaching moment if it's delivered Wednesday morning. It's a major correction if it's delivered three months later as part of an aggregated pattern.
This is why consistent feedback usually feels lighter for both parties. Each unit is small because each unit references something that just happened. The conversation is about a single instance, in its specific context, with the manager and the report still able to reconstruct what was going on. It costs the manager less to deliver, the report less to receive, and the team less to absorb.
By contrast, the major correction — the version where the manager has been silent for a quarter — is heavy in inverse proportion to how often it's used. Each conversation has to carry months of accumulated information. The manager arrives prepared for resistance. The report arrives feeling ambushed. Both sides spend the meeting managing the meeting instead of working on the thing that needs to change.
The coaching gap most managers walk into gets wider every week the manager doesn't make small corrections. The skill of giving small, in-the-moment feedback has to be practiced; the skill of delivering accumulated criticism has to be developed under pressure. Most managers, trained on the second, never build the muscle for the first.
Switching from major-correction mode to small-coaching mode is, mechanically, the only durable way to reduce the anxiety in the manager-employee relationship. The conversations get smaller because the unit of feedback gets smaller. The relationship gets calmer because nothing accumulates.
What a transparent growth culture actually looks like
The phrase "growth culture" shows up in HR copy on every B2B SaaS website. Most of the time it doesn't mean anything — it's a poster on the wall that nobody can describe in operational detail.
The version that operates rather than poses has specific features. They're easier to recognize than to install:
Feedback is two-way and asymmetric in honest ways. Managers ask direct reports for feedback on their management before reports ask managers for feedback on their work. The asymmetry — the manager initiating — establishes that feedback is normal in both directions, not a power move from above.
Observations are visible. When a manager captures an observation about an employee's work, the employee can see it. The observation file isn't a secret HR document; it's a shared running record. Performance Blocks is designed this way by default: when a manager captures an observation, it lands in the report's view as well, so there's no hidden draft of the relationship being maintained on one side.
Coaching happens in normal voice. The conversation about a missed deadline doesn't have a different tone from the conversation about which meeting to skip. Both happen in the same channel, with the same texture. The team doesn't develop a second emotional register reserved for the meeting where Something Big might be said.
Calibration is grounded. Compensation and promotion decisions reference the observations the employee has already seen, not a separate file generated for HR consumption. The employee's experience of their own performance and the org's documented version of their performance are the same document, with edits visible.
A growth culture isn't built by aspirational language about openness. It's built by removing the structural conditions for surprise. When the information flow is constant, low-stakes, and visible to both sides, trust isn't something the organization has to talk about — it's the texture of how the organization runs.
Frequently asked questions
Why do employees fear performance reviews?
Employees don't fear the substance of feedback — they fear the surprise of it. Annual review formats concentrate months of unspoken observations into one meeting, which means employees can't predict what's coming. The fear is about hidden information that's about to surface in a high-stakes setting, not about the feedback itself. Continuous communication removes the hidden-information problem and most of the anxiety with it.
What is psychological safety in the workplace?
Psychological safety, a term researched by Amy Edmondson at Harvard Business School, describes a team environment where people feel safe to take interpersonal risks — to disagree, ask questions, admit uncertainty, or raise problems. Edmondson's research has consistently linked psychological safety to higher team performance. The mechanism is predictability: when the social environment is predictable, the cost of speaking up is bounded, so people speak up more often.
How can managers deliver feedback without creating anxiety?
The most reliable way is to make feedback a regular, low-stakes occurrence rather than a periodic event. When observations are shared as they happen — in passing in Slack, briefly in a stand-up, structured in a 1:1 — the employee always has an accurate model of how their manager sees their work. There's no hidden information accumulating, no quarterly reckoning, no surprises. The texture of the conversation gets ordinary, and ordinary is what calmly receives feedback.
What's the difference between coaching and criticism?
Coaching is small, specific, and forward-looking — delivered close to the event so the recipient can reconstruct what happened and adjust the next instance. Criticism is large, general, and backward-looking — delivered after enough delay that the recipient can't act on it directly and can only respond to the label. The same content, delivered at different latencies, becomes coaching or criticism. The quality of the relationship depends on which mode the manager defaults to.
How do you build a culture of transparent feedback?
Make observations visible to both manager and employee. Make feedback two-way and have the manager initiate. Keep the tone of feedback conversations the same as the tone of work conversations. Tie compensation decisions to the same observations the employee already saw, not to a separate HR file. The infrastructure isn't aspirational language about openness; it's the structural removal of the conditions that produce surprise — primarily by ensuring that information flows continuously rather than in batches.
What we know — and what we're refining
If you manage people and you've felt the dynamic this post describes — the employee who goes quiet when the conversation shifts toward feedback — the move this week is small: at your next 1:1, name something you noticed in their work this week that you would not normally have brought up. One observation, factual, before any rating implication. Watch what happens to the rest of the conversation.
The thing employees actually dread isn't feedback. It's the asymmetric information that builds up between them and their manager when feedback only happens during formal cycles. Continuous communication isn't a soft cultural ideal; it's the structural fix to a structural problem. We've built Performance Blocks around making this kind of communication low-friction — Henry helps managers capture observations as work happens and share them with the report by default, so the file is genuinely two-way rather than a private HR document. The model is simple: when the information is at parity, surprise stops being the dominant feature of the relationship, and the conversation about performance becomes the conversation about work.
The detail we're still refining is how visible an observation should be at the moment of capture. Some teams want every observation immediately surfaced to the report; others want a layer of editorial filtering before the manager shares. The right default probably depends on the team's existing trust level — high-trust teams seem to handle full transparency well; teams in the middle benefit from a brief delay between capture and surfacing. We're still mapping where the line should sit by default. If you've thought about this, we'd genuinely like to hear what you settled on.
Employees don't hate feedback. They hate finding out, in November, about a thing that happened in March. Take the surprise out of the system and most of the anxiety leaves with it. The conversations get smaller because nothing has been allowed to accumulate. The trust gets larger because the information is shared instead of hoarded. Consistency isn't the soft virtue here. It's the architecture.