Every few years, someone in HR announces that annual performance reviews are broken and the company is switching to something better. Continuous feedback. Real-time check-ins. Always-on development conversations. The org takes a deep breath and feels hopeful.
Then the next comp cycle rolls around.
Suddenly, the question isn't "how do we give people better feedback?" It's "how do we justify these raise numbers?" And managers are back to the same place they've always been: trying to remember what someone did nine months ago, scrolling through Slack for evidence, and making calls based on gut.
The tools changed. The problem didn't.
Why annual reviews got replaced
Annual reviews had obvious flaws. They compressed a full year into a single stressful meeting. Recency bias was baked in. Whatever happened in the last 60 days loomed larger than what happened in January. The process took weeks of everyone's time and often produced ratings that felt disconnected from what people actually did.
The format also made honest conversations hard. Employees walked into a room knowing their raise was tied to whatever their manager said in the next 30 minutes. That's not a setup for candid development talk. It's a negotiation with bad information.
Companies moved toward continuous feedback to fix this. More frequent check-ins. 360 surveys throughout the year. Manager feedback delivered closer to the work.
That shift made sense. Feedback given six months after the fact isn't feedback; it's history. Getting closer to the work improves development conversations. Managers can course-correct before problems compound. Employees get to adjust in real time rather than learning at the end of the year that they'd been doing something wrong for months.
But there's a catch.
Continuous feedback created a new problem
When feedback is ongoing and informal, it doesn't accumulate in a structured way. By the time comp decisions come around, managers have a pile of signals (some documented, most not) and no framework for weighing them.
The result: comp decisions made on vibes.
Not because managers don't care or aren't trying. They genuinely want to be fair. But without a structured performance calibration process, they default to what they remember, what they feel, and what they can articulate under pressure in a room full of other managers.
That's not a people problem. It's a process problem.
The employees who get rewarded tend to be the ones whose managers are good storytellers, or who work on visible projects, or who happen to have had a strong Q4. The employees who don't advocate for themselves, or whose work is hard to quantify, get left behind.
This is where continuous feedback falls short: it improves the day-to-day, but it doesn't fix comp equity.
The real pain point: manager vibes run compensation
In most companies, here's how comp decisions actually happen:
A manager sits down to fill out ratings or write justifications. They pull from memory. They reference whatever feedback they happened to capture. They make a case for their people based on what they can recall and defend.
Then those decisions go into a calibration room, where managers compare notes and adjust based on budget, peer pressure, and whoever argues most convincingly.
The process is exhausting and the outcomes are inconsistent. Two employees with nearly identical performance records end up with different outcomes because of who advocated for them and who didn't.
HR teams know this. They talk about it every cycle. But without better inputs (more structured, more objective, less dependent on individual managers remembering everything), calibrations stay messy.
The solution isn't to tell managers to document more. They're already overwhelmed. They have one-on-ones to run, performance conversations to prep for, hiring to manage, and actual work to do alongside their team. Adding another documentation requirement doesn't fix the comp process; it just creates more admin with nowhere useful to go.
The solution is to change what data feeds into the process.
What a better performance calibration process looks like
Confirm approaches calibration differently, by combining three things that most tools handle separately:
Lightweight calibration sessions. Instead of marathon review meetings where managers debate ratings from scratch, Confirm structures calibrations around data. Everyone comes in with the same view of performance signals, not their own filtered memory. Discussions focus on exceptions and edge cases, not building the case from zero.
Organizational network analysis (ONA). Vibes-based decisions often fail because they rely on what managers see, and managers can't see everything. ONA surfaces who's actually doing the work by mapping collaboration patterns and peer influence across the org. An employee who's critical to how the team operates shows up in that data, even if they're not the loudest person in the room. This turns informal influence into a visible, measurable signal.
AI-assisted performance summaries. The volume of signals from continuous feedback programs is real, but humans can't synthesize dozens of check-in notes and peer comments in real time. AI can. Confirm's AI layer processes that data and produces structured performance summaries that go into calibrations, so managers are working from a complete picture, not a selective one.
Together, these tools shift comp decisions from "what does the manager feel like arguing for" to "what does the evidence support."
The shift that actually matters
The problem with annual reviews was never that they happened once a year. It was that they were the only structured moment to surface performance data.
Continuous feedback fixed the cadence but left the comp process intact. Managers still had to reconstruct a year's worth of signals from scratch, under time pressure, without a shared framework.
The performance calibration process is where that has to change. When calibrations pull from structured data (collaboration patterns, peer input, AI-synthesized summaries), they stop being a test of memory and advocacy and start being a conversation about evidence.
That's when compensation gets fair. Not because everyone agrees, but because everyone's working from the same information.
Confirm is built for this: the full arc from feedback to calibration to comp, without the vibes.
Want to see how Confirm's calibration process works? Book a demo and we'll walk you through it.
