Continuous feedback vs annual reviews at small companies

If you run a team between ten and fifty people, you've probably had the same uncomfortable conversation more than once. Someone — usually a thoughtful operator, sometimes a board member, occasionally a hire from a bigger company — points out that your feedback process is ad hoc, and asks what the plan is. The implicit choice on the table is binary: build out an annual review process (the formal, calendar-driven, documentation-heavy version), or buy a continuous feedback tool (the always-on, app-based, dashboard-driven version). Both options are wrong for small teams, and most managers who pick one come to regret it within a year. The third option is the one nobody pitches you, because it isn't a product. (This piece is one part of a broader argument about why performance review rituals fail at small companies — the pillar walks through the four jobs reviews try to do, and why none of them actually require the annual ceremony.)

Why annual reviews don't fit small companies

Annual reviews evolved inside organizations that had three things small companies don't: a legal department, a multi-thousand-person headcount, and a calendar synchronized to compensation cycles. The annual review was solving for those constraints — documenting performance for severance defensibility, calibrating ratings across managers who don't know each other's reports, distributing a fixed bonus pool. None of those problems exist in a forty-person company.

Removing the constraints doesn't make the form lighter; it makes it ceremonial. The manager and the report sit down once a year, fill in a template designed for someone else's compliance problem, and produce a document neither of them will read again. The conversation that was supposed to be about growth becomes about whether the rating matches the raise, and whether the raise matches the budget. The rest of the meeting is theater performed for an audience of zero, because at a forty-person company there's no calibration committee to perform it for.

There's also a more subtle problem: the cadence is wrong. A year is far too long to wait to surface a real performance issue, and far too short to evaluate whether someone has grown into a meaningfully different version of themselves. So annual reviews end up doing neither well. The people who needed feedback months ago got it months late, and the people who could use a real growth conversation get a recap of the year instead.

Why "continuous feedback" doesn't fit either, mostly

The standard pitch from continuous-feedback SaaS is that you replace the annual ceremony with a steady drip — weekly check-ins, in-the-moment praise, real-time peer recognition, dashboards that show "engagement" trending up or down. It sounds like the obvious upgrade. In practice, three things go wrong when you actually run it at a small company.

The signal-to-noise ratio collapses. When everything is feedback, nothing is. A manager who sends a thumbs-up emoji on Slack three times a day has not given feedback. They've given chat reactions. The continuous model conflates acknowledgment with feedback, and after a few months neither party can distinguish them. Real feedback — the kind that takes a moment to formulate and a moment to absorb — gets crowded out by the cheaper version.

It becomes surveillance-flavored. A platform that prompts your team to record their feelings about every meeting, log peer recognition with a structured tag, and rate their week on a 1–5 scale is one quarter away from feeling like soft monitoring. At enterprise scale this gets muffled by sheer process volume. At forty people, where everyone notices everything, the surveillance read shows up fast and quietly poisons trust.

It's tool-dependent. Most continuous-feedback platforms are priced and designed for org charts with HR roles. The features you actually use — 1:1 templates, occasional pulse surveys, peer feedback — are the cheap parts. The features that justify the price are the dashboards no individual manager looks at twice. Small companies end up paying enterprise-tier software to get IC-level value, and the contract starts feeling expensive about the time the novelty wears off.

There's a smaller version of continuous feedback that does work — short, intentional, behavior-specific, attached to actual work moments — but it's not what the SaaS vendors sell, and it doesn't need a tool. More on that in a second.

The third option: rhythm-based feedback

The framing that actually fits a small company is neither continuous nor annual. It's rhythm-based. You match feedback cadence to three rhythms that already exist in the company, and you stop trying to impose a fourth.

Rhythm 1: Work cycles. Projects, sprints, launches, quarters, account renewals — whatever the natural unit of work is for a given role, it ends. The end of a unit of work is the highest-signal moment for feedback, because everyone remembers the specifics, the outcome is observable, and the lessons are still hot. Anchoring feedback to work cycles instead of to the calendar is the single biggest move a small company can make — surgeon Atul Gawande argues the same for surgery and elite tennis: the highest performers refine through close observation of specific moments, not annual stocktaking.

Rhythm 2: Trust capacity. People can only absorb so much honest feedback before they stop being able to act on any of it. The cap is roughly one substantive round per quarter per person, with smaller in-the-moment corrections in between. That's the constraint, not how often a tool prompts you to log a check-in. Pretending the cap is higher just generates noise — which is exactly what continuous-feedback SaaS optimizes against.

Rhythm 3: Decision rhythm. Compensation changes, promotions, role shifts, exits — these decisions have their own cadence, and they're the only ones that legitimately need calendar-anchored timing (usually annual, sometimes biannual). The mistake is conflating the decision conversation (which is calendar-driven) with the development conversation (which should be work-driven). When you separate them — when comp gets its own dedicated meeting, distinct from the feedback conversation — both work better.

Three rhythms, three different cadences. Trying to compress all three into a single annual ritual is what makes annual reviews feel hollow. Trying to dissolve all three into a continuous flow is what makes continuous feedback feel like noise. The right move is to honor the rhythms separately.

What this looks like in practice for a 10–50 person company

Concretely, here's what rhythm-based feedback looks like operationally for a small company. None of it requires a platform.

End-of-project retros (work cycle). When a meaningful unit of work ends — a project shipped, a launch landed, a quarter closed — the manager runs a short retro with each contributor: what worked specifically, what you'd change, what you saw me do that helped or didn't. This is where most of the actual development feedback happens, because the specifics are still vivid. Small enough projects don't need this; big enough ones absolutely do.

Quarterly self-assessment with selective peer feedback (trust capacity). Once a quarter, each manager runs a structured self-assessment using a real template — not a free-form journal — and selectively asks two or three direct reports, peers, or cross-functional collaborators to answer the same questions. The point isn't volume; it's the comparison. The gap between your self-view and the aggregated peer view is where the actual signal lives. (For why the self-side has to come first: Why self-assessment comes before peer feedback.) This becomes the development conversation that annual reviews try and fail to be.

1:1s as ongoing micro-corrections (in between). The 1:1 is where real-time, low-stakes feedback happens — not as a check-the-box ritual, but as the channel for "I noticed something in Tuesday's review, want to talk about it?". This replaces the surveillance-flavored continuous feedback model. The volume is low, the specificity is high, and the cadence matches actual events.

Annual conversation, but only about the decision (decision rhythm). Once a year — or whenever your comp cycle runs — you sit down and talk specifically about role, level, comp, and the next horizon. Not about performance in the abstract. Not about strengths and weaknesses. About the decision in front of you. The performance context for that decision was already built up over the previous twelve months through the other three rhythms; you don't need to relitigate it in the meeting.

The total time investment for a manager running this rhythm is about the same as running annual reviews well — maybe slightly less. The output is dramatically more useful, because the feedback is anchored to moments instead of calendar pages.

Common mistakes when small companies try to "modernize" reviews

Three failure patterns show up reliably when a small company tries to evolve past the annual review.

Adopting an enterprise-grade tool to replace a small-team problem. Buying Lattice or 15Five at thirty people because "we should be more rigorous about feedback" almost always ends with the tool being half-used after eight months and abandoned after eighteen. The features you needed could have been a templated doc, a calendar reminder, and a quarterly hour of focus. The features you didn't need are what you paid for.

Letting "continuous" mean "constant". A platform that prompts daily reflection and weekly pulse surveys is generating data, not feedback. The discipline of small-company feedback is in the editing — what you choose not to formalize, what you let stay informal, what you save up for the substantive quarterly conversation.

Confusing development feedback with comp conversations. When the only formal feedback moment of the year is the comp meeting, every piece of development feedback gets contaminated by the question "is this going to affect my raise?". The fix is to separate the two — different meetings, different inputs, different stakes — and make the separation explicit to the report.

Trying to fix the form before the rhythm. Most companies start by redesigning the review template. Better questions, cleaner rubric, more open fields. It rarely changes anything, because the form was never the problem. The problem was that the cadence was wrong. Fix the rhythm first, then the form takes about an afternoon to write.

What to do next

If you're running feedback for a small company and the current process feels like overhead, take a quarter to test the rhythm-based version. Three concrete moves:

1. Pick one finishing project this quarter and run an end-of-project retro with each contributor. Twenty minutes per person, three behavioral questions, written notes you keep. Notice how much signal comes out of one twenty-minute conversation versus twelve months of "any thoughts on the quarter?".
2. Run one structured 360 round on yourself. Self-assessment first, then send the same six to ten questions to two or three people. Compare the gaps. (The starter list: 50+ 360 feedback questions that actually surface useful answers.)
3. Move the comp conversation off the development meeting. Even if the calendar still says "annual review", separate the two parts of the conversation into two meetings. Notice how the development half gets sharper when the money question isn't on the table.

If you want a structured way to run the quarterly self-and-peer round — same templates re-used round over round, your self-answers side by side with your respondents', tracking what changes across quarters — that's what Mirorly's templates handle. The Library is where the per-context templates live. None of it is built for HR; it's built for individual managers running their own development loop, which is exactly the small-company use case.

The third option isn't a product, isn't a process upgrade, and isn't a feedback philosophy. It's a permission to stop choosing between two systems that were both built for someone else, and to use the rhythms you already have.