Why Recency Bias Is the Single Biggest Distortion in Reviews

You started this quarter off amazingly. You were hitting your numbers and ahead of pace, but started to trail off near the end. Then you get your performance review. Even though you hit your KPIs and delivered on a major project at the beginning of the quarter, it felt like you fizzled out at the end. That resulted in “didn’t meet expectations” on your performance review. This comes because of recency bias.

Recency bias isn’t just a manager's problem; it’s a cognitive problem. Telling your managers to “be more aware of it” won’t solve it. There needs to be structural changes to your performance reviews to account for it. In this article, we’ll preview how recency bias shows up and practical ways to reduce it in your performance reviews. 

What Recency Bias Actually Is and How It Shows Up

Recency bias is the tendency to weigh recent actions or events more heavily than historical ones. For example, if someone asked you, “What music do you like?” You might default to responding with a genre that you recently listened to, and now what your actual favorite is. Recency bias isn’t a bad thing; however, it is something that we have to be aware of and control for - especially when doing an annual or semi-annual performance review.

The Cognitive Science

There’s a reason that people say, “Save the best for last.” It’s because we overemphasize recent experiences compared to those that happened in the past. Nearly 61% of employees said that their last performance review was biased.

In the book Understanding Performance Appraisals, Kevin Murphy and Jeanette Cleveland show that the average window for recency bias is between 30 and 60 days. That means that projects completed longer than 60 days ago will, subconsciously, carry less impact and importance in the review process. Just because managers are aware of recency bias doesn’t mean that they know how to overcome it. In fact, Tversky and Kahneman found that without proper training, managers can overcorrect with recency bias.

The common thought process is that managers fall into the recency bias because they’re lazy or don’t want to put in the effort to do it “correctly.” Behavioral research shows that even your best, most motivated managers can fall into the recency bias trap.

How Recency Bias Distorts Different Performer Tiers

One of your high-performers does exceptional work through all of Q1-Q3, but stumbles in Q4. They might get disproportionately punished for the shortcomings in Q4 when, in reality, they’ve had nearly an entire year of stellar results. This overemphasis on one bad quarter drops them more than their underlying performance warrants.

Recency bias can go the other way and distort an accurate view of a low-performer. A team member who has struggled all year but has a Q4 recovery might get a ‘save’ rating. This ignores issues that have lasted throughout the year and can set a bad standard for other employees. In fact, only 22% of employees believe that their performance review process is fair and transparent. It can also cause legal issues if you have to put them on a Performance Improvement Plan (PIP) in the future.

The Three Calibration Symptoms That Signal Recency Bias

The good thing is that you don’t need to wait for the end of your review cycle to see if recency bias is seeping into your review process. Here are three symptoms to watch out for to diagnose the deeper issue.

1. Symptom 1: Clustered Scores - As reviews start to come in, you’ll start to see scores clustered heavily around the middle of the chart. It will look a lot like a bell curve. There will be very few “exceeds expectations” and “does not meet expectations” in the report. This is usually because team managers can’t remember distinctive events, so they default to the middle.
   
2. Symptom 2: Late-Quarter “Saves” - Generally, with performance reviews, there are comments alongside the score. When these comments primarily reference Q4 accomplishments and projects, while Q1-Q3 projects are missing, it’s an immediate red flag that recency bias is creeping in.
   
3. Symptom 3: Narrative-Rating Mismatch - Likewise, when the number score and the written comments don’t match up, it’s a sign to dig deeper. When the comments praise their work in quarters one, two, and three, but the score only reflects the Q4 stumble, then the manager is rating with recency and not on performance.

The Practices That Reduce Recency Bias Without Eliminating Manager Judgment

So your managers are aware of recency bias, but how do they overcome it? And how do they overcome it while still retaining their own judgment and opinions? 

Continuous Evidence Capture

The most effective way is by continually capturing evidence of your employees' work. This responsibility should fall on the employee AND the manager. While this isn’t trendy, it is the most efficient way.

In practice, that means keeping documents around 1:1s, recognition events, peer feedback, and project artifacts. Capture these throughout the year, in real-time, so that it’s an easy reference at performance review time and you’re not having to dig through your memory for Q1 items in November. We recommend creating a career brag sheet to keep track of major accomplishments for easy reference.

NECTAR TIP: We’ve spoken to several Nectar customers who use the AI summaries in Nominations and Shoutouts to help them gather performance review content.

The Structured Review Prompt

Most managers, when they open their performance review software and see a blank screen, get writer's block and freeze up. The annual performance review becomes a daunting task. Then, slowly, they start to write about what they remember - and it’s always the last quarter of the year. 

To fix that, replace the blank screen with prompts that force the manager to surface citations from each period before writing the review. That can be as simple as: “Cite one specific behavior and measurable outcome from Q1.” Then repeat that process for each quarter (Q2, Q3, and Q4). This forces the brain to think back throughout the year and not just the most recent items.

The Calibration Meeting Design That Names the Bias

Another powerful practice to reduce recency bias is to have a calibration meeting to discuss the bias. This meeting allows HR leaders and managers to align on what recency bias is and whether recent performance reviews are biased. 

At the start of the meeting, the meeting leader sets the stage, making sure that everyone knows why they’re there. You’re there to discuss performance ratings. Recency bias means that you naturally overweigh events from the last 60 days. An important rule for the calibration meeting is that managers must cite evidence from at least three quarters. If a manager’s citations are all from Q4, they need to flag that employee review before it’s approved.

The calibration meeting makes it easy to answer the question, “Where is the evidence behind this rating?” This isn’t about removing manager judgment. It’s a forcing function that drives managers to think critically and look beyond the most recent events.

While this might feel like HR policing managers about employee reviews, it isn’t. The goal of this is to produce ratings that are defensible, consistent, and reflective of the entire performance period. This also, should help provide consistency across managers, departments, teams, and locations.

How Recognition Data Strengthens the Evidence Base

As managers start thinking and writing performance reviews, they’ll want to gather data from multiple sources. One of the richest, untapped performance review sources is your recognition data.

Recognition as Time-Stamped Behavioral Data

Recognition events can be a gold mine of behavioral employee data. When an employee receives a peer-to-peer shoutout from multiple colleagues about leading a cross-functional sprint, that event is time-stamped and highlights a specific behavior and project. While it happened in Q1 and it might fade from a manager’s mind, the record is still there. 

When you go to the employee profile, you can get an AI summary of the last year of recognition events. That way, you can analyze Q1, Q2, Q3, and Q4 events. The manager might not remember what happened in Q1, but the shoutout record or award given shows what happened, who noticed it, and the behavior of the employee.

Over 83% of employees say that recognition affects their motivation to succeed, and 71 would be less likely to leave if recognized more frequently. The motivational value of recognition is well documented.

Using Recognition Without Turning It Into a Popularity Contest

A legitimate concern HR leaders raise about using recognition data in reviews is that a quiet, high-output individual contributor who works asynchronously ends up with fewer recognition events than a chatty, average-output peer. That is a valid and real concern. Here are three guardrails to help prevent this:

• Weight it by giver density and not by raw count - An employee who receives 12 recognitions from 10 different coworkers has a richer record than an employee who receives 15 from the same three teammates. Shoutout diversity signals that the receiver is visible across the organization with different teams and departments and not within a close-knit cluster.
  
• Track company values attached -  Every shoutout should be tagged with a company value, so you can easily track what values employees are being celebrated for. The Meaningful Meter™ also helps employees provide richer, more thoughtful shoutouts instead of generic recognitions of “great job!”
  
• Cross-check against the project outcomes - Recognition data shouldn’t be the only data that you use in performance reviews. It should supplement the real-world outcomes from the projects your employees worked on. An employee with few peer recognitions but strong, documented project outcomes isn’t low performing; they’re just data-sparse. It falls on the manager to actively look for peer feedback from their colleagues.

The Recognition-Narrative Audit

For smaller teams, it's easy to audit all of the performance reviews to make sure that they’re aligned with company values. However, with larger teams, you will need to do a sample of reviews and check that the reviews match the narrative. A small sampling of HR, 10-15% of reviews, can help you get a grasp on whether managers are following the guidance.

For example, when going through the audit, you notice an employee with 22 peer shoutouts tagged with your collaboration core value. That’s amazing! However, when you look over the manager’s review, their feedback was that the employee “tends to work alone.” That’s a red flag that the manager isn’t looking at recognition and possible deeper issues of not knowing what the employee is working on.

This example is a perfect opportunity to coach the manager and show them how to use recognition analytics within their performance reviews to make sure that the narrative matches. Learning how to give constructive feedback to managers and employees is one of the most powerful ways to change behaviors.

The HR Ops Audit Pattern That Catches Recency Bias at Scale

The HR operations team is the backbone of your HR department. They’re the ones who make sure the day-to-day operations are happening smoothly. This team will be essential in training the company about recency bias and spotting the pattern before it becomes an issue.

The Quarterly-Citation Audit

If your team conducts annual performance reviews, the HR ops team will review a small randomized sampling of reviews and tag each event with its quarter. As they go through reviews, they should see approximately the same number of events cited in each quarter. 

For example, three events in Q1, three in Q2, three or four in Q3, and three or four in Q4. It’s when you find six or seven in Q4 while there are three or four in all of quarters 1 through 3 that it becomes alarming. When a review does have the majority in the most recent quarter, it should be returned to the manager to be re-reviewed.

With a company of over 500 people, taking a 10% sampling of performance reviews takes approximately 8 hours of the HR Ops team each cycle. If you have a lean HR team, those hours can be precious. However, taking the time to eliminate (or reduce) recency bias can help you save on recency-driven compensation decisions, flight risks from high performers, and even legal exposure from reviews that contradict documented evidence.

The Narrative-Rating Fit Audit

The narrative-rating fit audit helps catch reviews where the narrative and the points don’t match up. These types of reviews are the ones that cause the most issues if you have to fire an employee for poor performance in the future.

These types of reviews fall into the recency bias trap and are easily spotted when an employee has a “miraculous” turnaround during the last quarter of the year. You can generally spot these with instances where the narrative is positive (e.g., “demonstrated ownership all year and an ability to work well with teammates.”) but has a “below expectations” rating on their score. Vice versa, you could see a score where it says they exceeded expectations, but the narrative is thin, vague, or light on examples.

Narrative-rating mismatches are one of the most common ways businesses can get into trouble and can leave employers open to lawsuits. If a fired employee had stellar performance review scores but the narrative said the work was poor, that can open the door to a discrimination lawsuit.

The Calibration-Meeting Post-Mortem

Last but not least is the post-mortem. Post-mortems are a common practice in business where you evaluate how a process or event went. This is a valuable time to evaluate how the performance review process went, where recency bias showed up the most, and evaluate opportunities to improve during the next cycle.

This calibration meeting should help you identify where you had healthy alignment during the performance reviews compared to where drift started to happen. It is healthy to have reviews change when new evidence is presented. 

For example, if a manager gave a mid-tier rating for an employee, the calibration team evaluated it and surfaced a Q2 recognition moment and project the manager hadn’t considered, and then the manager adjusted the rating upwards. That’s the system working as designed.

Ratings that changed because of pressure from leadership, a manager, because the employee “really needs it”, or they’re a great culture fit, are examples of calibration drift. 

Three Moves Before the Next Review Cycle

Recency bias can’t be trained away; you need to develop an environment to take recency bias into account. Before your next performance review cycle, here are three changes to implement.

1. Install a structured quarterly review prompt - This single prompt forces employees and managers to evaluate each quarter and think back about everything they’ve accomplished this year.
   
2. Introduce the recency-bias prompt at the start of your next calibration meeting - At the beginning of each calibration meeting, recite the prompt, “Every rating must cite evidence from at least three quarters.” It may sound silly, but it will keep it top of mind for everyone involved.
   
3. Run the quarterly citation audit on your previous cycle’s reviews - You can start fixing things and training team members now. Review last cycle’s performance reviews and see where gaps are in your team.

Implementing these processes into your performance reviews will reduce recency bias in your performance reviews. Employees and managers will have a more comprehensive picture of the work they’re doing, and not just the last 3 months. 

Nectar’s recognition platform gives managers time-stamped recognition behavior that they can draw on to enrich and cite during their performance reviews. Schedule a demo to see how recognition can be used in your performance reviews.

Frequently Asked Questions

What is recency bias in performance reviews, and why does it happen?

Recency bias is the tendency to weigh an employee’s most recent work more heavily than their performance across the entire period. This is a common cognitive bias and happens not only in business but also in sports, music, and multiple other facets of life.

Does manager training alone reduce recency bias?

No, manager training raises awareness of recency bias, but does not fix the problem. To actually reduce recency bias, companies must introduce structural changes (e.g., continuous performance documentation, prompts that force full-period reviews, and calibration) to their company.

How do you spot recency bias in calibration meetings?

Some telltale signs to watch out for are:

• Evidence clusters in the last quarter of the review period
• A high score that’s justified almost entirely by a recent win (or a low score by a recent mistake)
• A narrative that is vague without any evidence (e.g. they’re really hitting their stride)
• Sharp wings from prior cycles that lack strong evidence

Can recognition data be used as evidence in performance reviews?

Yes! In fact, it’s one of the best antidotes to recency bias because recognition data is timestamped across the entire year. It makes it easy to find peer-to-peer recognition, manager shoutouts, nominations, and award activity to create a detailed list of their accomplishments for their performance review.

What review prompts force managers to look beyond the last 60 days?

Review prompts are great ways to help managers and employees remember what they’ve done beyond just the past 60 days. Here are some helpful prompts to get them thinking:

• “Describe one significant contribution from each quarter of the review period”
• “What was this person working on in [Quarter]? What was the outcome?”
• “Cite two examples of impact from before the last 90 days”
• “What have they improved on since the first half of the cycle?”

How should HR ops audit reviews for recency bias before they go to compensation?

HR ops is the backbone of a successful review process. A strong HR ops team will audit the reviews to make sure that recency bias isn’t slipping in, and that reviews are completed. During their audit, they’ll look to make sure cited examples in reviews aren’t heavily weighted in the most recent quarter, monitor large swings in employee ratings from previous quarters, and review a small sample size of reviews to make sure that the rating and narrative match.

Can continuous performance management eliminate recency bias entirely?

No. Recency bias is a natural bias that all people deal with. Continuous performance management can help reduce its presence, but can never fully eliminate it.