The manager brings up something that went wrong in March. It is now November. The employee remembers it differently, or does not remember it at all. There is a short pause while both of them decide whether to argue about it.
Neither one does. The form gets signed. Nothing changes.
That pause is worth understanding, because it is not a manager problem or an employee problem. It is a design problem, and it has a price tag.
The annual review is an expensive memory test
The traditional appraisal model was built for a slower world, and its running costs are easy to underestimate. Deloitte counted the cost of its own performance process and found the firm was spending close to 2 million hours a year on it. Not on coaching. On the process itself.
Adobe ran the same audit and found 80,000 manager hours a year going into reviews, which is roughly 40 full-time people doing nothing else.
The natural next question is what all that time buys. Gallup has asked employees directly, and the answers are not encouraging.
Three out of four employees do not believe their own review is accurate. Thousands of hours go into producing a document most people quietly disagree with.
Peter Cappelli, Director of the Center for Human Resources at The Wharton School, has spent years studying why these systems underperform. His diagnosis is that annual reviews:
That distinction is the heart of the problem. A healthy performance management cycle runs through planning, monitoring, developing, and reviewing, but most companies only staff the last stage. Reviewing is measurement. Managing is what happens in the eleven months between measurements. Most companies have automated the first and left the second to chance.
Srikant Chellappa, Co-Founder of Engagedly, describes the part that gets left to chance:
“Our job as leaders is to keep people focused, by giving them a sense of purpose and highlighting their individual accomplishments and its impact on the business.” Srikant Chellappa, Co-Founder, Engagedly
Keeping people focused is a weekly act, not an annual one. When it only happens once a year, five specific failures follow.
What goes wrong in the gap
Recency bias. A manager rating twelve months of work mostly remembers the last six weeks. Everything earlier blurs, so the review reflects the calendar rather than the contribution.
Goal drift.Clear goals are the backbone of the whole system, yet a goal written in January often describes a company that no longer exists by June. If nobody reopens the document, people spend half the year working toward targets that have quietly stopped mattering.
The February exodus. Adobe found voluntary turnover spiked every February, right after ratings landed and people processed a disappointing number.
Silent underperformance. This is the most expensive of the five, and the one that costs the most trust.
Kim Scott, Co-Founder of the management training firm Radical Candor and author of the book of the same name, spent years leading teams at Google and on the faculty of Apple University. She tells the story of an employee she calls Bob, whose work was poor for ten months while she reassured him to spare his feelings. When she eventually had to let him go, his reaction was the part she never forgot:
By the time most companies act on a case like Bob’s, the only tool left is a formal plan, which is why alternatives to the performance improvement plan are worth having before you need one. Scott’s conclusion applies to every review cycle that saves the hard news for December: “It sounds so simple to say that bosses need to tell employees when they’re screwing up. But it very rarely happens.”
Every one of these five failures comes from the same source, which is distance between the work and the conversation about it. Close that distance and the failures lose their oxygen. That is what real-time performance management is for.
What “real time” actually means
Real-time performance management is not a faster annual review. It is a different shape:
Feedback lands within days of the work, not months
Goals stay visible and get updated as priorities shift
Check-ins are short, frequent, and forward-looking
The formal review summarizes things that were already said
The last point does most of the work. In a real-time system, nothing in the review is new information, so the conversation stops being a verdict and becomes a summary.
Scott builds that same timing into her definition of useful feedback, using the acronym HHIPP:
“Radical candor is humble, it’s helpful, it’s immediate, it’s in person, in private if it’s criticism and in public if it’s praise, and it doesn’t personalize.” Kim Scott, Co-Founder, Radical Candor
Immediate is the word an annual cycle cannot satisfy by design. And the research on what immediacy is worth is unusually clear.
The evidence on frequency
Gallup found that employees whose managers give daily feedback are 3.6 times more likely to be motivated to do outstanding work than those receiving annual feedback.
Recognition follows the same curve. Daily shout-outs leave 98% of employees feeling valued, while annual feedback leaves 37% feeling that way.
Josh Bersin, founder of the HR research firm The Josh Bersin Company and previously of Bersin by Deloitte, has covered this market for two decades and reaches the same conclusion. Companies with a feedback culture consistently outperform those that do not encourage it. His term for the goal is enabling performance “in the flow of work”, meaning inside the tools and rhythms people already have rather than in a separate annual exercise.
The research is one thing. What makes the case harder to argue with is that several of the largest US employers have already run the experiment, at scale, and published what happened.
In 2012, Donna Morris, then Chief Human Resources Officer at Adobe, announced the company was scrapping annual reviews. She said it to a journalist, on a flight to India, before she had told her own team or the CEO.
Her reasoning:
“Adobe was founded on four core values: genuine, exceptional, innovative, and involved. Our old annual review process contradicted every one of them.” Donna Morris, former CHRO, Adobe
Adobe replaced it with Check-in, a rhythm of frequent, lightweight manager conversations about expectations, feedback, and growth. No ratings, no rankings, no forms.
The second row answers the objection most leadership teams raise first. Involuntary departures went up because honest conversations happened earlier, and performance problems that used to sit unaddressed for a year got surfaced in weeks. Removing the annual review did not remove accountability. It moved it forward.
Marcus Buckingham, then head of people and performance research at the ADP Research Institute, and Ashley Goodall, then Director of Leader Development at Deloitte Services LP, described the redesign in Harvard Business Review. No cascading objectives, no annual review, no 360 tool. Instead, weekly check-ins, plus four short questions each team leader answers about each person at the close of a project.
Their design principle was the same one Cappelli identified from the outside: stop assessing the past and start fueling the future.
General Electric: the company that invented ranking, abandoned it
GE built the stack-ranking system that the rest of corporate America copied for thirty years, which makes its reversal the most striking of the four.
It dropped forced rankings in the mid-2010s and replaced them with an app for continuous “touchpoints” between managers and employees, built around two recurring questions: what should I keep doing, and what should I change.
Netflix: informal 360s instead of a formal cycle
Netflix went further than most and dropped formal reviews without replacing them with another process. Patty McCord, Chief Talent Officer at Netflix for fourteen years and co-author of the company’s culture deck, described what took their place in Harvard Business Review:
“When we stopped doing formal performance reviews, we instituted informal 360-degree reviews. We kept them fairly simple: People were asked to identify things that colleagues should stop, start, or continue.” Patty McCord, former Chief Talent Officer, Netflix
The underlying assumption, in her words, was that people can handle anything as long as they are told the truth. The format survives outside Netflix even if the rest of that culture does not suit your company, and it sidesteps the question of which rating scale to use entirely. Stop, start, continue takes about four minutes to answer and produces more usable feedback than a five-point scale.
Across all four, the same trade appears. Less process, more conversation, and the conversation moved closer to the work.
What changes with a real-time system
The annual model
The real-time model
Feedback once a year
Feedback in the flow of work
Manager recalls from memory
Continuous record of actual work
Goals set once, forgotten
Goals visible and updated
Review is a verdict
Review is a summary
Problems surface at year-end
Problems surface in weeks
Recognition is annual and abstract
Recognition is immediate and public
Data lives in a spreadsheet
Data lives where the work happens
The right-hand column looks obvious on paper, which is why so many companies announce it and then watch it fade. None of it runs on goodwill. Managers are already stretched, and asking them to give more feedback produces three weeks of enthusiasm followed by silence.
It holds only when a system handles the remembering, the prompting, and the record-keeping. That is where software earns its place, and it is the problem Engagedly was built to solve.
How Engagedly does this
Engagedly did not start as a performance platform. It launched as a social collaboration tool, and the market pushed the founders somewhere else.
“We were initially focused on social tools in the platform as a product, but feedback from the market and our clients revealed they needed a more holistic approach that connected employee engagement with performance objectives.” Srikant Chellappa, Co-Founder, Engagedly
Give or request feedback in a few clicks, at any time
Every piece of feedback is stored and searchable
Public praise flows to the social feed, so recognition is visible
By review season, the manager has evidence rather than recollection instead of hunting for review phrases to fill a blank form
That last line is the direct answer to recency bias. A manager writing a review from a year of logged feedback is not reconstructing anything.
Goals and OKRs, for the drift problem
OKRs & Goals keeps objectives current and visible:
Cascading goals connect individual work to company strategy, using SMART goal formats people can actually measure
Progress updates happen continuously, not at quarter-end
Everyone can see how their work ladders up
Visibility matters as much as accuracy here, because employees cannot align to a strategy they cannot see. Chellappa makes that point about the leader’s side of the equation:
“Two important things you can do as an organizational leader is to make sure that your objectives and your company’s purpose is very clear and transparent to everyone, not just your direct reports. Also, employees should have a clear understanding of what they are working towards and how their contribution is moving the needle forward.” Srikant Chellappa, Co-Founder, Engagedly
Gallup suggests this is rarer than leaders assume. Only 26% of employees strongly agree they understand how their work connects to company goals.
Goals give the check-in something concrete to be about, which is what separates a useful employee check-in from a status update. Engagedly Meetings turns one-on-ones into a habit rather than an intention:
Shared agendas, so nobody arrives cold
Talking points and action items carried forward
A running history of what was discussed and decided
Reviews with a paper trail
By the time you run a performance review, the system already holds a year of feedback, goal progress, check-in notes, and recognition. The review becomes the summary described earlier rather than a reconstruction.
Adding 360 feedback, run to established best practice, widens the picture beyond one manager’s viewpoint, which is the most reliable way to dilute the rater bias that thin evidence encourages. It is the same instinct behind McCord’s stop, start, continue, with the collection handled for you.
AI that surfaces the pattern
Marissa™ AI, Engagedly’s AI layer, drafts feedback, summarizes review inputs, and flags patterns a manager may not have noticed. Our guide to using AI in performance reviews covers where that help is safe and where it is not.
The division of labor matters, given what the Amazon example showed about automated judgment. AI drafts, the human decides. The point is not to remove the manager from the conversation but to clear the admin work off their desk so the conversation can be the job.
There is a cultural condition attached to measuring people this often. It only helps if people are still allowed to take risks, which is the argument Chellappa made in TalentCulture:
A once-a-year verdict punishes a bad quarter. A continuous record can absorb a miss in March and still register a strong second half.
Signals from the whole employee
Performance problems are often engagement problems in disguise, which is why the record of work is only half the picture. Team Pulse and Employee Surveys catch a drop in sentiment while it is still a conversation rather than a resignation, well before the February exodus pattern has a chance to repeat.
It meets people where they work
All of the above depends on people actually opening the tool. Flow of Work integrations put feedback and goals inside Slack, Teams, and the applications people already have open, which is Bersin’s flow-of-work principle in practice. The mobile app covers frontline and field teams who rarely open a laptop.
Where teams get this wrong
Having the platform is not the same as having the practice. Four failure modes account for most stalled rollouts, and we have written a fuller list of performance management adoption barriers elsewhere.
Adding real-time on top of the annual review. Keep the twelve-page form and add weekly check-ins, and you have doubled the workload. Continuous feedback should shrink the formal review, not sit alongside it, which is exactly what Adobe and Deloitte did.
Skipping manager training. Frequent bad feedback is worse than infrequent bad feedback. Scott’s HHIPP standard is a usable checklist here, and the SBI model of situation, behavior, impact gives managers a repeatable structure for being specific, behavioral, and forward-looking.
Leaders who opt out. If the executive team skips check-ins, the layers below will too. Adoption is copied downward.
Avoiding those four is mostly a question of sequencing, which is what the next ninety days should look like.
A 90-day starting plan
Days
Focus
What good looks like
1-30
Goals
Every employee has 3 to 5 visible, current goals
31-60
Check-ins
Monthly one-on-ones happening for 80%+ of teams
61-90
Feedback and recognition
Most managers giving feedback monthly; praise visible publicly
Ongoing
Review
The annual review shortens, because the year is already documented
Goals come first for the reason given earlier. Check-ins need something concrete to be about, and feedback needs a shared definition of what good performance looks like before it can be useful.
The bottom line
The annual review asks a busy person to recall twelve months of someone else’s work from memory, then compress it into a number that affects someone’s pay. Cappelli’s objection, Scott’s story about Bob, and Gallup’s numbers all describe the same failure from different angles.
Adobe stopped doing it and cut voluntary turnover by 30%. Deloitte stopped and reclaimed hundreds of thousands of hours. Netflix replaced it with four minutes of stop, start, continue. GE, which invented the ranking model everyone copied, stopped as well.
Real-time performance management is not a gentler alternative to any of that. Problems get named while they are still small, good work gets recognized while it still feels recent, and the review, when it arrives, contains nothing anyone should be surprised by.
How does real-time performance management software work?
Managers and employees set goals in a shared system, update progress as work moves, and exchange feedback through the platform or through an integration with Slack or Teams. Check-in notes, feedback, and recognition are all logged. At review time, the system assembles that history into a draft.
What features should I look for in performance management software?
The five that matter most for a real-time process are goal and OKR tracking, continuous feedback, one-on-one check-in agendas, 360 or multi-rater reviews, and integration with the tools people already use. Recognition and employee surveys strengthen the picture, since disengagement usually shows up before a performance drop does.
How is this different from employee monitoring software?
Monitoring measures activity: keystrokes, hours, screen time. Performance management software captures outcomes, feedback, and goal progress. The distinction matters to employees, given that 61% of Americans oppose AI tracking their movements at work.
Does performance management software integrate with our HRIS?
Most established platforms do, and this is worth confirming before you buy. Employee records, reporting lines, and job data should sync from the HRIS automatically, otherwise HR ends up maintaining the same org chart twice.
Is real-time performance management suitable for small HR teams?
Yes, and arguably more so. A small HR function cannot manually chase a company-wide review cycle, so automating the reminders, the collection, and the record-keeping frees up more time proportionally than it does in a large enterprise.
Does AI in performance management create bias risk?
It can, if the model makes the decision. Amazon’s scrapped recruiting tool learned bias from its own hiring history. The safer pattern is the one described above: AI drafts and summarizes, and a human reviews, edits, and owns the outcome.
Jeevithan K is a Senior Solutions Consultant at Engagedly with extensive experience in HR technology, customer success, product implementation, onboarding, and training. He specializes in helping organizations adopt B2B SaaS and enterprise AI solutions effectively. His expertise also includes project management, account management, process improvement, and translating complex product capabilities into practical solutions that deliver meaningful value for customers.