“One of our best people walked in one morning and handed me a resignation letter, completely out of the blue.” It’s a story you hear again and again at Japanese companies operating in Vietnam. But the truth is that most departures are never actually “sudden.” People who leave start giving off signals months before they resign. The problem is that our instincts alone aren’t reliable enough to catch them.
What’s more, the things people commonly think of as “signs of an imminent departure”—a resume left out on a desk, showing up in an interview suit, a sudden increase in doctor’s visits—are, in fact, not warning signs at all in any scientific sense. The real signals appear much more quietly, buried in your HR data. In this article, we explain the “pre-departure signs” you can capture with data, not gut feeling, backed by research evidence. This is the second installment in our “Retention & Reducing Turnover” series. Where the first installment, “Vietnam’s Turnover Rate in Numbers,” covered the big picture, diagnosis, and countermeasures, this article zeroes in on one thing—early detection—and goes deep.
1. Why Catch the Early Signs — Once the Resignation Lands, It’s Too Late
Early detection matters because in Vietnam, trying to retain someone after the resignation letter is on the table is usually too late. As we saw in the first installment, Vietnam is a high-mobility market where 64.8% of employees are considering a job change within six months (Talentnet-Mercer). On top of that, Vietnam’s headhunting market has recently expanded from around 200 firms to roughly 500 (Talentnet), so the more in-demand the role, the more external offers an employee is exposed to—and the shorter the window between the first warning sign and an actual departure.
At the same time, turnover is not something that simply can’t be prevented. According to Gallup, 42% of voluntary turnover was preventable—in other words, the manager or the organization could have done something. Furthermore, 45% of people who left had almost no meaningful conversation with their manager in their final three months. Flip that around, and it means there is a very real category of departures that could have been prevented if the warning signs had been caught and a conversation had happened before the resignation. That is exactly why capturing “pre-departure signs” is worth the effort.
2. Myth vs. Reality — What the Real “Pre-Departure Signs” Are
Before we talk about the real signals, we first need to debunk some widely believed myths. Get this wrong, and you’ll end up monitoring things you shouldn’t while missing the signs that actually matter.
These Are NOT Warning Signs — Common Misconceptions
“Leaving a resume out on the desk,” “showing up in an interview suit,” “a sudden increase in doctor’s visits”—these visible behaviors are, in fact, not warning signs of turnover at all, as clearly disproven in research by Gardner & Hom (2016) of Cornell University. Managers who treat these as “red flags” are relying on assumption; the study found no statistical link between these behaviors and actual departures. Monitoring based on gut instinct only generates false alarms.
So what are the real signs? The same research identified 13 “pre-quitting behaviors.” They included a decline in work performance and productivity, reduced engagement with the team and the job, doing only the bare minimum, more frequent expressions of dissatisfaction, and a loss of enthusiasm and idea generation. Employees whose scores on these behaviors reached an average of 4.2 or higher were reported to be roughly twice as likely to leave compared with those who did not.
In other words, the genuine signals are not “flashy, visible behaviors” but rather a quiet decline in engagement and output. Precisely because they aren’t dramatic, they are easy to overlook—which is exactly why capturing them with data matters. Academic meta-analysis (Rubenstein et al. 2018) likewise confirms that the strongest leading indicators are the intent to leave and job-search behavior itself, followed by changes in performance and attitude.
3. Different Types of Signs Appear at Different Lead Times
One thing worth keeping in mind when tracking warning signs with data is that “how many months in advance a sign appears” varies by the type of sign. The earlier a sign shows up, the larger the window you have to intervene.
| Type of Sign | How Long Before Departure | Source |
|---|---|---|
| Decline in engagement (involvement, loyalty) | ~9 months before | Peakon (analysis of 33M+ data points) |
| Decline in performance / productivity | 2–6 months before | Ko & Trevor 2026 |
| Pre-quitting behaviors (lower engagement, more dissatisfaction) | Predicts departure within 12 months | Gardner & Hom 2016 |
In Peakon’s “9-Month Warning,” which analyzed more than 33 million data points, it was shown that departing employees’ sense of loyalty and perceived managerial support begin to decline continuously about nine months before they leave. Performance decline, by contrast, appears much closer to the departure—becoming significant 2–6 months out (Ko & Trevor 2026). If you can catch the shift in engagement early, you have more than half a year to act; by the time performance decline is visible, your time is already running short.
4. The Concrete Signs That Show Up in Your HR Data
So where in your day-to-day HR data should you actually look? Here we organize the observable warning signs into four categories—attendance, leave, payroll, and evaluation—with a note on “why it’s a warning sign” for each.
Attendance Signs
- A sudden spike in overtime: A heavier workload leads to burnout, which can be a trigger for departure.
- A sudden drop or irregularity in overtime: An employee who used to work overtime suddenly leaving on time day after day can be a sign of declining commitment to the job.
- Increased absenteeism: Meta-analysis finds that absenteeism (ρ=.23) is a stronger leading indicator of turnover than lateness (ρ=.14). Concluding that “if lateness increases, it’s automatically dangerous” is a mistake—absenteeism should be weighted more heavily.
- Unlogged overtime: A situation where overtime is actually happening even though it hasn’t been filed can be a sign of dissatisfaction with the system or an unreasonable workload.
Leave Signs
Leave is an indicator where you should watch the two extremes. An employee who continues to take 0% of their paid leave is at risk of burnout. Conversely, an employee who never used to take leave suddenly starting to burn through it can indicate a job search or preparation to resign. Rather than looking at just one end, what matters is paying attention to any change that is “unlike usual” for that particular person.
Payroll Signs
- A downward trend in take-home pay: When take-home pay keeps falling due to reduced overtime, higher deductions, and the like, it feeds dissatisfaction and the desire to change jobs.
- High dependence on overtime: A situation where base pay is low and income is propped up by overtime is two sides of the same coin as dissatisfaction with base pay.
- A gap versus the internal average: Employees paid below the average for their grade or department are prone to a sense of unfairness driven by comparison.
Evaluation Signs
- Declining or stagnant ratings: A falling evaluation score—or one that stays flat for a long time—reflects declining motivation or a career that has hit a dead end.
- A drop in rank or role: The period immediately after a demotion or a reduction in role tends to carry a heightened risk of departure.
- A gap between evaluation and reward: When someone earns high ratings but no raise or promotion follows, it translates directly into the feeling of “not being valued.”
5. Don’t Judge on a Single Sign — Improve Accuracy by “Combining” Them
We’ve listed many signs so far, but there is an important caveat. You must not conclude “this person is leaving” from a single sign alone. Not everyone whose overtime drops is going to quit, and taking paid leave doesn’t necessarily mean someone is preparing to resign. Judging on a single indicator generates a flood of false alarms.
The key to greater accuracy is to look at multiple signals in combination. In fact, using its AI “Watson,” IBM was reported to have combined more than 34 HR variables to predict departures within six months at roughly 95% accuracy, and as a result cut turnover-related costs by about $300 million (CNBC). Signs that are weak on their own become powerfully predictive when multiplied together.
Doing this by hand, however, is unrealistic. Evaluating every employee across dozens of indicators—including the changes over time—every month is a task that’s next to impossible to run manually at a company of several hundred people. That is precisely where machine learning and automated analysis come in. Some studies suggest that introducing predictive analytics can cut voluntary turnover by up to 20%, but in any case, “looking at many indicators in combination, continuously” simply cannot happen without a system to support it.
6. What to Do Once You Catch a Sign — Detection Is Only the Entry Point
Even if you detect a warning sign, detection alone won’t stop the departure. Early detection is only the entry point; it only means something when it connects through to the actions that follow. The basic flow has three steps.
- (1) Detect the signs with data: Identify the high-risk employees and the factors driving that risk.
- (2) Manager 1-on-1s and stay interviews: Instead of asking after they’ve left, ask directly while they’re still on board—”what’s keeping you here, and what’s frustrating you?”
- (3) Individual review of compensation, career path, and workload: Take targeted action based on the factors you’ve surfaced.
At the center of this flow is the direct manager. Gallup finds that 70% of the variance in team engagement can be explained by the manager. Manager-driven dialogue is the single biggest lever for stopping turnover. Some studies suggest companies that conduct stay interviews saw turnover fall by 14.9%, yet only 17.2% of companies actually run them—making it a measure with a great deal of untapped upside for most organizations.
In Vietnam, there are particular moments when you should be especially deliberate about holding these conversations: around the 6-month mark after joining (when early departures cluster), in-demand roles (the segment facing strong external pull), and the quarter before Tet (Lunar New Year).
A Vietnam-Specific Consideration — Tet and the Turnover Cycle
In Vietnam, the typical turnover pattern has long been a “golden window” in which employees collect their Tet bonus and then resign, changing jobs after the holiday. It’s rational behavior—move only after receiving something you’re entitled to, the bonus. For that reason, the quarter before Tet is the period when you should be most sensitive to warning signs.
That said, even this “typical pattern” isn’t monolithic. For 2026, some reports point to a trend toward post-Tet turnover actually stabilizing, against a backdrop of rising living costs and other factors. Rather than taking the old conventional wisdom at face value, it’s important to verify the actual cycle with your own data.
7. Conclusion — From Gut Feeling to Proactive, Data-Driven Retention Management
Turnover doesn’t happen out of nowhere one day; people who leave have been leaving signs in your HR data for months beforehand. Capture those signs with data rather than gut feeling, and you can act before the resignation letter arrives—that is the core message of this article. Let’s recap the key points.
- Look at data, not myths: A resume left out or an interview suit is not a warning sign. The real signals are a quiet decline in engagement and output.
- Different types of signs have different lead times: Declining engagement shows up ~9 months out, performance decline 2–6 months out. The earlier the sign, the larger your window.
- Combine, don’t judge alone: A single indicator produces too many false alarms. Multiply multiple signals together to raise accuracy.
- Detection is the entry point—follow through to action: Connect detection to manager 1-on-1s and stay interviews, then to individual reviews of compensation, career, and workload.
Make the Warning Signs Visible — with EST
As this article has shown, manually tracking every employee across a large number of indicators—including the changes over time—every month is simply not realistic. That is exactly why you need centralized HR data and automated analysis of it. When your data is scattered across attendance software, Excel, and paper request forms, you can neither look at the signs in combination nor track risk continuously.
“EST,” an HR management system built for Vietnam’s labor practices, centralizes attendance, payroll, and evaluation data on a single platform, and comes equipped with “HR Insight,” where AI analyzes the accumulated data every month to detect high-turnover-risk employees early.
The warning signs cited in this article—spikes in overtime, disrupted attendance, low leave usage, stagnant evaluations, and more—are visualized by EST as quantitative risk scores. Because it also presents risk trends by department, identifies high-risk individuals, breaks down the risk factors, and surfaces signs of overwork (burnout), you can see at a glance where the risk is and why.
Once you can make the warning signs visible, managers can move before the resignation letter arrives. Move from retention efforts driven by gut feeling to proactive, data-driven retention management. Start by capturing your own turnover signs in “numbers.”
Related Articles
- Vietnam’s Turnover Rate in Numbers (Retention Series, Part 1)
- Vietnam’s 2026 Minimum Wage and Social Insurance Caps
- Guide to Vietnam’s Leave Entitlement System
