Turnover in Startups: How to Calculate the Rate and Retain Talent in 2026
High turnover is not an HR problem: it is a unit economics, product speed, and team morale problem. This guide shows how to calculate the monthly and annual rate, how much it really costs to replace a person, what signs anticipate resignation requests, and a 30, 60, and 90-day plan to retain talent.
In a startup, losing a person rarely costs only the replacement process. It costs context: the person who leaves takes the memory of decisions that were never documented, the agreements with clients, the shortcuts of a code that only they knew. When this happens two or three times in the same quarter, the roadmap stops moving even with the team full on the organizational chart.
That is why turnover deserves the same rigor as a financial indicator. It has a formula, it has a measurable cost, it has antecedent signs, and it has correction levers. Treated like this, it stops being a hallway conversation topic and becomes a number that enters the board meeting alongside churn, CAC, and burn.
The consolidated research on the topic is clear on one point: most voluntary exits are avoidable and are linked to career, direct management, and work environment — not to an isolated salary below the market. The Work Institute's Retention Reporthas been pointing out this pattern for years, and Gallup's State of the Global Workplace reinforces the other side of the equation: low engagement precedes the intention to change jobs.
In this guide, the logic is practical. First we define what turnover is and separate voluntary from involuntary turnover. Then we calculate the monthly and annual rate, measure the real cost of replacing someone, list the signs that anticipate resignation requests, explain how onboarding, leadership, and career plans change the game, and we close with a 30, 60, and 90-day plan and with the indicators you need to follow besides turnover.
💡 The founders' biggest mistake: treating each exit as an isolated case ("she received a better offer") instead of looking at the pattern. Pattern is data; an isolated case is an anecdote — and an anecdote cannot be corrected.
What is turnover and what is the difference between voluntary and involuntary turnover?
Turnover is the rate of movement of people entering and leaving an organization in a period. It is a flow indicator, not a stock indicator: it doesn't tell you how many people you have, it tells you how fast they are exchanged. And, like any flow indicator, it only makes sense when it is segmented. Aggregate turnover hides the information that matters.
Voluntary turnover
The person decides to leave. It is the most diagnostic number that exists about the quality of your management, because it means that someone you chose, trained, and who already knew the context preferred to start over somewhere else — assuming the risk of trading the known for the unknown.
Within it, there are two very different species: the regretted exit (a person you wanted to keep) and the unregretted exit (a person whose exit relieves the team). Startups that do not make this distinction celebrate a drop in turnover that is actually the retention of low performance.
Involuntary turnover
The company decides to end the bond: dismissal for performance, restructuring, end of a project, or cost-cutting. Contrary to what many founders think, a high number here is not a sign of rigor — it is a sign of a previous systematic error, almost always in hiring or in the alignment of expectations.
If half of the people hired in the last twelve months were dismissed by the company, the problem is not with them: it is in the designed profile, in the selection process, in the job description, or in the absence of a structured adaptation period with explicit goals.
Healthy turnover exists
Zero turnover is also a bad sign. A totally static team in a company that changes stages every year usually indicates accommodation, absence of a performance bar, or salaries above the market without compensation. The goal is never to eliminate turnover; it is to reduce regretted and avoidable turnover, keeping the natural renewal that brings a new repertoire. A good cut for startups is to track separately: total turnover, voluntary turnover, regretted turnover, and turnover in the first 90 days.
The three Shinier working principles
Every person who joins Shinier — and every developer trained and allocated by us — goes through the same onboarding, conducted by the CEO himself. We don't start with a tool or a process: we start with three principles that support intensity without breaking people. It is our practical response to the problem of turnover.
Work when you are well
Mental Health & Respect
Productivity is not physical presence or logged hours. We prefer that the person stops, warns, and returns whole than pushes a bad week into the codebase.
Intensity and Constancy
Sustained Pace without Burnout
Constancy beats heroism. Four regular weeks deliver infinitely more than one all-nighter followed by three weeks of exhaustion and bugs.
Daily Feedback without Guilt
Transparency without Guilt
Daily reporting of what worked and what failed. Daily alignment destroys the silent accumulation of dissatisfaction that usually culminates in resignation requests.
These three principles are not a wall poster: they appear in the daily ritual, in the way the leader reacts to an error, and in the explicit right to stop. That's why we manage to keep lean teams on long projects — and that is exactly what we take to the startups in the acceleration program.
Bring these principles to your team now
In Shinier's acceleration program, every dev trained and allocated by us operates under the same three principles — and your startup receives the OKR, Kanban, and team management tools to apply the same method in retention.
Get to know the acceleration programHow to calculate the monthly and annual turnover rate?
The classic turnover formula considers entries and exits because it was born to measure movement in companies with a stable headcount:
Turnover (%) = [ (admissions + dismissals) ÷ 2 ] ÷ average headcount × 100
Except that a growing startup hires a lot. If you use this formula during an expansion, the indicator goes up for a good reason — which hinders the reading. Therefore, for retention, prefer the dismissals formula:
Exit turnover (%) = dismissals of the period ÷ average headcount × 100
The average headcount is the average between the number of people on the first and last day of the period. In a month that started with 40 people, ended with 44, and had 2 exits: average headcount 42, monthly turnover of 2 ÷ 42 = 4.8%.
For the annual, do not naively multiply the month by 12 if the base varied a lot — sum the dismissals of the twelve months and divide by the average headcount of the year. In the example above, if the year closed with 22 exits over an average headcount of 45, the annual rate is approximately 48.9%: almost half of the company exchanged in a year.
Three precautions that change the quality of the number: keep the same formula over time (changing methods invalidates the historical series), calculate by area and by seniority (the aggregate hides the team that is bleeding), and also follow the survival curve — what percentage of the hired people continue after 3, 6, 12, and 24 months.
How much does it cost to replace a person in a startup?
The account that most people do is the severance pay. The real account includes at least five blocks, and the most expensive of them never appears on the bank statement: the productivity that did not exist while the vacancy was open and while the new person went up the ramp. Combined, these blocks usually fall between 50% and 200% of the annual salary of the position — the more senior and more specific the knowledge, the closer to the ceiling.
Exit cost
Severance pay, proportional vacation and 13th salary liabilities, team hours in knowledge transfer, and often, emergency hiring of a freelancer to plug the hole.
Recruitment and selection
Ads, ATS tools, headhunter (when applicable, from 15% to 25% of the annual salary), and the hours of founders and leaders in screening, technical interviews, and cases — the item that consumes the most time from those who should be selling or building a product.
Onboarding and training
Setup, access, documentation, mentoring from the follower, and the time of colleagues dedicated to explaining context. In engineering teams, each new person consumes senior hours for weeks.
Productivity ramp
The interval between admission and full performance varies from 1 to 3 months in operational functions and reaches 6 months in technical or complex sales functions. During this period you pay 100% and receive a fraction.
Team overload
While the vacancy is open, the work does not disappear: it is redistributed. Sustained overload generates delivery delays, drop in quality, and — the worst effect — new exits, creating the classic domino effect of startups in crisis.
Knowledge loss
Undocumented decisions, relationships with clients, context of technical debts, and informal agreements. It is the most difficult cost to estimate and the one that most delays the roadmap in the following three months.
An honest numerical example
Consider a mid-level developer with a total monthly cost (salary + charges + benefits) of R$ 15,000 — R$ 180,000 per year. Applying a conservative estimate of 80% of the annual cost, the replacement comes out to about R$ 144,000: approximately R$ 20,000 in exit and severance, R$ 24,000 in recruitment (including leadership hours), R$ 30,000 in onboarding and mentoring, and R$ 70,000 in lost productivity during an open vacancy and ramp.
In a 25-person startup with an annual turnover of 30%, there are 7 or 8 replacements per year. Even using a lower average cost, the bill easily exceeds half a million reais annually — usually more than the company's entire marketing budget, and without a single line in the P&L called "turnover". It is this number that turns retention into a board agenda item.
What signs anticipate resignation requests?
The turnover literature describes the exit as a process, not an event. It begins with a persistent dissatisfaction or a "shock" (a denied promotion, a change of manager, a canceled project, a received proposal), passes through an alternative evaluation phase, and only then becomes the intention to leave. Between the shock and the request there are usually weeks or months — and it is in this interval that leadership can still act.
Behavioral signs
- Visible reduction in initiative: the person delivers what was agreed, but stops proposing improvements and disagreeing in meetings.
- Disinterest in long deadlines: avoids taking on projects with delivery six months from now or treats annual planning with irony.
- Social withdrawal: disappears from informal rituals, conversation channels, and spontaneous interactions with the team.
- Generic feedback in 1:1s: short answers, 'everything is fine', lack of requests and complaints — silence usually comes after the request was not met.
- Increase in short absences and isolated days off, often for interviews.
Structural and health signs
- More than 12 months without a formal career conversation or without any movement in scope, range, or responsibility.
- Recent change of direct manager — one of the biggest predictors of exit in any size of company.
- Chronic overload: recurring overtime, informal shifts, and vacations postponed for two consecutive cycles.
- Burnout symptoms: persistent tiredness, cynicism towards work, and a drop in perceived effectiveness — the picture that the WHO associates with psychosocial risks in the workplace.
- Accumulated salary lag compared to the market, especially after a hiring round with higher ranges.
The tool almost no one uses: the stay interview
An exit interview is an autopsy: it informs, but it does not save the patient. The stay interview is the preventive exam — a 30-minute conversation, every quarter, outside the performance review, with four questions: what makes you stay here? what almost made you leave in the last few months? what would you like to learn or take on in the next semester? what should I, as a leader, start, stop, and continue doing? Record the answers and review them in the following quarter. The simple act of asking beforehand already reduces the surprise effect.
How do onboarding, leadership, and career plans affect retention?
Three levers explain most of the turnover variation between teams in the same company — same remuneration, same product, same declared culture, completely different exit rates.
Onboarding: the first 90 days decide
A disproportionate share of exits occurs in the first year, and a good part of them in the first three months. The reason is rarely competence: it is misaligned expectation. The person was hired for a position that, in practice, is another; no one explained how success is measured; they spent two weeks without access to systems; they didn't know who to ask. A minimum viable onboarding has four items: a document of expectations with 30/60/90-day goals, a sponsor outside the hierarchical line, a first real delivery in the first week, and a formal check-in on day 30 and day 90.
Leadership: the variable with the greatest impact
In a startup, a leader is usually the best specialist promoted — without any training to give feedback, conduct 1:1s, or deal with conflict. The result is predictable: good people stay away from bad management even if they like the product. Investing in leadership training (with a weekly cadence of 1:1s, explicit evaluation criteria, and practice of difficult conversations) is the most cost-effective intervention of the entire retention theme.
Career plan: predictability instead of promise
A startup doesn't need a multinational's job plan, but it needs to answer three questions: what are the levels of this role, what differentiates one level from the next in observable behavior, and when the evolution is evaluated. Without this, promotion becomes an individual negotiation — and individual negotiation feeds the perception of injustice, which is a direct fuel for exit.
Tying everything together is the organizational culture. Schein remembers that culture is not the values on the wall, but the assumptions that leadership demonstrates when it needs to decide under pressure. If the stated value is balance and the team is praised for pulling an all-nighter, the real assumption won — and the turnover responds to it, not to the poster.
30, 60, and 90-day action plan to reduce turnover
Do not try to solve everything in one culture initiative. Turnover drops with sequence: first measure, then diagnose, then change the system. This is the roadmap we apply with the accelerated startups at Shinier.
Measure & See Patterns
Calculate total, voluntary, and first-90-day turnover over the last 12 months. Build the survival curve and review exit interviews.
Listen & Correct Fast
Run stay interviews and short surveys. Perform salary benchmark, standardize weekly 1:1s, and adjust onboarding with visible deliverables.
Change the System
Create the first explicit matrix of levels for the most critical areas. Train leaders in feedback and difficult conversations.
Do not look only at turnover
Turnover is a lagging indicator: when it goes up, the problem happened months ago. To act proactively, track leading indicators: eNPS (Employee Net Promoter Score) to measure the general temperature, the percentage of 1:1s held versus planned (to measure leadership commitment), the rate of acceptance of job offers (to measure market attractiveness), and the absenteeism rate (which usually precedes the burnout exit).
Referências
- GALLUP. State of the Global Workplace. It is a reference because it tracks, year after year and in more than 140 countries, engagement, stress at work, and the intention to change jobs with the same methodology — which allows you to compare your startup with a consistent global base instead of with internal perception. Access the Gallup report
- WORK INSTITUTE. Retention Report. It is a reference because it classifies the stated reasons for leaving based on large-scale exit interviews, showing that most exits are avoidable and linked to career, management, and environment — not to an isolated salary. Access the Retention Report
- HOM, P. W. et al. One Hundred Years of Employee Turnover Theory and Research. It is a reference because it consolidates a century of academic research on turnover in the Journal of Applied Psychology, organizing the models of exit intention, shocks, embeddedness, and exit paths that support any serious turnover diagnosis. View on APA PsycNet
- WHO. Mental Health at Work. It is a reference because it brings the technical position of the World Health Organization on psychosocial risks, burnout, and productivity loss — the bridge between mental health, absenteeism, and turnover that many startups ignore until it becomes a crisis. Access the WHO fact sheet
- SCHEIN, Edgar H. Organizational Culture and Leadership. It is a reference because it defines organizational culture as layers of artifacts, espoused values, and basic assumptions — the model that explains why culture rituals and speeches do not hold people when the real assumptions of leadership say otherwise. View on Amazon