Seasonal attrition in travel is structural, not personal. People do not leave because of the peak - they leave immediately after it, and by then the decision was made weeks earlier.
If you run reservations, guest contact, sales or front-line service in a seasonal travel business, you already know the pattern. The team gets through the peak. Then, in the weeks straight after it, the resignations arrive - and disproportionately from the people you least wanted to lose.
For a summer-season operator that is September and October. For a ski business it is April. For an operator whose peak is a single event or a school holiday period it is whenever that ends. The month moves; the pattern does not.
It is tempting to read this as recovery: they were tired, they took stock, they moved on. That reading leads to interventions that do not work, because it locates the cause in the individual. The cause is structural, it is visible in advance, and by the time the resignation arrives the decision has usually been made some weeks earlier.
The decision is made in the peak, not after it
People rarely resign at the point of maximum pressure. They are too busy, and there is a strong social contract in seasonal operations about getting the team through. What happens during peak is that the decision is formed. What happens afterwards is that it is executed, once it is defensible to do so.
That gap matters practically. Anything you do after the peak addresses people who have already decided. The intervention window is the peak itself, which is exactly when nobody has capacity to think about retention.
What actually causes it
· The job narrows.Under peak load, the interesting parts of the job disappear first. Coaching stops. Development conversations are postponed. What remains is the highest-volume, most repetitive contact types, back to back, for weeks. Good agents are not exhausted by volume so much as by the disappearance of everything that made the role tolerable.
· Autonomy is withdrawn.In peak, agents lose discretion. Scripts tighten, handling times are watched, and the latitude to actually resolve something well is withdrawn precisely when customers are most upset. For experienced people this is the single most corrosive change, because it removes the part of the job they are good at.
· The tooling stops coping.Systems that are merely irritating at normal volume become genuinely unworkable at peak. Slow lookups, multiple screens, no context on the caller. An agent handling a hundred contacts a day feels every second of a five-second lookup, and feels it as the organisation not caring.
· The apology tax.In a bad peak, a large share of contacts begin with the agent apologising for something the organisation did - an overbooking, a cancellation, a delay, a queue. Sustained over weeks, absorbing displeasure you did not cause and cannot fix is the most reliable predictor of people leaving a front-line role.
Building your own replacement cost
The replacement cost of an experienced front-line agent is routinely underestimated, because most calculations capture recruitment and miss the larger components. There is no reliable published figure for this in European travel, so build it from your own five inputs. Every one of them exists somewhere in your business already.
1. Recruitment and agency cost per hire. Source: your own HR or finance reporting. This is the only component most models include, and it is usually the smallest.
2. Onboarding and training days before the new person takes live contacts unsupervised. Source: your own training schedule. Count trainer time as well as trainee time.
3. Time to competence, in weeks, and the productivity curve across it. Source: your own quality and handling data, comparing new starters against the team median. If you cannot produce this curve, that is itself worth reporting.
4. Supervisory hours absorbed by onboarding, taken away from the existing team. Source: ask your team leaders to estimate over one cycle. An estimate they own is more defensible than a number you assert.
5. Quality and rework cost during ramp - escalations, corrections and service recovery attributable to inexperience. Source: your own quality sampling.
State every assumption. A fully loaded replacement cost that finance can interrogate is far more useful than a headline multiple of salary, and it survives the meeting.
One component is almost always missing and is usually the largest: the load transferred to the people who stayed. That is the mechanism by which one departure becomes three.
Why hiring more seasonal staff makes it worse
The instinctive response to peak pressure is more seasonal hires. Done at scale it frequently increases attrition among permanent staff rather than relieving it.
Seasonal hires cannot take the complex contacts. So the complexity concentrates further on the experienced staff, whose job narrows in the opposite direction from the one described above: they get only the difficult work, with less support, while also being expected to coach the new arrivals. The permanent team ends up doing more of the hardest work and more of the training, in the same hours.
There is a version of seasonal hiring that works, and it depends entirely on whether the additional people absorb genuinely simple volume or whether they add a training load without removing work. That is a question about contact routing, not about recruitment.
What actually reduces it
· Take the repetitive volume out of the peak rather than adding people to absorb it.The objective is that experienced agents spend peak on work that uses their judgement, not that they handle more of everything.
· Protect coaching through peak explicitly, in the plan, as a non-negotiable.It is the first thing cut and the loss compounds.
· Give agents context before the conversation starts.Reducing the effort of every interaction is worth more to retention than most engagement initiatives, and it is measurable.
· Preserve some autonomy on resolution.The latitude to fix something properly is what makes the difficult contacts bearable.
· Run the retention conversation during the peak, not after it.Afterwards you are conducting exit interviews.
How to frame it internally
One caution. This case is frequently made in language that guarantees it will fail - efficiency, deflection, headcount. Framed that way it lands with the team as a plan to replace them, and it becomes self-fulfilling, because the fastest way to lose your best agents is to tell them their work is being automated.
The accurate framing is also the more persuasive one: the objective is to remove the work people hate so that experienced staff spend peak doing the parts of the job they are good at. That is what actually happens when it is done well, and it is what the people affected need to hear first - from their own leadership, before they hear it anywhere else.
Read the country research: workforce pressure, seasonality and the contact layer