Every educational institution loses students, yet few can explain why. The question always arrives at the end of the term, when the number has already closed and the room to act is gone.
The issue is not lack of interest, it is lack of organized data. The school dropout rate shows up in reports as a loose percentage, stripped of the signals that produced it, and without those signals no retention action has a target.
This article reorganizes the subject in two parts. First, what national research shows about the size of the problem. Then, which internal variables you can actually measure and change starting Monday.
The school dropout rate is the share of enrolled students who leave a program without completing it within a given period. In higher education, the most common calculation compares one year's enrolled students with those who continued the following year, minus graduates and internal transfers.
There is a second, harsher and more useful reading: cumulative desertion by cohort. It follows one intake group across the full program cycle and measures how many reached a diploma.
The two measures tell different stories. The annual figure looks manageable, the cumulative one reveals the real size of the loss.
Dropout and abandonment are also not the same thing. Abandonment is departure without any formal notice, and it is the most common case in the private network, because the student simply stops paying and stops showing up.
That distinction matters operationally. Institutions that count only formal cancellations underestimate the problem and react late, when the record reaches the system weeks after the last class attended.
Brazilian higher education loses roughly a quarter of on-campus students and more than four in ten distance learning students every year. The figures come from the 16th Higher Education Map, published by Instituto Semesp in 2026 with 2024 as its base year.
Broken down by delivery mode and network, the concentration of the problem becomes visible:
|
Delivery mode |
Private network |
Public network |
Total |
|
On-campus |
26.6% |
21.4% |
24.8% |
|
Distance learning |
41.9% |
32.2% |
41.6% |
Table: Annual undergraduate dropout rate by delivery mode and network in 2024, according to Instituto Semesp.
The cohort reading is even more severe. Among 2020 entrants in the private network, 64.7% had dropped out by 2024, and in that network's distance learning programs the figure reached 68.1%.
Institution size also shows up in the data. At mega institutions, cumulative desertion was 69.2%, against 53.3% at smaller ones.
That contrast matters because it dismantles the idea that dropout is simply market destiny. There is a gap of almost sixteen points between groups of institutions operating in the same country, in the same period, under the same regulatory framework.
The weight of distance learning explains part of the urgency. The 2024 Higher Education Census, run by Inep, shows the format already accounts for 50.7% of undergraduate enrollments in the country.
In other words, the fastest growing format is also the one losing the most students. Any retention strategy designed for on-campus programs has to be rebuilt for this context.
When a student chooses an institution, they build an expectation about faculty, facilities and class format. The school dropout rate rises when the delivered experience falls short of that expectation, not when teaching is objectively poor.
The difference is subtle, but it changes the action. An institution can have strong faculty and high dropout, because it promised flexibility and delivered a fifty-minute lecture on video.
In one of the webinars of our permanent student retention forum, professor Leonardo Vils summed up the mechanism in a single line: “Satisfaction is a disconfirmation of expectation.”
That reading forces satisfaction to be measured stage by stage, not only at the end of the term. A short survey in the second week of class captures misalignment while it is still fixable.
In distance learning, the risk is higher because the student has less social anchoring. With no classmate nearby and no campus routine, dissatisfaction meets no friction and turns into a silent exit.
Formats that reduce this effect share a pattern: short blocks, hands-on activity from week one and some recurring human touchpoint. None of them depends on a new platform.
Inability to pay tuition is the most decisive variable behind the school dropout rate in the private network. It is also the easiest to detect early, because it leaves a trail in late payments, renegotiations and falling attendance well before formal withdrawal.
The problem is that this trail usually lives in finance and never talks to academic coordination. By the time the information reaches someone who could act, the student has already decided.
External factors amplify the effect. Job loss in the family, moving city and rising transport costs all enter the equation without ever appearing on an institutional form.
One simple question is worth checking in your operation: how many days pass between the first late payment and the first retention contact? If the answer is more than two weeks, the useful window has already closed.
All of these actions depend on identifying the right student at the right moment, and none of them works from a report closed at the end of the term.
In basic education, the same mechanism wears different clothes. Children's education is rarely the first expense cut, but it joins the queue when household income drops abruptly.
Weak academic performance and long programs form the quietest combination behind dropout. A student who fails a course in year one and sees four more years ahead starts comparing effort against return, and that comparison usually ends in withdrawal.
Semesp data helps size the effect. Between 2023 and 2024, the number of on-campus graduates fell 6.9%, while new entrants declined 1.0%.
The likely reading is that the loss concentrates mid-journey, not at the front door. Those who leave had already invested several terms, which means money lost on both sides.
Two pieces of information solve most of the diagnosis: failure history by course unit and completion rate by program. The first shows where the academic bottleneck sits, the second shows which program is bleeding.
Course units with failure rates well above the program average deserve their own treatment. Targeted tutoring, alternative scheduling and a revised syllabus attack the cause instead of reacting to the number.
It also matters to measure satisfaction with the program itself, not only with the institution. A student unhappy with their career choice needs guidance, not a discount.
Student retention strategies that work share three traits: they start from data, they act before the cancellation request and they treat different causes with different responses. A blanket discount for everyone is expensive and solves little, because it addresses only one of the variables.
A lean design starts with risk segmentation. Financial, academic and engagement are three distinct tracks, each with its own owner, deadline and offer.
The contact cadence comes next. A detected risk signal should generate a task with a deadline, not a monthly report for later reading.
The third element is the one most often forgotten: closing the loop. Every retention contact needs to record the reason the student reported, because that field feeds the next term's diagnosis.
A fourth element is reviewing the first weeks of experience. Much of the abandonment is decided before the first assessment, while the student is still testing whether they belong there. A structured onboarding script has a disproportionate effect on student retention across the term.
Institutions that apply artificial intelligence to retention service can cover a larger volume of students without growing the team, as long as automatic triage hands complex cases to a person.
Using a CRM for educational management means gathering, in a single record per student, the signals currently scattered across the registrar, finance, the virtual learning environment and support. That unification, rather than the software itself, is what makes risk predictable and early action possible.
In practice, four sources cover most of the diagnosis: payment status, attendance, grades and interaction history. With all of them in one base, risk stops being an opinion and becomes a queryable field.
From there, the CRM takes on three jobs. It consolidates the data, triggers the retention task when the condition is met and records the outcome of every contact.
Predictive models sit on top of that layer to order the queue. Instead of treating all late payers alike, the team receives a list of who is most likely to leave and most likely to respond.
There is one non-optional caution. Student data is personal data, so the operation has to observe legal basis, purpose and retention period.
The most common mistake here is deploying the tool before defining the indicator. Without prior agreement on how the school dropout rate is calculated, each department keeps its own number, and the discussion never leaves the report. Settling the most relevant metrics for higher education institutions before implementation saves months.
Educational marketing strategies affect the school dropout rate before enrollment, at the moment the student's expectation is built. A campaign that promises what the program does not deliver buys enrollment expensively and returns dropout within three months.
That makes campaign promises a retention item. Describing real workload, class format and academic demand reduces lead volume and improves the quality of what comes in.
The second touchpoint is communication with students already enrolled. Most institutions switch off their communication cadence on enrollment day, exactly when the bond is weakest.
Support content, useful notices and fast response channels perform a retention function without feeling like collections. The same repertoire used to attract works to retain, with a different objective.
The third point is alignment between departments. When enrollment is measured only by volume and retention only by dropout, the two targets contradict each other, and the institution pays the difference. A strategic educational marketing design measures both stages on the same dashboard.
Start with the indicator, not the tool. Defining how the school dropout rate is calculated at your institution, with a single formula accepted by academic, finance and marketing teams, solves more in the first month than any system rollout.
With the indicator settled, the next step is bringing the four signal sources into one base and choosing a single trigger to operate. A first late payment and two consecutive weeks of absence are good candidates, because they appear early and allow action.
After that, the retention cadence can grow by segment. What does not work is starting with a blanket discount, which burns margin and tells you nothing about the cause.
If you want to discuss how this applies to your scenario, with your dropout figures and your data maturity, talk to our team.