Learning Mode Adoption Index 2026
The relative share of online and in-person private tutoring across regions and levels, and how learning settled after the remote-learning surge.
Abstract
In-person learning still leads Indonesia's private tutoring in 2026, holding around seven in ten sessions, while online learning has settled near one in five after the pandemic-era surge receded. This report builds a learning mode adoption index across regions, levels, and subjects using indexed data and official statistics.
Executive Summary
How a family chooses its learning mode reveals a great deal about access, trust, and a child's needs. This report focuses on a single axis: the balance between in-person and online learning within Indonesia's private tutoring during the first half of 2026. Rather than counting absolute totals, we build an adoption index that reads the relative share of each mode across regions, levels, and subjects.
The broad picture is clear. In-person learning remains the backbone, holding around seven in ten sessions, because families still value a tutor's physical presence and the ease of watching over a child. Online learning occupies roughly one in five sessions, supported by national internet penetration that has passed three quarters of the population. Between the two, a blended model is growing as a middle path that combines the strengths of both.
Online adoption differs sharply by geography and age. Metropolitan areas with mature infrastructure lead, while regional cities and districts move more slowly, mirroring the digital divide recorded in official statistics. The older the learner, the larger the online share, and the type of subject matters too: digital skills and foreign languages move to the screen with ease, while music, Quran recitation, and early literacy stay anchored in physical presence.
Most striking is the post-normalization pattern. After the pandemic-era surge in remote learning subsided and schools returned to in-person classes, online adoption did not fall back to where it began. It settled at a level far higher than before the pandemic, forming a stable new balance. All internal figures are presented as indices to protect confidentiality, while factual claims rest on official sources. As a snapshot of the first half of 2026, this study will continue in a later edition with a longer time window so its long-term direction reads more sharply.
Key Findings
- 1In-person learning holds around seven in ten private tutoring sessions, while online learning takes roughly one in five and blended models fill the remainder.
- 2Online adoption peaks in the greater Jakarta metropolitan area and falls sharply toward regional cities and districts, tracking the national digital divide.
- 3The higher the level, the larger the online share: university students and professionals lead adoption, while early learners lean almost entirely on in-person sessions.
- 4Subject shapes mode: programming, artificial intelligence, and foreign languages are most at home on screen, while music, Quran recitation, and early literacy hold to the in-person room.
- 5The post-normalization pattern confirms online learning holding well above its pre-pandemic point without repeating the restriction-era peak, forming a stable new balance.
Table of Contents
- 01Introduction: Learning Mode as a New Axis
- 02Context: The Digital Foundation of Online Learning
- 03From Remote Learning to a New Normal
- 04Learning Mode Composition 2026
- 05Online Adoption Index Across Regions
- 06Online Adoption Index Across Levels
- 07Subject Affinity and Learning Mode
- 08The Rise of the Blended Model
- 09Drivers of Online Learning Adoption
- 10Access, Equity, and the Digital Divide
- 11Implications and Recommendations
- 12Outlook: Where Learning Modes Head in 2026 and Beyond
- AAppendices: Methodology & Limitations
- BAppendices: References
Introduction: Learning Mode as a New Axis
The choice between in-person and online has become a meaningful family decision, and an adoption index reads the relative share of each mode without disclosing absolute totals.
For many years, private tutoring meant a teacher coming to the house and sitting beside a student. That way remains alive and strong, yet it now shares space with a newer form: a class that unfolds through a screen, linking a student in one city with a teacher in another. Choosing between the two has become part of every family's learning plan.
This report centers on that comparison. The core question is simple yet layered: how far have Indonesian families shifted toward online learning, in which regions adoption is fastest, at which age levels it strengthens, and which subjects move most easily to the screen. We weave the answers into a learning mode adoption index.
Why an index, not absolute numbers
Every internal figure in this report is presented as an index: a value relative to a base, expressed as a share or ranking, with session counts and rupiah amounts withheld. This approach protects operational confidentiality while surfacing what serves the reader most: the structure and direction of adoption. Measures such as shares and indices reveal national market patterns without exposing raw data.
Facts beyond the platform, such as internet penetration and education policy, are cited directly from official institutions and clearly flagged. The pairing of these two lanes, indexed on the internal side and cited on the external side, forms the report's foundation of credibility.
Context: The Digital Foundation of Online Learning
Online learning stands on a steadily strengthening digital network, with national internet penetration past three quarters of the population yet uneven between city and village.
Adoption of online learning rests on one basic prerequisite: connection. The Indonesian Internet Service Providers Association (APJII) 2024 survey recorded national internet penetration at 79.5 percent, equal to 221.5 million people. This figure gives ample room for classes conducted through a screen.
Data from Statistics Indonesia (BPS) reinforce this picture from the usage side. According to the 2024 Susenas, 72.78 percent of the population had accessed the internet, up from 69.21 percent in 2023. Mobile phone ownership reached 68.65 percent of the population, the primary device that lets an online class take place from almost anywhere.
| Indicator | 2024 value |
|---|---|
| Population accessing internet | 72.78% (2023: 69.21%) |
| Mobile phone ownership | 68.65% |
| Household computer ownership | 18.52% |
The gap between city and village remains an important caveat. BPS recorded that among people aged five and over, 79.13 percent accessed the internet in urban areas against 63.71 percent in rural areas. This divide then shapes the adoption map: online learning takes root most easily where the connection is most mature.
Source: APJII, Indonesia Internet Penetration Survey 2024; BPS, Telecommunication Statistics in Indonesia 2024.
From Remote Learning to a New Normal
After the pandemic-era surge in remote learning subsided and schools returned to in-person classes, online adoption settled far above its pre-pandemic point, forming a stable new balance.
Adoption of online learning in Indonesia did not grow gradually, it was forced to leap. When the pandemic closed classrooms in 2020, remote learning became the only option for nearly every student. That sudden jump introduced millions of families to screen-based classes within a matter of weeks.
The next phase was the return of live meetings. Through a Joint Decree of Four Ministers, the government pushed in-person learning to resume in the 2021/2022 school year, part of an effort to recover the learning outcomes eroded during the remote period. Physical classrooms returned to the center, and in-person learning recovered as the primary mode.
Illustrative index, the 2024-2026 post-normalization level is set at 100. The shape of the curve is anchored to the official narrative of the transition from remote learning to in-person schooling, not an internal absolute time series.
The key point of the story lies in what happened after recovery. Online adoption did fall from its restriction-era peak, yet it did not return to the starting point. Families who had tasted the flexibility of online learning kept part of that habit, especially for needs that suit the screen. The result is a new balance: online learning holds well above its pre-pandemic level without repeating the dominance of the restriction period.
Source: Joint Decree of Four Ministers on learning during the COVID-19 pandemic, Ministry of Education, Culture, Research, and Technology; internal EduPoint analysis (indexed).
Learning Mode Composition 2026
In-person leads with around seven in ten sessions, online learning takes roughly one in five, and the blended model fills the remainder.
In the first half of 2026, the composition of private tutoring modes shows a clear direction. In-person remains the majority choice for families, followed by online learning that has settled as the leading alternative, then the blended model as a maturing middle path.
- In-person≈71%
- Online≈23%
- Blended (hybrid)≈6%
Relative share across modes (percent), based on internal demand composition February-June 2026. Absolute figures are not shown.
These figures place Indonesia in a distinctive position. In-person learning endures as the backbone because parents value a teacher's physical presence, the ease of watching over a child, and warmer interaction. Online learning, at roughly one in five sessions, has moved past the trial phase and become a permanent part of the learning landscape.
The blended model deserves attention despite its small share. It reflects families who do not pick one mode outright but mix the two as needs demand, for instance in-person for concept reinforcement and online for routine practice. This composition becomes the starting point for reading adoption by region, level, and subject in the chapters that follow.
Source: Internal EduPoint analysis (indexed), private tutoring demand composition February-June 2026.
Online Adoption Index Across Regions
Online learning adoption peaks in the greater Jakarta metropolitan area and tapers toward regional cities and districts, following the national digital divide.
The adoption map almost copies the map of digital infrastructure. The greater Jakarta metropolitan area, with the most mature connectivity and the densest mobility, holds the highest adoption index. Large cities on Java follow, then Bali and major metropolitan centers beyond Java, while regional cities and districts sit below the national average.
Relative index, national average = 100. Values above 100 indicate online adoption higher than average; absolute figures are not shown.
This pattern intersects with BPS data on internet access across provinces. Jakarta was recorded as the region with the highest share of internet users, around 87.84 percent of the population, while the highland areas of Papua sit at the lowest point, around 6.76 percent. A range this wide explains why adoption of online learning cannot be uniform.
There is a promising other side. It is precisely in regions where local teacher supply is still developing that online learning offers the greatest value, bridging families to quality teachers from other cities. As digital infrastructure catches up, these regions stand to close the adoption gap quickly, turning the online mode into an equalizer of access.
Source: Internal EduPoint analysis (indexed); BPS, Telecommunication Statistics in Indonesia 2024 (share of internet users by province).
Online Adoption Index Across Levels
The higher the level, the larger the online share: university students and professionals lead adoption, while early levels lean entirely on in-person learning.
A learner's age is one of the strongest determinants of online adoption. Independence, focus span, and device familiarity all grow with level, and together they make screen-based learning feel increasingly natural. University students and professionals preparing certifications or job skills lead the online adoption index, well above the national average.
Relative index, national average = 100. Higher values indicate a greater lean toward online; absolute figures are not shown.
Senior high school and test preparation sit above average, because dense material and repeated practice suit the flexible online format. Moving down to junior high, adoption nears the average, then weakens at primary school. At the early-years level, in-person learning is nearly universal.
This gradient is consistent with the logic of child development. The younger the learner, the greater the need for physical presence to hold attention and build learning habits. The older the learner, the more capable of managing an online session independently. The adoption index across levels, therefore, maps a curve of learning independence.
Source: Internal EduPoint analysis (indexed), private tutoring demand February-June 2026.
Subject Affinity and Learning Mode
The type of subject sets the mode: digital skills and foreign languages are most at home on screen, while music, Quran recitation, and early literacy hold to in-person sessions.
Not every subject moves to the screen with ease. The nature of the material sets how comfortably it is taught online. Programming and artificial intelligence, born in a digital environment, hold the highest lean toward online. Sharing a screen, writing code together, and reviewing projects feel natural at a distance.
- 1Programming & AI
- 2Foreign languages (English, Mandarin)
- 3Test preparation
- 4Math & Science
- 5Quran recitation
- 6Music (piano)
- 7Early literacy (PAUD)
Relative index, the cluster most at home online = 100. It shows mode lean, not demand volume; absolute figures are not shown.
Foreign languages follow closely. English or Mandarin conversation can flow smoothly over video, and access to speakers from various cities lets the online mode widen the range of choices. Test preparation along with Math and Science sit in the middle, using online for practice and in-person for reinforcing concepts that call for step-by-step explanation.
At the other end, music, Quran recitation, and early literacy hold firmly to in-person learning. Teaching finger placement on piano keys, correcting the articulation of letters, or guiding a child's hand through a first letter calls for physical presence that a screen struggles to imitate. This subject affinity explains why the national mode composition will never be uniform.
Source: Internal EduPoint analysis (indexed), private tutoring demand February-June 2026.
The Rise of the Blended Model
The blended model grows as a middle path that pairs the warmth of in-person learning with the flexibility of online, mixing two modes within one learning arc.
Between the two poles of in-person and online, a third way is drawing growing interest. The blended model combines both within one learning arc, giving families the latitude to choose the mode that best fits each need. Its share is still small, around one in sixteen sessions, yet its direction of growth is worth reading.
Its forms vary. Some families use in-person sessions to introduce new concepts that call for deeper explanation, then switch to online for practice and repetition. Others schedule live meetings on weekends and short online sessions on weekdays, adjusting to the rhythm of a busy week.
The maturity of the blended model rests on two things: a teacher's readiness to manage two formats at once and a family's comfort in moving between modes. When both are met, blended learning can become the default choice for families in cities with mature connectivity, offering a balance that no single mode provides.
Source: Internal EduPoint analysis (indexed), private tutoring demand composition February-June 2026.
Drivers of Online Learning Adoption
Adoption is driven by distance to quality teachers, schedule flexibility, the nature of the subject, device availability, and cost considerations.
A family's decision to move toward online learning is not the result of a single factor, it is a blend of several considerations that reinforce one another. Understanding these drivers helps read where adoption is heading.
- Distance to quality teachers: in regions with limited local teacher supply, online learning opens access to teachers from other cities.
- Schedule flexibility: online sessions cut travel time, making it easier for families to fit learning into a busy day.
- Nature of the subject: material built on discussion, practice, and digital projects moves to the screen smoothly.
- Device and connection availability: widespread mobile phone ownership lets online classes run from any home within internet reach.
- Efficiency considerations: without a teacher's travel cost, some families view online as the practical choice.
These factors work differently for each family. For families in large cities, schedule flexibility is often the main reason. For families in developing regions, access to quality teachers from other cities stands out most. This difference in weighting shapes the range of adoption indices mapped in the earlier chapters.
Source: Internal EduPoint analysis (indexed); BPS, Telecommunication Statistics in Indonesia 2024.
Access, Equity, and the Digital Divide
The digital divide between city and village limits online adoption in developing regions, making infrastructure equity the key to equalizing access to learning.
Online learning carries a double promise. On one side, it can shorten the distance between families and quality teachers, especially in regions where local teacher supply is still growing. On the other, it only works when a connection exists, so the digital divide can move an old inequality into a new space.
| Area | Population accessing internet |
|---|---|
| Urban (age 5+) | 79.13% |
| Rural (age 5+) | 63.71% |
A gap of about fifteen percentage points between city and village is no small number. It marks millions of families who, though they need learning support, cannot yet make full use of online classes. For them, adoption of the online mode is held back not by interest but by the limits of connection.
The good news is that the direction of the data points upward. The share of the population accessing the internet rises year on year, and network expansion programs continue across many regions. Every advance in infrastructure stands to widen the room for online adoption, turning the promise of equalization into a more even reality.
Source: BPS, Telecommunication Statistics in Indonesia 2024; internal EduPoint analysis (indexed).
Implications and Recommendations
The adoption index findings steer the decisions of families, service providers, and policymakers as they choose and balance learning modes.
For families
- Match the mode to a child's age: early levels thrive best with in-person learning, while older learners can draw on the flexibility of online.
- Match the mode to the subject: choose in-person for music, Quran recitation, and early writing, and consider online for programming, foreign languages, and test practice.
- Use the blended model to pair the warmth of live meetings with the efficiency of online sessions.
For service providers
- Strengthen online capacity to reach regions with limited local teacher supply, making the online mode an equalizer of access.
- Prepare teachers skilled in managing both formats, so the blended model can be offered with quality intact.
- Differentiate the approach by level and subject, tailoring the mode to fit each need.
For policymakers
- Expanding rural digital infrastructure will directly widen the room for even adoption of online learning.
- Digital literacy programs for families and teachers can speed a healthy uptake of the online mode.
- Attention to access gaps across regions keeps the shift toward online from widening inequalities that already exist.
Source: synthesis of report findings; internal EduPoint analysis (indexed).
Outlook: Where Learning Modes Head in 2026 and Beyond
In-person learning will remain the backbone, online adoption will widen as infrastructure matures, and the blended model will strengthen into an increasingly common choice.
Three directions appear to strengthen in the years ahead. First, in-person learning will remain the backbone, especially at early levels and for subjects that demand physical presence. Second, online adoption will keep widening as digital infrastructure matures, catching up regions long held back by connection. Third, the blended model will strengthen, growing from a middle path into an increasingly common choice in cities with mature networks.
Because this study captures a single half-year, the direction it reads is indicative. A later edition will stretch its time window so seasonal patterns and year-to-year shifts stand out more clearly, building the foundation for a periodic and richer learning mode adoption index.
Methodology & Limitations
This report combines two lanes of data. The internal lane comes from the EduPoint platform and is presented as indices, rankings, or shares, with no absolute figures. The external lane comes from official sources and is cited directly. This separation preserves confidentiality alongside credibility.
Data window
The internal learning-mode composition covers February to June 2026. Because this window is short, the report is framed as a snapshot of the current state, not a year-over-year trend analysis. The post-normalization pattern rests on the national learning-transition narrative from official sources.
Mode definitions
In-person covers sessions held with the physical presence of teacher and student. Online covers remote sessions through digital tools. Blended covers learning arcs that mix the two. Adoption is measured from the composition of incoming demand.
Masking & confidentiality
All internal magnitudes are converted into indices, rankings, or relative shares. No session counts, rupiah rates, or business metrics are shown. The aim is to present the structure and direction of adoption without exposing operational data.
Illustrative index
The post-normalization pattern curve is illustrative, depicting the shape of the surge and settling of online adoption anchored to the transition from remote learning to in-person schooling. It is not an absolute time series from internal data.
External alignment
The reading of adoption across regions is aligned with official internet-access distributions (APJII and BPS) so it reflects the structure of national digital readiness and guards against internal sample distortion.
Limitations
As a snapshot of one period, this report does not yet capture full seasonality or year-to-year change. Data cells with small samples are suppressed to protect representativeness. The next edition will extend the time window so long-term direction can be read more reliably.
References
- [1]Indonesia Internet Penetration Survey 2024, Indonesian Internet Service Providers Association (APJII) (2024)
- [2]Telecommunication Statistics in Indonesia 2024, Statistics Indonesia (BPS) (2024)
- [3]Proportion of Individuals Using the Internet by Province, Statistics Indonesia (BPS) (2024)
- [4]Percentage of Households that Accessed the Internet by Province and Area Classification, Statistics Indonesia (BPS) (2024)
- [5]Joint Decree of Four Ministers on Learning During the COVID-19 Pandemic, Ministry of Education, Culture, Research, and Technology (2022)
- [6]Minister of Education Regulation No. 12 of 2024 on the Curriculum for Early Childhood, Primary, and Secondary Education, Ministry of Education, Culture, Research, and Technology (2024)
- [7]Law No. 20 of 2003 on the National Education System, Government of the Republic of Indonesia (2003)
- [8]PISA 2022 Results, Country Note: Indonesia, OECD (2023)
Appendices A
Indexed Data Tables
A summary of all indices used in this report. Every value is relative (indexed) and does not represent an absolute total.
| Mode | Share |
|---|---|
| In-person | ≈71% |
| Online | ≈23% |
| Blended (hybrid) | ≈6% |
| Region type | Index |
|---|---|
| Greater Jakarta metro | 133 |
| Large Java cities | 110 |
| Bali & metros beyond Java | 98 |
| National average | 100 |
| Regional cities & districts | 76 |
| Level | Index |
|---|---|
| University students & professionals | 158 |
| Senior high & test prep | 121 |
| Junior high | 96 |
| Primary school | 67 |
| Early years (PAUD/TK) | 39 |
| Cluster | Index |
|---|---|
| Programming & AI | 100 |
| Foreign languages | 82 |
| Test preparation | 61 |
| Math & Science | 54 |
| Quran recitation | 29 |
| Music (piano) | 22 |
| Early literacy (PAUD) | 14 |
Appendices B
Glossary
- Adoption index: a value relative to a base (e.g. national average = 100) that describes the level of use of a learning mode, not an absolute total.
- Learning mode: the form in which tutoring is delivered, covering in-person, online, and blended (hybrid).
- Post-normalization: the phase after the pandemic-era surge in remote learning subsided and learning modes sought a new balance.
- Digital divide: the difference in internet access and use across regions, especially between urban and rural areas.
- Share: the relative proportion of a mode against total sessions, expressed as a percentage.
Cite this report
Adhel Pradita Sugito, & Edupoint Researcher Team. (2026). Learning Mode Adoption Index 2026: Online vs In-Person. EduPoint Education Research. EduPoint Indonesia. https://edupoint.id/en/research/report/indeks-adopsi-mode-belajar-2026
@techreport{edupoint-indeks-adopsi-mode-belajar-2026-2026,
title = {Learning Mode Adoption Index 2026: Online vs In-Person},
author = {Adhel Pradita Sugito and Edupoint Researcher Team},
institution = {EduPoint Indonesia},
type = {EduPoint Education Research},
year = {2026},
url = {https://edupoint.id/en/research/report/indeks-adopsi-mode-belajar-2026},
}Learn in the mode that fits your child best
EduPoint connects families with quality private tutors across Indonesia, in person or online, for academics, test preparation, and new skills.
Find a Private Tutor