Online Learning Statistics 2026: 100 Growth Data Points

In fall 2024, 26.5% of all US postsecondary students took every one of their courses at a distance (NCES, Digest of Education Statistics 2025). The most repeated of the online learning statistics, a 15% MOOC completion rate, comes from a dataset last updated in 2015.

Key Takeaways

Online is no longer a segment of US higher education, it is most of it. 54.7% of US postsecondary students took at least one course at a distance in fall 2024, and 26.5% took every single one that way.
Completion ended lower than it started across every year anyone measured it, and the famous 15% is a 2015 number. The most rigorous long-run study puts edX certification among all registered participants at 3.13% by 2017-18. Both figures are correct. They count different people.
Paying changes behavior more than any course design tweak. In one edX dataset, certification runs from 3% among everyone who registers to 56% among learners who paid for a verified track.
The clearest case of AI disruption in this sector is Chegg. Its academic services revenue fell 61% year over year in the quarter to June 2026, while its non-academic business held flat.
Student AI use has already outrun institutional policy. 84% of US high schoolers used generative AI for schoolwork by May 2025, while only 28% of colleges had a formal AI policy in place that spring.
There is no trustworthy “e-learning market size” number. Every figure in circulation comes from a paid analyst firm that does not publish its method, and the full reports sit behind paywalls.

Top Online Learning Statistics for 2026

If you only carry four numbers into a meeting, carry these.

Online learning is now enormous and finishing is still rare: 26.5% of US college students take every course at a distance and 205 million people have registered on Coursera, while the average MOOC completion rate is 15% and 64% of US teens use an AI chatbot.

Six Core Online Learning Statistics

1. 26.5% of all US postsecondary students, 5,214,325 people, took every one of their courses at a distance in fall 2024 (NCES, Digest of Education Statistics 2025, Table 311.15).

2. 54.7% of US postsecondary students, 10,761,754 people, took at least one distance course in fall 2024 (NCES, Table 311.15).

3. The average completion rate across MOOCs is approximately 15%, and can approach 40% in the best cases (Katy Jordan, MOOC Completion Rates: The Data, data last updated 12 June 2015).

4. 205 million people had registered on Coursera as of 31 March 2026, after a first-quarter record of 7.6 million new registrations (Coursera Q1 2026 results).

5. 64% of US teens aged 13 to 17 use an AI chatbot, and roughly three in ten use one every day (Pew Research Center, December 2025, n=1,458, fielded 25 September to 9 October 2025).

6. 273 million children and youth were out of school worldwide in 2024 (UNESCO).

Origins of the 15% Benchmark

7. The widely cited 15% completion rate traces to a 2018 Columbia University Teachers College report, where it appears in one sentence: an anonymous survey at a September 2018 webinar in which ten respondents reported Specialization completion of 15% or less at their own institutions, eight of them 5% or less (Hollands and Kazi, Benefits and Costs of MOOC-Based Alternative Credentials, CBCSE, Teachers College, Columbia University, October 2018).

8. Ten people answering a webinar poll is a weak foundation for an industry benchmark. The same 15% figure is independently supported by Jordan’s aggregated dataset of published MOOC enrollment and completion figures, which is the citation I would use going forward (Jordan, 2015).

9. The NCES Fast Facts page on distance learning, which currently ranks first in Google for this topic, still reports fall 2021 figures (NCES Fast Facts: Distance learning). The underlying Digest table has been updated three enrollment cycles further. Cite the table, not the summary page.

Online Course Completion Rate Statistics

This is the section other sites link to, so it gets the most detail. The short version: in the datasets that measure it over time, completion has not improved, and every number you see is a different population wearing the same label.

Course completion rates ranked by which learners are counted: 56% among paid verified-track edX learners, 48% among learners who intended to finish, 43% among learners who started, 30% of everyone who enrolled, 15% average MOOC completion, and 3% of all registered participants.

Online Course Completion Rates Measured

10. Certification among all registered edX and MITx participants fell to 3.13% in 2017-18, down from 4.96% in 2013-14 (Reich and Ruiperez-Valiente, “The MOOC Pivot,” Science, 2019).

11. That series peaked at 5.91% in 2014-15 and declined in every subsequent year the study covers (Reich and Ruiperez-Valiente, 2019).

MOOC certification fell every year after its 2014-15 peak, from 5.91% to 3.13% of all registered edX participants by 2017-18.

12. Among participants who stated an intention to complete, the same study reports certification between 11.5% and 17.7% depending on the year (Reich and Ruiperez-Valiente, 2019).

13. Among paid verified-track learners, the same study reports certification between 44% and 56% (Reich and Ruiperez-Valiente, 2019).

14. The most recent peer-reviewed completion study available reports a raw completion rate of 30.02% across 15,805 enrollees in four MOOCs (Celik and Cagiltay, Open Praxis, August 2024).

15. In the same study, completion rises to 43.08% when only learners who started the course are counted (Celik and Cagiltay, 2024).

16. It rises again to 48.13% among learners who said they intended to complete (Celik and Cagiltay, 2024).

17. Per-course raw completion in that study ranged from 12.3% to 44.05%, across only four courses (Celik and Cagiltay, 2024).

Where the 96% Dropout Figure Comes From

18. A roughly 96% dropout rate is the most-repeated framing of MOOC completion, and it is a fair paraphrase of one specific measure: the 3% to 6% certification rate among everyone who registered on edX between 2013-14 and 2017-18 (Reich and Ruiperez-Valiente, 2019).

19. It is not a fair description of learners who paid, where certification reached 56% in the strongest year of the same dataset (Reich and Ruiperez-Valiente, 2019).

20. Registration on a free platform is closer to bookmarking than to enrolling. The 2019 study counts every account that ever registered for a course, and reports that 52% of registrants never enter the courseware at all (Reich and Ruiperez-Valiente, 2019).

21. For comparison, formal degree completion is also far from universal. Only 43% of new entrants to bachelor’s programmes across 32 OECD and partner countries graduate within the expected duration (OECD, Education at a Glance 2025, page 5).

22. That rises to 59% after one additional year and 70% after three additional years (OECD, 2025).

23. Men complete a bachelor’s within three years beyond the expected end date at 63%, against 75% for women (OECD, 2025).

What Increases Online Course Completion

24. Paying for a certificate raises engagement measurably. A study of more than 70 Coursera courses found a certificate effect worth 8% to 9% additional engagement and a transient sunk-cost effect worth 17% to 20% (Goli, Chintagunta and Sriram, Journal of Marketing Research).

25. Intent matters as much as price. Only 35% of participants in the Columbia study planned to earn the alternative credential at all, and a quarter planned to take every course in a series without earning the credential (Hollands and Kazi, 2018).

26. The average cost of a MicroMasters at the time of that study was $960, and one Boston University MicroMasters cost $900 for five courses of six to eight weeks (Hollands and Kazi, 2018).

The gap between 3% and 56% inside a single dataset is a pricing question before it is a teaching question. Everyone who registers on a free platform is counted, whether or not they ever opened a lesson. Paying is the cheapest available proxy for intent.

If you are building a course, the practical read is simple. Charge for it, give it a deadline, and stop comparing your numbers to free MOOC benchmarks. I go into which platforms make that easiest in the best online course platforms roundup.

Online Education Statistics: How Many People Learn Online

US Distance Learning Statistics and Enrollment

In fall 2024, 26.5% of US postsecondary students took every course at a distance, 28.2% took some, and 45.3% took none.

US distance learning statistics all trace back to one federal table, and it is worth reading directly rather than through a summary.

27. 19,659,633 students were enrolled in US degree-granting postsecondary institutions in fall 2024 (NCES, Table 311.15).

28. 5,547,429 of them, 28.2%, took some but not all of their courses at a distance (NCES, Table 311.15).

29. 8,897,879, or 45.3%, took no distance course at all (NCES, Table 311.15).

30. In fall 2023, the most recent year with final rather than provisional data, 4,982,327 students, 26.2%, studied exclusively at a distance (NCES, Table 311.15).

31. Total US postsecondary enrollment grew from 19,038,434 in fall 2023 to 19,659,633 in fall 2024, and the fully-online headcount grew faster than the total, at 4.7% against 3.3% (NCES, Table 311.15).

Undergraduate Versus Graduate Online Enrollment

32. 16,385,717 US undergraduates were enrolled in fall 2024 (NCES, Table 311.15).

33. 3,888,681 undergraduates, 23.7%, took every course at a distance (NCES, Table 311.15).

34. 8,951,854 undergraduates, 54.6%, took at least one distance course (NCES, Table 311.15).

35. 3,273,916 postbaccalaureate students were enrolled in fall 2024 (NCES, Table 311.15).

36. 1,325,644 of them, 40.5%, took every course at a distance, roughly 70% higher than the undergraduate rate of 23.7% (NCES, Table 311.15).

37. Only 484,256 graduate students, 14.8%, mixed distance and in-person courses (NCES, Table 311.15). Graduate students tend to be all in or all out; undergraduates blend.

That split matters if you sell education. The fully-online graduate student is far more often a working adult buying a credential on a schedule, which is a different buyer from an 18-year-old taking one asynchronous elective.

Global Enrollment and Access

38. 269 million students were enrolled in higher education worldwide, roughly 43% of the 18 to 24 age group, more than double the figure of 20 years earlier (UNESCO).

39. 79 million children of primary school age were out of school in 2024 (UNESCO).

40. 64 million of lower-secondary age and 130 million of upper-secondary age were out of school in the same year (UNESCO).

41. Global out-of-school numbers fell 33% between 2000 and 2015, then rose 3% by 2024 (UNESCO).

42. In low-income countries the out-of-school population has increased 29% since 2015, while staying broadly flat elsewhere (UNESCO).

43. Average adult literacy proficiency among 25 to 64 year-olds across OECD countries with data for both survey cycles fell by 9 score points since 2012, to 260 points (OECD, Education at a Glance 2025).

44. 13% of tertiary-educated adults across 29 OECD countries failed to reach even the baseline literacy proficiency level in 2023 (OECD, 2025).

Online Learning Market Size Statistics

What Market Size Figures Measure

45. The global education and training market is projected to reach almost $10 trillion by 2030, growing at 4.4% a year (HolonIQ, 2025 Global Education Outlook). That figure covers all government, employer and household spending on education, not digital learning.

46. No government statistical agency publishes an e-learning market size. Every figure in circulation comes from a paid analyst firm whose full report is gated and whose method is not published.

47. HolonIQ’s projection is the most defensible number I found, and it still needs its caveat printed next to it. Its full report runs to 180 pages behind a paywall, and the headline figure counts all education and training spending, including public school systems (HolonIQ).

EdTech Venture Funding Decline

48. Global EdTech venture funding was $1.8 billion in 2024, the lowest level since 2014 (HolonIQ).

49. EdTech has taken roughly 2% of all global venture funding since 2010 (HolonIQ).

50. HolonIQ recorded 300 M&A deals and 4 IPOs in education in 2024 (HolonIQ).

Market-size decks keep going up while the money going into new companies keeps going down. Both can be true, and the second one is a better guide to what the next two years look like.

Online Learning Platform Statistics and Users

Coursera and Udemy Are Now One Company

51. Coursera completed its combination with Udemy on 11 May 2026 (Coursera). Any article still describing them as competitors is out of date.

52. The combined company describes itself as connecting learning signals from more than 290 million learners and 18,000 enterprise customers (Coursera).

53. Each Udemy share converted to 0.800 Coursera shares, leaving former Coursera holders with about 59% of the combined company and former Udemy holders with about 41% (Coursera).

54. The companies expect $115 million of annual run-rate cost savings within 24 months of closing (Coursera).

55. Coursera’s revenue for the quarter to March 2026 was $195.7 million, up 9% year over year (Coursera Q1 2026 results).

56. Coursera reaffirmed a full-year 2026 revenue outlook of $805 million to $815 million (Coursera).

57. Udemy’s final year as an independent public company produced $789.8 million of revenue, an increase of just 0.4% on the prior year (Udemy FY2025 Form 10-K).

58. Enterprise revenue was $524.1 million, 66% of the total, while consumer marketplace revenue fell 9% to $265.8 million (Udemy FY2025 Form 10-K).

59. Udemy averaged 40 million monthly unique visitors during 2025 and has recorded more than 1 billion cumulative course enrollments since inception (Udemy FY2025 Form 10-K).

60. Udemy Business net dollar retention was 93% at the end of 2025, and 97% among its large customers (Udemy FY2025 Form 10-K).

A note on the biggest number in this section. Registered learners is a count of accounts ever created. It never goes down, nobody ever gets removed from it, and it is not a measure of anybody using anything. Coursera does not publish monthly or daily active users.

Duolingo User and Revenue Growth

61. Duolingo averaged 133.1 million monthly active users in the fourth quarter of 2025, up 14% year over year (Duolingo Q4 and FY2025 shareholder letter).

62. Daily active users reached 52.7 million, up 30% (Duolingo).

63. Paid subscribers reached 12.2 million at period end, up 28% (Duolingo).

64. Full-year 2025 revenue was $1,037.6 million, up 39%, with total bookings of $1,158.4 million (Duolingo).

65. Duolingo reported $414.1 million of net income for 2025, a figure that includes a one-time $256.7 million tax benefit from releasing a deferred-tax valuation allowance (Duolingo).

66. About 9% of Duolingo’s monthly actives pay, which is the conversion rate to beat in consumer learning (Duolingo, derived by dividing 12.2 million period-end paid subscribers by 133.1 million averaged quarterly monthly active users, so treat it as an approximation).

Duolingo is the only large online learning company reporting daily active users at all, which tells you how the rest of the sector prefers to be measured.

Khan Academy and the Nonprofit Numbers

67. Khan Academy had 189.6 million registered users at the end of the 2024-25 school year, after adding 20.9 million (Khan Academy Annual Report SY24-25).

68. It counted 104.9 million yearly active learners, people who used the platform for any amount of time during the school year (Khan Academy).

69. Learners spent 66.8 billion minutes on the platform, up 8.1 billion year over year (Khan Academy).

70. 1.7 million learners became proficient in 60 or more skills during the year (Khan Academy).

71. 2.0 million people had access to Khanmigo, its AI tutor, at some point during the school year (Khan Academy).

72. Khan Academy is available in more than 190 countries and 55 languages (Khan Academy).

Note the gap between 189.6 million registered and 104.9 million active. Khan Academy publishes both, which is more than most platforms do.

Course Platforms Built for Creators

73. Thinkific processed $460.0 million of gross merchandise volume across its creators in 2025, flat year over year (Thinkific Q4 and full-year 2025 results).

74. Thinkific’s own revenue was $73.2 million, up 9% (Thinkific).

75. Payment volume through Thinkific Commerce grew 36% to $273.5 million (Thinkific).

76. Thinkific reports more than 35,000 customers (Thinkific).

Thinkific is the only creator course platform of any scale that files public financials, which is why it is the only one with a number in this section. Kajabi, Teachable, Podia and Skool are private and publish marketing milestones rather than audited figures. If you are comparing them, the online learning platforms breakdown covers what each is actually good at, and there are full reviews of Kajabi and Thinkific.

AI in Online Learning Statistics

How Students Use AI Tools

AI use falls as students get older: 84% of US high schoolers used generative AI for schoolwork and 69% used ChatGPT, against 42% of college students and 30% of college instructors using it weekly or daily.

77. 84% of US high school students reported using generative AI tools for schoolwork by May 2025, up from 79% in January of the same year (College Board, U.S. High School Students’ Use of Generative Artificial Intelligence).

78. 69% of US high school students reported using ChatGPT specifically to help with school assignments in May 2025 (College Board).

79. 42% of US college students used generative AI weekly or daily in spring 2025, up from 14% in spring 2023 (Tyton Partners, Time for Class 2025).

80. 30% of instructors and 40% of administrators reported the same weekly-or-daily cadence (Tyton Partners).

81. ChatGPT is the chatbot US teens name most often at 59%, followed by Gemini at 23% and Meta AI at 20% (Pew Research Center).

82. 16% of US teens use an AI chatbot several times a day or almost constantly (Pew Research Center).

83. Computer science students accounted for 38.6% of student conversations in an analysis of roughly one million anonymized Claude conversations, despite being 5.4% of US bachelor’s degrees (Anthropic Education Report: How University Students Use Claude). The same report gives 36.8% for that share in its summary, so quote the figure with the sentence it came from.

84. That study filtered about one million conversations down to 574,740 judged relevant to student academic work (Anthropic).

The Anthropic figures come from the company whose product they measure, over an 18-day window, on one platform. Anthropic says so in the report. Treat them as a description of Claude users, not of students.

Institutional AI Policy Adoption

85. Only 28% of US institutions had a formal institution-wide generative AI policy developed and rolled out as of spring 2025 (Tyton Partners).

86. Curriculum development accounted for 57% of educator conversations in an analysis of about 74,000 anonymized Claude conversations from higher-education accounts (Anthropic Education Report: How Educators Use Claude).

87. Academic research accounted for 13% and assessing student performance for 7% of those conversations (Anthropic).

Students moved first, instructors moved second, and policy is still catching up. The gap between 84% student adoption and 28% institutional policy is the single clearest number in this whole article.

AI Impact on Chegg Revenue

Chegg's academic services revenue fell from $87.9 million to $34.3 million between the June 2025 and June 2026 quarters, while its skilling business stayed near $17 million.

88. Chegg’s total net revenue for the quarter ended 30 June 2026 was $51.8 million, against $105.1 million a year earlier, a fall of 51% (Chegg Form 10-Q).

89. Its Academic Services revenue fell 61%, from $87.9 million to $34.3 million, over the same period (Chegg Form 10-Q).

90. Its Chegg Skilling segment held roughly flat, at $17.5 million against $17.2 million (Chegg Form 10-Q).

91. Over the first six months of 2026, total revenue fell 49% to $115.1 million (Chegg Form 10-Q).

92. Chegg names the cause directly in its filing, writing that students “see generative AI products like ChatGPT and others as strong alternatives to vertically specialized solutions for education” and that this “has negatively impacted our industry” (Chegg Form 10-Q).

93. Chegg filed an antitrust complaint against Google on 24 February 2025 over the expansion of AI Overviews in search (Chegg Form 10-Q).

This is the clearest case study in the sector. A business whose value was answering student assignment questions lost most of its revenue in roughly two years, and the part of the business that teaches skills rather than supplies answers did not move.

Corporate and Workplace Learning Statistics

Skills Change and Reskilling Demand

94. Employers expect 39% of workers’ core skills to change by 2030, down from 44% in the 2023 edition of the same survey (World Economic Forum, Future of Jobs Report 2025).

95. 170 million new jobs are projected to be created and 92 million displaced by 2030, a net increase of 78 million (WEF).

96. That amounts to structural churn of 22% of the 1.2 billion formal jobs in the study’s dataset (WEF).

97. 63% of employers name skills gaps as the biggest barrier to business transformation over 2025 to 2030 (WEF).

98. 85% of employers plan to prioritize upskilling their workforce (WEF).

99. 70% expect to hire staff with new skills, 51% plan to move staff from declining to growing roles internally, and 41% foresee staff reductions as skills become obsolete (WEF).

The Future of Jobs survey covers more than 1,000 employers across 22 industry clusters and 55 economies. These are employer projections about the future, not measured outcomes, and the respondent base shifts between editions.

Corporate Learning and Development Benchmarks

100. Only 36% of organizations qualify as career development champions on LinkedIn’s own index, with 31% showing limited adoption (LinkedIn Workplace Learning Report 2025).

101. 49% of learning and development professionals say their executives are concerned that employees lack the skills to execute business strategy (LinkedIn).

102. 51% of those career development champions describe their organization as leading or accelerating on generative AI adoption, against 36% of organizations with weaker career development programs (LinkedIn).

LinkedIn sells learning software, career development champion is a category LinkedIn invented, and the respondents are LinkedIn’s own users. The direction is probably right; the precision is not.

Course Pricing and Creator Economics Statistics

Average Online Course Price

103. Across 132,009 course sales on one creator platform, the average course price is $137 (Podia).

104. When this page was first published, it cited an earlier version of that same Podia analysis, reporting an average of $182.59 and a median of $76.50 across roughly 133,000 courses. Podia has since revised the dataset. The current published average is $137 (Podia). Both figures are Podia’s; only one is current.

105. 63% of US households participated in at least one creative activity in the previous 12 months, based on a survey of more than 6,000 US adults (Association For Creative Industries, 2016 Creative Products Size of the Industry Study).

Creator Course Earnings Data Gaps

106. Kajabi, Teachable, Podia and Skool publish no audited figures. Every creator earnings number attributed to them traces to a press release or a third-party estimate.

107. The widely quoted “average online course costs” figure has no single traceable origin. Podia’s is the only first-party dataset published in the open (Podia).

If you are pricing a course, the only published benchmark covers one platform and sits near $137. That is not a reason to price there. It is a reason to notice how much of that platform’s volume sits at commodity prices, which is the market position a first-party audience makes unnecessary. More on building that audience in the creator economy statistics page.

For the tools side, see the best online course platforms roundup and the online learning platforms comparison. Related reading: content marketing statistics.

Frequently Asked Questions

What Is the Dropout Rate for Online Learning?

Between 94% and 97%, if you count everyone who ever registered for a free MOOC. Certification among all registered edX and MITx participants ran from 4.96% in 2013-14 down to 3.13% in 2017-18.

That is honest but nearly useless for planning, because registering on a free platform costs nothing. Among learners who paid for a verified track in the same courses, certification ran between 44% and 56%. Among learners in a 2024 study who started the course, completion was 43.08%. Pick the population that matches yours first.

What Percentage of Students Complete an Online Course?

The most defensible general figure is still approximately 15%, from Katy Jordan’s aggregated dataset of published MOOC enrollment and completion data, last updated in June 2015. Nothing comparable has replaced it at that scale.

For a paid course with deadlines, expect much higher. For a free self-paced course, expect much lower.

How Many People Take Online Courses?

In US higher education, 10,761,754 students took at least one course at a distance in fall 2024, and 5,214,325 took every course that way. Globally there is no equivalent count, because no statistical agency collects one. Platform numbers are the closest proxy and they overstate: Coursera’s 205 million is a cumulative count of registered learners, not people currently learning.

How Big Is the Online Learning Market?

Nobody credible knows. The most defensible number is HolonIQ’s projection of almost $10 trillion for the total global education and training market by 2030, growing at 4.4% a year, and that figure includes public school systems, university tuition and corporate training budgets, not just digital learning.

The standalone “e-learning market” figures that circulate come from paid analyst firms that do not publish their methodology. They disagree with each other by wide margins.

How Much Does an Online Course Cost?

Podia’s analysis of 132,009 course sales puts the average price at $137. That is the only first-party dataset on course pricing that could be traced to its publisher and read.

Marketplace courses sell for far less, and cohort-based courses for far more, so the average describes almost nobody. Price against your own audience and the outcome you can prove, not against a market average.

Is Online Learning as Effective as Classroom Learning?

No primary source answers this cleanly, and any page giving you a single percentage is quoting a meta-analysis from a different technological era.

The question is also becoming academic. In fall 2024, 54.7% of US postsecondary students took at least one course at a distance and only 45.3% took none. Traditional degrees are not a high bar either: only 43% of bachelor’s entrants across OECD countries graduate on time, which is not obviously better than the paid online cohorts in the edX data.

Is AI Making Online Learning Better or Worse?

The adoption data is clear and the outcome data is not. 84% of US high schoolers used generative AI for schoolwork by May 2025 and 42% of college students used it weekly or daily by that spring, while only 28% of institutions had a formal policy.

What no primary source yet shows is whether any of that improves learning. Vendor-published usage reports measure message volume, not understanding. Treat any claim that AI raises completion or comprehension as unproven until somebody runs an independent trial and publishes it.

Where Can I Find Official Online Learning Statistics?

For US enrollment, the NCES Digest of Education Statistics tables, which are free and updated annually. Use the Digest tables rather than the Fast Facts summary pages, which lag by several years.

For international comparisons, OECD Education at a Glance and the UNESCO Institute for Statistics data portal, both free. For platform scale, SEC filings, which carry the only audited financial statements anywhere in this field.

Sources

Figures on this page come from the primary sources below, checked in September 2026.