I’ve been doing some analysis recently of the costs of a fully online master’s degree program from a major research university (which for the moment will remain nameless, although this has been a very successful program, both academically and financially). I was doing this for our forthcoming book.

Unlike face-to-face teaching, where, after the initial planning phase, costs remain pretty much the same from year to year, the costs of online programs vary considerably over time. In the case I’ve been studying, if we take a point seven years after the program was approved, and five years after the first full cohort of students were enrolled, the costs break down something like this:


Cost (7 years) Percentage




Program administration and overheads















We can see that course development costs at 13% are quite a small proportion of the overall costs, while delivery costs constitute just over a third of all costs. Development costs occur early in the program. Annual maintenance costs are quite small, at less than 10 per cent of the total.

Figure 1 below shows that delivery costs constitute an even greater proportion after seven years.

Graph of costs of an online masters program over seven years

The other point to note are overheads and administrative costs, which with planning total 42% of all costs over seven years, and don’t decrease in subsequent years.

Several points occur to me from this analysis.

  1. Open content is not going to lead to major cost savings in online learning. Even without creating new content, someone will have to select, assess and modify open content, or provide some kind of curriculum framework or guide for students studying a subject or topic.
  2. What universities and colleges are really supplying with online learning is not content but service. The delivery costs in this program are mainly interaction with students through online discussions and direct e-mail communication, and student assessment. This is the quality part of online teaching, and a major cost (36%). If this ‘service’ is cut back, quality suffers, drop-out increases, and the credibility of the program suffers. (Think about telecommunications. Costs of transmission are now very low, at least in North America. What will increasingly differentiate between telecommunications carriers will not be price but service. Pity the big telephone companies have not understood this yet – ever tried getting past their voice mail? Service is seen as a ‘cost’ that should be avoided, not a competitive advantage.)
  3. Increasingly, institutions or faculty do not control content. Content is ‘out there’. What matters is how faculty and students select and use that content, and this requires relatively high quality interaction between students, and between students and instructor.
  4. If we want to bring the costs of online teaching down, without sacrificing quality, we need to focus on administration and overheads. These indirect costs still exceed 40 per cent of all costs (and in this case did not even include technology costs, such as servers, networking costs and LMS licensing fees.)
  5. We now know the costs of online teaching (just under $3 million after seven years, with an annual intake of 67 students). Where though are the analyses of the costs of different forms of face-to-face teaching? (Lectures, lab classes, etc.).
  6. This analysis takes an activity-based approach to accounting (a program as an activity), as distinct from normal university and college accounting, which is departmental (how much did we spend on salaries). A bit more useful, isn’t it?
  7. The ‘true’ cost per student for this program is $12,500 per student by year 7. This was a fully cost-recoverable program, so students paid full tuition. In this program, the ‘breakeven’ point (when revenues exceeded expenditure) came after three years. But it’s not until after year 7 that it starts to make a profit (because in the first few years it ran a deficit that had to be covered – see graph below), which was ploughed back into new programming. However, the same model could be used for grant- or base-funded programs. Base funding would cover so much per FTE, with the balance being made up from tuition fees.
  8. In a full cost recovery program, fees are a function of costs divided by course enrolments. Know any two of these and the third can be calculated. Thus if fees are determined by market competition, and enrolments can be manipulated by marketing, then the costs are determined and you have to design the program within those fixed costs, or you’ll go bust.

Revenues and expenditures over seven years

I’d really be interested if anyone else is willing to trade ‘program cost’ stories, either for online or face-to-face programs. In particular, does anyone have an activity-based cost model for a hybrid program?


  1. sadly poignant – an institution that has an economic model that focuses on delivery and quality service will be a survivor – “the university” has already passed from the living to the “living dead” and as a glacier will ssssllllooowwwwllllyyyy fade away rarely recognizing the general useless of its present modus operandi. Today’s educational institutions needs to morph as have many companies into a sleeker, more focused toward the consumer (as in “student”and more nimble to make the necessary changes to continue into the future. Problematic issues are “old school” faculty, even “older school” administrators who like many of today’s “old school” media firms can’t see past their “mission statements” to see the needs of students – GREAT post – thanks thanks and thanks!!

  2. […] Tony Bates has a new post relating the costs of an online Master’s program. From the post: I’ve been doing some analysis recently of the costs of a fully online master’s degree program from a major research university (which for the moment will remain nameless, although this has been a very successful program, both academically and financially). I was doing this for our forthcoming book. […]

  3. Dear Tony
    1.- I agree with your numbers except delivery expenses.
    2.- You do not divide these total costs by number of students.
    3.- Share this program with several universities in the USA and the world, collect several 1.000 students then cost per students is much less then f2f.
    4.- I have studies for the cost of online for 15 years. I will share those with you.
    Thanks billion

  4. Thanks for the comment, Muvaffak

    In response, you may not agree with the delivery costs, but that’s what they are. Yes, I think we can reduce delivery costs further, and maintain quality, but it’s not easy. They could be reduced, by increasing class size and cutting down interaction between instructor and students, but this is a graduate program, and interaction and discussion were considered essential to maintain quality, in terms of developing research, analysis and critical thinking skills. If you have found a way of automating this, please let us all know!

    Second, the cost per student gradually decreases each year the program is offered. Since the projected breakeven point was year 7, the cost per student was $12,500 in year 7 – which was the fee charged. The cost per student will be higher in years 1-6, and lower in years 8 onwards, if enrolment targets are met. (In fact the breakeven point came in year 5, because enrolments were higher than expected, so after nine years the cost per student has dropped to just over $10,000).

    The program was in fact developed in partnership with a university in Mexico. The business model presented in my posting does not take account of the costs and savings associated with the partnership. However, the Mexican university benefited in terms of costs and ‘perceived’ quality of content, because the Canadian university covered most of the development costs.

    Third, 30 per cent of enrolments in the Canadian offering of the program are international. Furthermore there have been several thousand Mexican enrolments in the version offered by the Mexican university.

    However, it is a popular misconception that there are millions who are qualified, who have the time and resources, and want to take an online program such as this, and that there will be enough qualified instructors to teach the program if there were. Furthermore, programs such as this need to be adapted to local cultures and circumstances (which was the real benefit of the Mexican partnership and probably why it hasn’t been taken up in the USA).

    The real question is: do people in developing countries want lower quality programs than in developed countries, but at less cost? And wouldn’t it be better for quality programs to be developed in their own countries, rather than import them from abroad, if at all possible?

    Finally, yes, please share your own cost analysis, but also we should see the benefits and limitations side as well. You can always cut costs by reducing quality, in any enterprise.

    Thanks very much for this comment, Muvaffak. I may disagree with you, but you have raised excellent questions.

  5. Absolutely fantastic article, with conclusions that I completely agree with, many of which go beyond just costs.

    I run an online institution, and my gut feel pretty much resonates with all the points you have made.

  6. I wonder what college instructors use for content other than “open source” material? If they don’t use “open source” content, then are they using their own content or are they paying a content provider year after year? In the K-12 online learning world, open source or open educational resources provide variety for teachers to pull from without a an ongoing content or subscription cost year after year. Granted, the way college and K-12 schools are funded are different, but it seems like whatever any of us do to not pay yearly subscription fees would reduce the cost of online program over time.

    Secondly, I wonder where professional development costs fall in your analysis…perhaps that is part of the “delivery” cost? Because we know it takes many years to become an effective online instructor at any level.

  7. Hi, Rob

    Thanks for two great questions.

    What do college instructors use for content? It varies. Some research professors will draw on their own or their students research, but the majority will draw mainly from text books and journal articles, ‘remixing’ and remashing in the form of lecture notes, Powerpoint slides, etc. Some even build their courses around content provided by learning management systems companies such as Blackboard (it’s the students who pay though not the instructor). The point here is that most of instructors’ work is not creating ‘original’ materials, but organizing already existing content. This won’t change if the materials are ‘free’ open content.

    You are right to point out that the business plan doesn’t include professional development costs. However, it does include the costs of instructional designers and web designers. It could be argued that it would be cheaper to ensure that instructors are properly trained in the first place, then instructional designers wouldn’t be necessary, but in this particular program we were not in that position (even though we were working with instructors from the Faculty of Education). In any case, I suspect (but have no data) that it would still be cheaper to use instructional and web designers, because they take a lot of the workload off more highly paid faculty – it doesn’t make sense to train a $150,000 a year top professor to do html mark-up or edit video.

    Anyone else who would like to add to this discussion of where instructors get content and the cost advantages of professional development vs working in a team with other professionals? (Or both!).

  8. Thanks for an interesting analysis.
    Would you think that one can reduce the cost of delivery (and still retain the high quality) by making (more or better) use of peer tutoring or group activities in which the instructors need not be so active?
    In my own experience the main cost and quality of online delivery is indeed in the work of the instructors. We used 2 instructors for each group of 15 students in a short-term programme.


Please enter your comment!
Please enter your name here