OpenFiction [Blog]

A record month for MIT OpenCourseWare – 1.7 M visits

Posted in Evaluation, MIT OpenCourseWare, Open Educational Resources, OpenCourseWare, web metrics by scarsonmsm on November 1, 2011

October is traditionally OCW’s annual high-water mark for traffic, and last month was no disappointment in that regard.  The site received a record 1,733,198 visits from 1,026,004 unique visitors.  This eclipses the previous high of 1,602,561/1,015,112 from August last year, and is a 12.4%/12.8% increase over last October.

A few more October numbers:

  • Average visits per day: 55,909
  • Page views:  8.8 M
  • Top course: 6.00 Intro to Computer Science and Programming – 104,096 visits

It’s great to see continued momentum as we swing into a new school year.

MIT OpenCourseWare vs. Spiderman

For what it’s worth:

Metric MIT OpenCourseWare Spiderman: Turn off the dark
Years in production 8 1
Cost to date $40 M $65 M
Audience to date 100 M 20,000 (est.)
Ticket price Free $70-$250
Irish rock icons 0–so far 2 (Bono, The Edge)
Major production delays None 4
Serious injuries to principals 1 (see link below) 4
Fires 1 (2006) 0 that we know of
Critical reviews Largely positive Mixed
Web reach World-wide Stage-wide
Spandex-clad heros 2 (don’t ask) 1
Ratio of great power to great responsibility 1:10 10:1

100 Million Served

As we swing into MIT OpenCourseWare’s 10th anniversary year, I’ve been taking a look at the data from the past decade.  Really, this is the data from October 1, 2003 to present, as the program was announced April 4, 2001, and it took 2 1/2 years to get to the official launch.  Looking at the numbers, there is a good case to be made that OCW has reached as many as 100 million people in our first decade.  Here’s how:

  • Direct use: We’ve welcomed 50 million unique individuals directly to the site since 2003.  Web metrics always come with caveats, and I could interpret the stats 10% higher or 10% lower, but this is a good round number.
  • Translation site visits: Our translation affiliates maintain their own sites, and some are better than others about reporting traffic, but we’ve had 32 million visits reported from them.  Divide this by the 1.8 visits per visitor historic ratio we have and you get just under 18 million visits.  This doesn’t reflect significant numbers of unreported visits, so let’s round up to 20 million.
  • Redistribution sites and programs: Since we began redistributing our content through iTunes U and YouTube, we’ve received reporting indicating 17 million downloads from iTunes U and 15 million YouTube views.  The ratio of views/downloads per visitor is going to be higher than the visits/visitor above (plus there is overlap with some folks watching embedded YouTube videos on our site) so let’s say 2.5 views/downloads per visitor.  This would mean roughly 13 million more.  We are redistributed by a number of other sites (AcademicEarth, and we have sent mirror sites to 250 universities in bandwidth-constrained regions, so lets add another 2 million for this.  This brings our running total to 85 million.
  • Secondary uses: OCW materials are licensed for reuse, and we actively encourage visitors, especially educators, to download and reuse the materials.  We know that many educators do, and are thus bringing OCW materials to large numbers of students who never visit our site.  It’s a bit of a finger in the wind to figure out how many people this might represent, but let’s take a stab.  Educators are 10% of our visitors historically, so 5 million educators.  Forty-six percent of faculty responding to surveys indicate they have reused OCW materials.  This is almost certainly an overcount, as the survey respondents are self selected, but how many would have to have reused OCW content to have shared it with 15 million students? Assuming each teacher might share the materials with 50 students (a conservative number I hope), that would mean that 300,000 educators would have had to reuse OCW materials in the past ten years.  That’s 6% of the educators that our metrics indicate have been to the site.   This seems reasonable, if not conservative.

Fifty million direct, 20 million to translations, 15 million to redistribution sites, and 15 million through secondary use = 100 million.   All of these numbers are squishy and there’s some overlap between them. Maybe it’s in the 80 million range; maybe it’s higher than 100 million.   Even in rough numbers, it’s truly humbling to roll it all up this way, and a testament to the generosity and vision of the MIT faculty.

Year-end numbers for MIT OpenCourseWare

Posted in Evaluation, MIT OpenCourseWare, Open Educational Resources, OpenCourseWare, web metrics by scarsonmsm on January 3, 2011

Some high-level numbers from 2010:

  • 17.5 M visits
  • 9.6 M visitors
  • 1.82 visits per visitor
  • 98.3 M page views (actually a little lower than last year, but our site redesign is helping folks to find content faster)
  • 5.63 page views per visit
  • 1.9 M zip files downloaded
  • 11.8 M files downloaded from iTunes U
  • 7.3 M videos viewed on YouTube
  • 275 K visits from the MIT community (from GA, not WebTrends; WT is not reading this one right)
  • 446 K visits referred by StumbleUpon; 172 K by Reddit; 112 K by Wikipedia; 95 K by YouTube; 78 K by Facebook
  • 38% of visits used Firefox; 33% used IE; 15% used Chrome; 10% used Safari

Goodness all around.

OCW Zeitgeist: A year in search

Posted in Evaluation, MIT OpenCourseWare, Open Educational Resources, OpenCourseWare, web metrics by scarsonmsm on December 10, 2010

Apropos of the Google Zeitgeist, here’s the year in top search terms that brought people to MIT OpenCourseWare.*

  1. history of computer (24,424)
  2. mathematics (21,399)
  3. calculus (16,958)
  4. economics (12,166)
  5. differential equations (9,846
  6. linear algebra (9,516)
  7. architecture (7,975)
  8. political science (7,618)
  9. chemistry (7,382)
  10. math games (6,508)
  11. English speaking (5,790)
  12. box jellyfish (5,474)
  13. physics (5,181)
  14. biology (5,157)
  15. physical education (5,139)
  16. civil engineering (4,626)
  17. electrical engineering (3,906)
  18. anthropology (3,520)
  19. history of India (3,463)
  20. AP physics (3,342)
  21. communication skills (3,186)
  22. electricity and magnetism (3,155)
  23. math puzzles (2,982)
  24. Gilbert Strang (2,940)
  25. science and technology (2,898)
  26. wonders of the world (2,872)
  27. chemical engineering (2,843)
  28. AP calculus (2,791)
  29. multivariable calculus (2,747)
  30. physics lectures (2,624)
  31. science (2,520)
  32. Walter Lewin (2,416)
  33. history of mathematics (2,296)
  34. lighting (2,227)
  35. spiral model (2,102)
  36. high school physics (2,188)
  37. marketing management (2,082)
  38. nuclear engineering (1,947)
  39. nanotechnology (1,896)
  40. Gilbert Strang calculus (1,872)
  41. electrical projects (1,838)
  42. projects (1,818)
  43. computer science (1,791)
  44. biomedical engineering (1,790)
  45. random matrix theory (1,765)
  46. English dictionary (1,738)
  47. environmental engineering (1,734)
  48. organic chemistry (1,729)
  49. math worksheets (1,679)
  50. Walter Lewin lectures (1,676)
  51. Hindi poems (1,645)
  52. Strang calculus (1,610)
  53. materials science and engineering (1,576)
  54. math courses (1,576)
  55. street fighting mathematics (1,552)
  56. guitar (1,522)
  57. food web (1,518)
  58. electronics (1,515)
  59. calculator (1,515)
  60. human body (1,487)
  61. audio lectures (1,477)
  62. Monet (1,467)
  63. history of the internet (1,458)

MIT OpenCourseWare:  What the world is learning.

*Omitting terms with “MIT,” “OpenCourseWare,” “open course,” etc. out of the top 200

Education, 8 cents a page

In my undergraduate days, I used to go to the West Virginia University Library and (quite illegally) copy large sections of books in the collection, feeding dime after dime into the photocopier to pay the 10 cents a page. And that’s back when a dime was a dime.

I realized today in looking at some of the OCW statistics that our project has become more cost effective than that older method of redistributing educational materials. We’ve topped 540 million page views since launch on a total investment of 40 million dollars (both round numbers).

That comes out to under 8 cents a page (and that’s just our HTML content—even more cost effective if we counted the PDFs that hold most of the content).

Haunted by the ghosts

A couple of difficulties is talking about OER:

One: “Open educational resources” represents a broad and diverse field of practices that function in a range of ways, and to speak of them as a unified whole (unintentionally or intentionally) or expect them all to function in the same way is to confuse apples with oranges.

Two: Since OER is a relatively new field, and grew out of a number of precedents including open source software, open content and learning objects, one of the consistent issues I see arise is that people tend to projects their agendas from other realms onto the relatively blank slate of OER, rather than looking at what is actually going on.

I see both of these issues at work in David’s latest installment on OER adoption. Being unable to travel either to OpenEd or the Asia OCW Conference this year, I find myself with time to unpack how the issues above are at work in his response.

Disposing of the first issue above, throughout the post, David treats OER as a unified whole that must have a unified goal–in this case saving money. I do hope that OER will be able to generate some cost savings, and I see the best opportunities for this in open access journals and open text books. These would seem to offer savings opportunities, though, not because of their relationship to formal education, but because of their relationship to a media industry that is struggling to make its peace with the digital environment, and not doing a good job.

OpenCourseWare is different (at least in the higher ed realm) in that there is no significant media industry out there making money on the sales of course content alone. Outside of The Learning Company’s videos of lectures and a few other niche products, OCW hasn’t been trying to supplant some other for-profit producer that’s been raising prices on paying customers to offset losses due to digital copying as journal and textbook publishers have. That’s what has enabled OCW to spread as widely as it has and what makes it an effective wedge to begin discussions of open sharing on campuses. The success of OCW at MIT no doubt made the discussion of an MIT open access publishing policy easier to have. To lump OCW together with other OER and say the point of the thing is cost savings is a mistake.

On the second issue above, a common rhetorical device I often see in discussions of OCW, and one David uses here, is:

1) The value of OCW is (insert agenda here).
2) There is no evidence of (agenda).
3) Therefore OCW is unsuccessful.

David employs this in asking:

How many displacing adoptions are happening inside MIT thanks to the existence of MIT OCW? Since their OCW is the largest of them all, they could potentially be saving their students more money than anyone else. I’d love to see some data on this out of MIT OCW.

Is there evidence that MIT is saving MIT students money with OCW?  Not a lot. Does this lack of evidence mean that OCW doesn’t benefit MIT students? No, and we’ve actually documented considerable benefits for MIT students generated by the site. 93% of undergraduates and 82% of graduates use the site; 70% of MIT students use the site materials to compliment those they receive in class, 46% use it as a course planning tool, and 39% use it for personal learning; 58% of students rate the site’s impact on student experience as “extremely positive” or positive” and only 4% indicate no positive impact. Lots of benefit, just not the particular benefit of cost savings, which is David’s interest.

From a broader perspective, there’s no doubt that OCW and other OER are producing tangible benefit on a relatively large scale. Millions of people are accessing the materials and hundreds of universities are sharing open educational resources. I suppose its possible that these millions are people we don’t really want to serve and that all these universities have been duped and deluded into believing there is some benefit for them in sharing their materials as they are when there really isn’t. Or it’s possible that lots and lots of people are being helped by the open sharing of educational materials and that there really are tangible benefits for universities even if they turn out to be not the ones that we were expecting.

I learned this early on in my experience with OCW. If your asked 2003-vintage me what the benefits of the project would be, I’d have said, “Oh, definitely, with faculty reusing the content. This is stuff someone is going to have to take and modify and teach in a classroom. It’s not stuff you just dive into without guidance.” When the early returns from our surveys indicated half the people visiting our site had no connection to a university, either as a faculty member or student, I decided it was a better idea to look at what the data was saying rather than guess how I thought the resource would be useful. After all, you can beg a chicken all day for milk and she’s only ever going to give you eggs, but eggs are pretty good too.

Is OCW useful to formal higher education? I’d say the data indicates it is. After all, if 50% of our visitors are not associated with a university, that means that about 50% are. That’s a lot of people, 400-500K a month to the MIT OpenCourseWare site alone, and from virtually every higher ed institution out there. We’ve gotten 2,600 visits this year from BYU, not all of which I assume are coming from David. Have we documented a raft of cost-saving opportunities for faculty and students at other universities? Again, not a lot. One third of students at other schools are using OCW to complement materials from their enrolled classes, and 12% of those indicate the site has saved them money in doing so. That’s a relatively small portion of the overall use. Does this mean there are no benefits?

If I had to put a stake in the ground on how OCW generates benefit for others, right now I’d say primarily as a reference tool that is used for a range of academic activities, including independent (not distance) learning, curricular planning and development, supplements to classroom learning, academic planning, and professional development and problem-solving. Interestingly, many of these benefit from accessing the materials in situ, embedded in the OCW site and MIT curricular structure, rather than disaggregated and localized. Thus, linking is a better strategy to support many of these activities.

David discusses at length the evils of linking, making absolute (and intentionally provocative) statements that in his mind follow from reliance on it at the expense of remix/reuse:

If linking is going to constitute the primary method of adopting OER, every penny spent on the process of openly licensing material for OCW or OER publication has been wasted.


When you define “adoption” as linking, there is literally no need to concern yourself with licensing or openness. When you define adoption as linking, you undermine everything that separates OER from the other resources on the web.

Here again, only true if you believe adoption and localization by individual faculty members is the primary reason for the open licenses. If you take a view that reference uses are a large part of the benefit, and you look at the area where the open licenses have been employed to greatest effect, it becomes clear that the licenses are quite important. Since we launched the site, more than 800 translations of our courses have been made into a range of languages. My best estimate is that back-of-the-envelope these represent about $10 million in funds and effort contributed by other organizations, and all made possible by the open licenses. They’ve attracted a huge amount of traffic–at least 30 million visits–and I would guess that most of the use on the other side of the language divide has been largely reference rather than remix as well. Licenses are a vital part of providing additional access to the content.

How does OCW benefit MIT if not through cost? Right now I’d say largely though transparency. MIT as an institution has better visibility into what it teaches and how; faculty teaching advanced courses understand more about what their students learn in foundational courses. Faculty also likely improve their materials in preparing them for open publication; they definitely make new connections based on the open publication; they increase their own professional standing and that of their departments. Students at MIT make much better academic decisions and understand how their chosen field relates to other disciplines (crucial in addressing cross disciplinary challenges like energy, cancer and environmental preservation); before they come to MIT, they have a better idea of what the academic experience at MIT will be like through visiting the site. Does any of this make MIT any cheaper? Maybe at the margins. Does it make MIT better? No doubt.

Update on OpenStudy

The OpenStudy study groups continue to grow at a blistering pace:

MIT 6.00 Intro Computer Science – 1254 subscribers
MIT 18.01 Single Variable Calculus – 1084 subscribers
MIT 21F.101 Chinese I – 303 subscribers

Keep on rockin’ in the (open) world

August is usually kind of a sleepy month for MIT OpenCourseWare traffic-wise, with folks in many parts of the world enjoying vacations and schools being out of session. This makes it all the more surprising that MIT OpenCourseWare set a monthly traffic record last month, with more than 1.6 million visits. We also topped 1 million monthly unique visitors for the first time. The final numbers:

Visits: 1,602,561
Unique Visitors: 1,015,112
Ave. Visitors per Day: 51,695

Articles from Lifehacker, Hacker News and TIME helped, as did significant traffic from and Good momentum to carry into fall.

Driving traffic to OCW

One facet of my job is to manage media relations for MIT OpenCourseWare, doing the typical work of responding to inquiries and writing press releases, and also working proactively to get coverage of the site. Historically, media coverage has been a pretty good driver of traffic to the site, with articles in the New York Times, AP, and Christian Science Monitor leading to big spikes in traffic. The big jumps in the chart below at the beginning of 2007 and 2008 are just such media hits. In the last couple of years, however, media coverage hasn’t resulted in the kinds of big traffic jumps we’ve seen in the past.

Graph of MIT visits since 10/2003, starting at 250,000 and ending at 1,500,000

In part, this is because Open Educational Resources has become a much larger and more vibrant field than when we started our OCW, and I’ve always said this was the only business where you could feel good about losing market share. In many of the recent articles, we are either one of a number of resources discussed, or simply included as a background reference when talking about another resource. I think we’ve also managed to saturate the audience that frequents major news outlets as well.

What has been driving traffic to OCW lately is social media sites such as Reddit and StumbleUpon. A good example is this Reddit hit from last week:

Redit tag reading in part "MIT offers free online course materials including video lectures..."

Not calling out a specific resource, just a general, “Hey, didja know?” but nonetheless got voted up and drove an impressive amount of traffic to our site:

Screenshot of Google Analytics 8/15 to 8/20 showing 100% increase in traffic on 5/18

We more or less doubled traffic to the site on the day it hit, 82% of the Reddit visitors had never been to the site before, and 54% stayed for more than one page. I am not sure how you build an awareness strategy around this, as I’m sure you can really piss off a web community if they feel like you are gaming their referrals, but it definitely shows the power of these social media sites in driving high-quality traffic.