Waypoints in the MOOC Debates, Part III: The Udacity-Georgia Tech Contract

Photo: Edu-Factory Congress, U Paris VIII, February 13, 2011
I have spent some time trying to understand the MOOC business model, and yesterday Inside Higher Ed published one result, my 2000 word study of the Udacity-Georgia Tech contract, "Where are the Cost Savings?"   There I try and fail to use two public spreadsheets to discover where Udacity saves money for Georgia Tech's Computer Science department-- beyond what it could save with its own "blended" online master's program.  I focus on a MOOC company's cost claims because these have turned the heads of politicians and academic managers across the land. 

Yesterday afternoon, Udacity founder Sebastian Thrun blogged on the Georgia Tech deal and commented on some of the points of my article. My MOOC research has been motivated by a concern about the gap between MOOC rhetoric and MOOC reality, and I certainly stand to learn from Dr. Thrun's explication of his own contracts, about which he knows quite a bit more than I do.

I haven't found Dr. Thrun's post so helpful about the numbers.  But it does offer an important retrenchment in MOOC rhetoric.

The first of these consists of Dr. Thrun's remarks that online education is "nothing new" and "will not replace face to face education."  This is an about-face from six months ago, when Dr. Thrun told The Economist that the coming of the MOOC meant that "in 50 years there will be only ten universities left in the world."

The second retrenchment is Dr. Thrun's repudiation of Daphne Koller's claim at the UCLA MOOC summit in January that MOOC's can teach at "effectively zero dollars marginal cost per additional student."  This challenges the promise that Udacity and its rivals will disrupt the cost structure of a higher education sector, though this promise was the basis of the international passion for MOOCs.

The third retrenchment is the post's claim that "meaningful and high quality education" requires "human contact and mentoring."  The MOOC pedagogical revolution turns on powerful kinds of instructional automation, and yet here we are back in the familiar world of colleges and universities where "blended" courses use technology as a supplement to labor-intensive hands-on contact.

The fourth is that Udacity's technology won't furnish Georgia Tech with a super-cheap yet excellent infrastructure. Year 1 costs are as high as I said--which is indeed normal for a product that still needs lots of customization labor.  Cost stay high in Years 2 and 3, a fact that I would have preferred Dr. Thrun to discuss.  His remark about "variable costs" in out years doesn't specify what those later per-student costs might really be.  Costs in years 1-3 still sadden me: in exchange for offering up 20 courses worth of intellectual content, faculty might like to get a true labor saving device, which has been the promise of MOOCs, as of all  technology. They don't get this.

The fifth retrenchment is that Udacity is going to get non-degree students to pay tuition, who will furnish Udacity with "the majority of tuition revenue."  The original MOOC claim to democratize global higher education access, however, depended on MOOCs being free.

These concessions, taken together, mean that MOOCs haven't solved "Baumol's cost disease" in universities. They mean MOOCs will be less than free and thus less globally democratic than claimed. They also mean that Udacity isn't going to be a MOOC after all, but some kind of MOOC-online college hybrid. 

Perhaps I have misunderstood. In any case, my first instinct is to welcome Sebastian Thrun and the Udacity crew back to the real world of educational problem solving. I  am moved by Dr. Thrun the educational humanist, who expresses a desire to use his company to help "educate 1% of underemployed Americans to become proficient in computer science."  Truly open access has and is the best part of the MOOC agenda.

On the other hand, Dr. Thrun is the founder and CEO of a startup company.  Since he operates in a capitalist economy, he is obligated to maximize returns to his company and not to other people, including his partners and the public.  He must also market his products as of enormous benefit to every customer and to all mankind.  To simplify, this business context commits him to the following positions, which are liabilities in his post.

1. Budgetary nontransparency. Dr. Thrun doesn't work through the contradictions in the budget and enrollment models. The main new information is (a) that Udacity will be controlled the costs of contact hours by possibly doubling them; and (b), as mentioned,that they will have very high enrollments while nonetheless charging tuition.  If we're being offered transparency, we should get the full spreadsheets on this deal, including out-year cost estimates.  But such spreadsheets are considered proprietary information, trade secrets retained for the advantage of Udacity over its competitors--including public universities. It's economically logical that Dr. Thrun isn't offering real transparency here. He is of course functioning as a businessman.
2. No public funding crisis.  Dr. Thrun makes no mention of the dramatic funding cuts that have degraded the educational services he claims to fix. Nor, naturally, does he mention his own role in convincing major state politicians like California's governor and Senate President Pro Tem that they don't need to reverse cuts to improve public university quality. The Georgia Tech spreadsheets are (incomplete) proof that online technology cannot make up for massive public funding cuts. So now, in many states, public universities are in limbo--not saved by a MOOC cost revolution that has not materialized, and yet impoverished by a MOOC-fed consensus that the "traditional" college is dead. For example, Iowa's governor invoked online technology in vetoing capital expenditures for pharmacy, biosciences, and education buildings ( (h/t Teresa Mangum).
3. No detente with the labor to be replaced.  Large MOOC savings will occur only if tenure-track faculty are fired and/or not replaced, either in existing colleges or new quasi-universities.  MOOC marketing has turned in large part on descriptions of faculty as having put their vested interests ahead of students and progress in order to protect professional skills that technology has made obsolete.  It has paralleled attacks on "teachers' unions" in K-12 "reform," and appeals directly to executives in academia and government while ignoring grounded descriptions of current instructional practice even from a department at Udacity's most important university partner (Philosophy at San Jose State University).  Now Dr. Thrun grants the importance of "human" teaching, but still not of faculty teachers.

This is to say that although the financial claims of online technology are being hollowed out, the MOOC storyline will continue as it has. This is to be expected of Dr. Thrun: it is his job. 

Yet though we shouldn't expect a company CEO to protect the public interest, we can and should expect it of public officials. After 18 months of MOOC-promises of a cost revolution, the public discussion of MOOC budgetary detail boils down to one intrepid reporter, Ry Rivard, who got the spreadsheets through a public records act request, one professor who spent hours thinking about them, and one company executive who replied.

If I err, please enlighten me.  But Harvard and MIT committed $60 million to edX, tens of millions have gone to other companies, and yet university managers have failed to provide concrete, public due diligence. Before SJSU or UCSD or Georgia Tech go further, their administrations need to quantify the workforce implications and the precise expected cost savings for their communities.

Dr. Thrun ends his post with a dig at existing universities: so maybe Udacity isn't that cheap, but neither are you. He asks rhetorically, "Where does all the tuition money and the state support go?"

This is easy to explain.  The state funding and tuition money go to:

• "Human contact." The student services and personal, individualized instructor responses that increase retention and success
• Research.  This is expensive for creators and cheap for free riders. But somebody still has to pay to do this, and to maintain, administer and upgrade the infrastructure, which are noble and necessary activities.
• Innovation. For instruction and research, this means time and effort on top of the task at hand.
• Middle class academic jobs.  This means good wages, plus benefits and retirement, which allow stability and the freedom to be creative, to try new things, to do exploration rather than just the self-justification and defense that public colleges have been saddled with. 

I completely reject the ambivalence of this society--and especially of Silicon Valley--about whether we need and can afford these things. We do and we can.

Campus Administrators and State Legislators Really, Really, Really Need to Read This Article

Chris has an analysis in today's Inside Higher Education that takes a careful look at the financial numbers of the Georgia Tech/Udacity deal creating an online Masters of Computer Science.  Based on the contract information obtained by Ry Rivard of IHE, Chris shows that while this is clearly a good deal for Udacity it is much less clear that Georgia Tech is gaining much in terms of revenue or capacity.   The bottom line seems to be that Universities would be much better off generating their own platforms and deploying their own faculty, staff, and graduate students if they want to move more fully into online experimentation.  The other bottom line is that the most fundamental value is being created not by the platforms, or by administrators overseeing platforms, but by the faculty and course designers working at ground level.

Chris's analysis can be found HERE.

Rivard's article on the contract can be found HERE.

Will UCOP's Latest Efficiency Tax Employees?

By Onuttom Narayan (UCSC)

Under the name of UCPath, the University of California will move to a single payroll and HR system for all campuses and medical centers “that meets the core needs of each location while capturing the efficiencies, improved data and cost-savings associated with unified systems.”
Unfortunately, one of the likely consequences of this will be a new “tax” on faculty research grants. This is because one aspect of the new system is the introduction of “Composite Benefit Rates” (CBR), where employees are divided into a small number of groups, and the benefits charged to the budget for each employee are based on the average cost for their group instead of the actual cost of their benefits. If all salaries were being paid from state funds, this would be a harmless accounting change. But this is not the case, and as the CBR Steering Committee's final recommendation admits, there will be winners and losers.

Most notably, for faculty members who are academic-year employees, funding sources (e.g. research grants) that pay their summer salary will be charged for benefits at the same rate as for academic year salary. At present, the summer benefit rate is much less (except at UC Davis which has already implemented CBR), because there are few additional benefits associated with summer salary: medical insurance coverage is already provided for the whole year for academic year employees and therefore there is no additional employer contribution for the summer; summer salary is not covered compensation for UCRP, so there is no employer contribution to UCRP for summer salary (there is a small contribution to a defined benefit plan that is included in current summer benefit rates). Because of these, for every dollar in faculty summer salary from a research grant, currently the University only charges the grant an extra 10-14 cents for benefits. Under the new model, the same grant will be charged the 30-35% academic year benefit rate, i.e. approximately 20 to 25 cents more without any increase in benefit expense to UC.  (1) With indirect costs (overhead) charged for the benefits, the new “tax” becomes 30 to 35 cents on the dollar. Summer salary from other sources, such as from summer teaching, will be affected in the same way: the funding source will have to allocate more money for benefits, but it is only for research grants that the faculty will have to figure out where to find the money.

The implementation of the system, which was originally scheduled for this year, will happen with very little warning. Faculty members with current research grants will face the choice of reducing their summer salary or reducing support for graduate students or post-docs.

The CBR Steering Committee euphemistically describes the inclusion of faculty summer salaries as “a topic of interest during this project by many constituents.” This is another way of saying that committees of the Academic Senate strongly resisted this plan, were initially told that this would not happen (and even, at some point, that no benefits would be charged for summer salaries) before discovering at the last minute that the Steering Committee's recommendation had changed. When this issue is raised by the Senate, senior management complains that UC receives far less indirect cost than do private universities—an issue that has nothing to do with benefit costs and is an indirect confession that the plan all along has been to squeeze extra funds from faculty grants.

The change is claimed to be driven by the Division of Consumer Affairs (DCA) of the federal government, with which universities negotiate the charges they levy on research grants. In December 2012, the University Committee on Planning and Budget (UCPB) of the Academic Senate was told that UC representatives – including the Chair of the Academic Senate – had met with DCA and found them receptive to separate benefit rates for summer salaries. In February 2012, UCPB was told that in subsequent meetings with UC representatives – this time without Senate members – DCA eventually decided that the University would have to charge either the same benefit rates for summer and academic-year faculty salaries, or charge no benefits for summer salaries. Nothing in between was permissible, including a rate that would approximate actual charges, and the University eventually decided that charging nothing for summer benefits was unaffordable. At present, the University calculates benefits individually for each employee, a process that the CBR Steering Committee describes as equivalent to “hundreds of thousands” of benefit rates. No explanation was given why federal agencies would, when presented with a proposal to reduce these to a handful of rates, insist on a further reduction that would eliminate a separate summer salary rate, a reduction which would cost federal agencies more than what UC proposed!

Apparently the UC Berkeley administration has gone to DCA to request that Berkeley be allowed to set a separate summer benefit rate. If this is successful, it is possible that other campuses will follow suit. Otherwise, the faculty at large will learn about the new system when a fresh bite is taken out of their research grants.

That Was the Week That Was

As you know, the California Legislature --following negotiations with the Governor--passed the 2013-14 budget and follow up bills this weekend.  In the end there were few real surprises (although Jerry does possess a line item veto so stay tuned).  There were a few highlights concerning both the CSU and UC budgets that are worth noting.

1) Each segment will receive the proposed $250M conditional on foregoing tuition increases for the year.  In practice this condition means that the overall increase in higher education budgets will be somewhat less than $250M since a sizable portion is a buyout for the tuition.  But the State's contribution to the UC and CSU budgets are going up and students will receive some relief from the seemingly continuously increasing financial burdens of attending either of the segments.

2) Brown backed off on his demands that funding increases be tied to a series of mechanical metrics that he and his staff had proposed.  Both UC and CSU administrators and faculty spoke out against these metrics and their concerns were echoed by leaders in both the State Senate and Assembly.  That the Brown administration was attempting to increase burdens on Higher Education at a moment when the budget is--at best--backfilling several lost years of funding was not lost on the legislature.  As the LAT reported, Susan Bonilla, who Chairs the Assembly's budget oversight of Higher Education took the lead in questioning Brown's logic:

Bonilla said she was concerned that Brown's higher-education proposal was generated by number crunchers without enough input from educators. During an April hearing, she said the governor made it sound "like we're talking about a factory" rather than a university.
And universities are in a rebuilding phase, she said: "I don't think that's the time to say we're going to withhold your money."

 3) AB94, Speaker Perez's proposal to create a "Middle Class Scholarship Fund." passed both houses.  Perez's Bill will establish--starting in 2014--a fund through which California Residents students whose family incomes are less than $150,000 a year will be eligible for scholarships of up to 40% of their mandated tuition.  Certain restrictions--concerning GPA, application for other grants, percentage of total scholarship counting other grants, etc--will be put into place.  But the Bill recognizes that, despite the claims of the Segments to provide sufficient aid to ease the burden on lower and middle-class students, the State's turn to a higher tuition/higher aid model has placed tremendous burdens on the majority of students and their families.

These developments followed, of course, a less positive example of Sacramento politics: the passage of Senator Steinberg's SB520 by the Senate.  Despite the growing evidence that Coursera and Udacity do not, in fact, offer a smart option for public higher education, Steinberg has continued to offer one version after another of his effort to find ways to press universities to partner with for-profit ed-tech companies.  I have offered analyses of 520 and its ever changing contents here, here, and here.  As I have indicated throughout, SB520 is a bad idea, displaying a remarkable ignorance of the way higher education actually functions and serving more to provide the appearance of tackling the problem of access and achievement than to actually do something meaningful for access and achievement. 

But I want to make a different point here.

As bad a proposal as SB520 remains, it was much more destructive in its earlier forms.  Steinberg has been pushed to drop many of the worst elements of his bill and to lessen, if not eliminate, the more intrusive and onerous mechanisms to open up a new market for Silicon Valley venture capitalists.  And while it is clear that the Administrators of the 3 segments played a part in pushing back against Steinberg, it is worth stressing that faculty organizations applied continual and effective pressure against 520.  I am less knowledgeable about the faculty efforts in CSU and CCC but in UC, the Academic Senate, The Council of University of California Faculty Associations, and The UC-American Federation of Teachers all organized against the Bill, met with legislators, wrote critiques, and in many cases testified against it (and Jerry's metrics proposals) in Sacramento.  Although they were not able to stop its passage in the Senate (where Steinberg dominates proceedings) they were able get significant changes. 

In addition, the Santa Cruz Faculty Association (the only FA with collective bargaining rights) was able to use its status to demand collective bargaining over any contract that UCSC sought to sign with Coursera.  Given that the administration had been proceeding with the contract negotiations with little, if any, faculty oversight this was a significant achievement.  

I do not want to overstate things here.   520 did pass out of the State Senate and administrators in both UC and CSU continue to pursue connections with Coursera and Udacity.   But the effect of the efforts of faculty organizations suggests that they are not as marginal as we sometimes fear.  Indeed, now is the time to build upon their success and deepen their effectiveness.  In the short term, 520 still has not passed the Assembly or been signed by the Governor--effective opposition may prevent it from becoming law.  More demands for oversight over contracts with online providers is a logical and necessary component of faculty commitment to the curriculum.   It is time to increase these efforts and increase the strength of these institutions.  They remain our best counter-balance to those who would redefine higher education.  Or at least they will be if we use them that way.

New Waypoints in the MOOC Debate, Part II: Lowering Revenue Expectations

In Part I, I described MOOCs as a symptom of the absence of educational ambition among politicians. Since then, the MOOC situation on the ground has been shifting day by day.  It's been driven by major deal announcements and by the first real studies of neo-MOOC learning patterns. It will likely be driven in the weeks to come by the revelation of the National Security Administrations's Prism program to use companies like Google to gather in secret all communications metadata on pretty much everybody.  I'll address the first of these issues in this post.

In May, Udacity and Coursera described what were at first greeted as landmark partnerships with major universities.  The press take on the Coursera contracts with "ten public institutions" was that

Under the new deals, Coursera is recasting itself as a platform for credit-bearing courses that would be offered to students enrolled at multiple campuses within a public-university system.

 The cognoscenti's view was that Coursera had "jumped the shark":

Yesterday, Coursera did a weird strategy about-face by announcing that, rather than competing with public colleges, it’s going to start competing with Blackboard instead.

. . . Coursera has simply never had a coherent plan to generate revenue.  Oh sure, it had a bunch of ideas about how to do it, which were outlined in this leaked MOU with the University of Michigan, but few seem to have panned out.  The only thing we’ve heard from Coursera is that their idea for charging people for certificates of completion netted $220,000 in Q1 of this year.  Given that Coursera’s annual burn rate seems to be in the neighbourhood of $10M (that’s on top of their partners spending $50K/course to place it on the Coursera platform), this is peanuts. 

In the UK, Martin Weller announced that "You Can Stop Worrying about MOOCs Now." Noting Coursera's intention to explore "MOOC based learning on campus," he asked,

what does that even mean? That's blended learning, or e-learning, and again, not exactly new.  If you take the MOO out of MOOC you're left with just a C, and no-one's that interested in just a C. This follows on the back of Georgia Tech's not free, online Masters MOOC, which looks awfully like a not very well supported elearning course.

Back in the edu-consultancy world, University Ventures described the Udacity-Georgia Tech deal as a case study for Bad Judgment magazine:

The first instance of bad judgment is the University’s announcement that it will be hiring only eight new instructors to teach as many as 10,000 new students and may rely in part on peer grading. The “massive” component of the masters degree stems from its origin: a meeting between the Georgia Tech Dean and Sebastian Thrun at Udacity. As one of the three horsemen of the MOOC movement, Udacity is banking on “massive” despite the fact that the more Udacity tailors its approach to approximate existing online programs (with instructors and interaction, as with its partnership with San Jose State), the more it seems to gain traction. 

The reality is that (1) instruction and interaction are critical to successful student outcomes and (2) the cost of instruction and interaction is not what makes higher education unaffordable.

There have been a group of us, including for example Bob Samuels, that has been making these same two points for quite some time.  Last week, I wrote a couple thousand words about the Udacity-Georgia Tech budget projections that I'm hoping CHE will get around to publishing sometime soon.  Suffice to say that major cost savings cannot be the rationale for the Georgia Tech arrangement. In the ramp-up period, terribly high per-MOOC costs could be justified by mass enrollments, but unfortunately from the VC point of view the masses take these courses for free. These production costs also collide with increasing awareness of large faculty time inputs: Duke's Dan Ariely and Cathy Davidson report 150 hours of their time per hour of "actual MOOC."  Prof. Davidson's phrase in a subsequent post is "insanely labor intensive"  -- in exchange for a $10,000 stipend that she spent entirely on assistants. Many MOOC watchers are now concluding, as she does, that MOOCs do not have a way of making up for massive public funding cuts.

Unfortunately, the idea of exponential cost savings based on zero marginal cost per additional student, as Coursera's Daphne Koller promised, has never been credible. As we exit the marketing phase and start looking at real cost performance, this news will move up the administrative and political food chain.

We need to get back to fixing huge budget shortfalls before we lose another couple of budget years. The spring's new cost discussions have been a help.

Is Climate Change A Hoax?

By Jorge Mariscal (UCSD)

During the same week that the President’s Advisory Council on Campus Climate, Culture, & Inclusion met in Oakland, a meeting where “community” representatives presented out-going UC President Mark Yudof with an award for “his commitment to our issues,” the UC Associated Students issued a statement that lamented the hostile environment across the UC system.

The “Resolution Recognizing and Condemning the Acts ofAnti-Black Racism and Racial Prejudices at the University of California Irvineand San Diego, while Affirming Support for the Black Student Unions at UCI andUCSD” begins with a sentiment few on the President’s Climate Council wanted to hear:  “Whereas, instances of Anti-Black racism are common on all UC campuses, yet [the] administration continues to not adequately address these incidences and carry out adequate measures against those who enact anti-Black rhetoric and actions.”

The irony was clear to anyone who was paying attention.  The Climate Council had come into being in 2010 shortly after UC San Diego was shaken by a series of events known as the “Cookout/noose” episode.  Now, three years later, awards were being handed out at the Council’s final meeting just days after viral videos captured members of the UC Irvine chapter of the Lambda Theta Delta fraternity cavorting in Blackface.  Things had come full circle and nothing of substance had changed.

The mission of the original Council was “to identify, evaluate, and share best practices in order to ensure a welcoming, inclusive, and nurturing environment across UC’s ten campuses.”  Local councils were mandated for each campus. They developed randomly according to existing campus structures.  Some were effective; some were a complete waste of time because they recycled existing approaches that were ineffectual. The system-wide Council began on a positive note with guest speakers such as Sylvia Hurtado from UCLA explaining to Council members the close connection between diversity and campus climate. 

It wasn’t long, however, before the attention of the system-wide Council was diverted away from the recent traumatic events.  Former Provost Larry Pitts announced that he did not want the climate Council to be a diversity council (guess he missed Hurtado’s presentation).  Council members then learned that the next meeting would not be held at UCOP offices. Instead we would meet in Los Angeles at the Museum of Tolerance where we would take a tour of the exhibits.

Clearly, the Council’s business had been caught up in a hazy notion of tolerance that more often than not, as Wendy Brown has taught us, “iterates the normalcy of the powerful [and] regulates the presence of the Other.”  (Regulating Aversion, 8) The hard work of trying to understand campus climate on the ground would not be encouraged.  The outright hostility some student groups had recently experienced and still had to cope with on a daily basis on UC campuses was washed away in a flood of feel-good “can’t we all get along” liberalism.

Through the efforts of a small group of Council members, an attempt was made to tackle real issues.  With the support of Dean Chris Edley, working groups were formed on faculty diversity, the status of LGBT communities, and other pressing issues.  Working group members dedicated dozens of pro-bono hours preparing reports that, as is most often the case, were submitted, circulated through Senate circles, commented upon, and filed away for the edification of some future scholar.  Minor recommendations were accepted at some campuses, but overall the Council’s impact was minimal.  The climate had not changed.

Just five days after President Yudof received his commendation on May 2, a Black student at UC Irvine reached into her bag and found a note that read, “Go back 2 Africa slave.”  UCI administrators reacted with appropriate horror and “we will not tolerate such behavior” pronouncements.  Racism and sexism on UC campuses seemed to be caught in a never-ending loop of the film Ground Hog Day.

No one, not even the students who wrote the AS resolution, would argue that administrators have the power to stop racist or even garden-variety stupidity.  But clearly something has to change.  What can be done?  No “climate” or “diversity” commission can hope to improve that situation until faculty members are educated about what their women and minority students experience outside of the classroom and lab.  No student will be dissuaded from staging a racist and sexist “cookout” until campus demographics include more students, faculty, and administrators from historically underrepresented communities.  No campus will make progress on any equity issue without strong and fearless advocacy emanating from the top of the administrative chain.

Yesterday, I attended a meeting of more than a dozen departmental representatives who had been asked to poll their colleagues about “diversity” issues.  The most repeated comment was “We don’t know what diversity means.” The psych and anthro folks offered arcane definitions from within their disciplines; one engineer laughed and asked if a candidate who grew up with five sisters could list that fact as part of his contributions to diversity.  No one in the room seemed to know that there is an elaborate scholarly bibliography about what diversity means in higher education, how it affects the campus environment, and how it can function to either impede or improve student academic success. 

In order to avoid a heated debate, the “diversity officers” who ran the meeting declined to clarify their language.  No one pointed out that until diversity and equity are understood in an historical frame that reminds us that for decades specific groups were excluded from higher education we will continue to have pointless meetings like this one.  In the meantime, random incidents of racism, sexism, and homophobia will break out periodically, temperatures will rise across the UC system, and climate change will continue to be a hoax perpetrated by the keepers of the status quo.