Description

I think there’s a good chance TWOU disappoints on aEBITDA margins this year and next and at 14x TTM revenue that wouldn’t be good for the stock. They report after the close, but I don’t have any special insight into the announcement. Since 2018s aEBITDA is very backend loaded, I don’t see why they’d guide down now, but who knows given the issues below. The high valuation, increasing competition, good chance of margin disappointment, aggressive accounting, weakening pro forma growth, heavy insider selling and large secondary offering at much lower prices together make this a good candidate for a short in my book. The risks are obvious: it’s a momentum stock lumped in to the SaaS category (even though it shouldn’t be due to its business model) and there is seemingly high visibility into revenue growth.

2U is growing at 30%+ and is a leader in the online program manager (OPM) industry, where it competes primarily with Pearson Embanet, Wiley, Bisk and Academic Partnerships. It provides and runs online degree programs for universities, and takes about 60% of tuition revenue for its services under contracts that have 10 to 15 year terms. Under these contracts, 2U is typically responsible for porting university curriculum to an online format, marketing to and recruiting students, counseling students with things like application completion and financial aid, and supporting faculty. With some contracts, 2U is on the hook for payments to universities if the online program enrollments disappoint.  

Most universities were late to the game in putting degree programs online, meaning they had a lot of pressure to launch quickly. This put OPMs like 2U in a good position to get universities up and running fast. In return OPMs negotiated aggressive revenue share contracts where they take 60% or more of the tuition. While many universities appear pleased with 2Us offerings, there seems to be some resentment over 2Us high level of fees.

“The revenue share is outrageous,” comments NYU Steinhardt’s Brewer, who partnered with 2U and HotChalk to launch several degrees. “But, of course, we couldn’t have done it ourselves.”

There is also a conflict of interest between the non-profit university and the for-profit OPM, which is laid out in detail here and here.

“The revenue-sharing model of fee structures utilized by OPMs such as AP, Pearson, Wiley, and 2U is particularly problematic because it establishes clear financial incentives for OPMs to make online programs larger and more expensive for students, while simultaneously reducing expenditures.”

And from someone that “helped dozens of presidents and provosts design, launch, and market online degree programs”:

“Over the years it became evident to me that in the rush to move courses online, the revenue-share model encouraged a cookie-cutter approach to online learning that, all too often, misaligned incentives in ways that put scale ahead of quality and revenues before outcomes. The dean of a top 20 business school once put it to me this way:”

"I'm not giving you 50 percent of revenue and I don't need all of your services. It sounds like your primary value proposition is marketing support. Building high-quality offerings and providing thoughtful instructional design support to our faculty is what is most paramount to us, not scale. I need to be able to select the services I need on an a la carte basis and pay you a fee for that service."

These issues have opened the door for new business models and competition. Indeed, 2U co-founder Katzman left 2U in 2012 to help build Noodle, which allows universities to start an online program at half the cost or less, with much more flexibility. Katzman doesn’t mince words regarding the revenue share model at 2U.

“For a fee, we help schools assemble the tools, services, and tech to run great programs without taking money from students—it’s more flexible and transparent and wildly less expensive,” Katzman argued. “The only real question is, how quickly will the old revenue-sharing model die?”

Noodle’s far lower fee structure and flexibility, other competitors, and increasing questions about the OPM model have to be putting pressure on 2U’s model.

Did these pressures have anything to do with 2Us contract renegotiations with its largest customer in November 2015 and April 2016?  2U says the renegotiations were initiated by 2U improve 2Us flexibility. However, these renewed contracts reduced 2Us tuition revenue share, provided for 2U to make significant additional payments to USC and inserted a clause where 2U would be responsible for making additional payments to USC if certain net program proceeds fell below a threshold.

Georgetown School of Nursing launched its online program with 2U in 2011. Revenue ramped from $8m in 2012 to $13m in 2013 and to $15m in 2014. But then it fell to about $12m in 2015 and Georgetown fell off the >10% customer list in the 2015 10-K. Perhaps not coincidentally, Georgetown renegotiated its contract with 2U in 2015 (2U called it a “contract extension” in its 3Q earnings release). The almost 20% revenue drop in 2015 with a growing program raises the question if there was a huge revenue share reduction in the revised contract.

I think 2U is becoming increasingly aggressive with its accounting and disclosures to hide weakening trends.

• Some hide the football. 2U reduced its disclosure to investors significantly this year. They stopped disclosing the revenue growth from various program cohort years. And with the upswing in capitalized technology and content and payments to university clients this year, 2U removed both of these categories from the balance sheet and lumped them into other categories. https://www.youtube.com/watch?v=X7oKBjMwS30

• Sharply rising payments to university clients that aren’t being recognized on the income statement. There maybe legitimate reasons for capitalizing these skyrocketing payments, but this is a new trend it’s not good. YTD through Sep2017 payments to university clients totaled $12m, a $13m swing from the prior year period when these payments flowed to 2U instead.

• DSO is climbing sharply. More aggressive revenue recognition perhaps.

• The company focuses investor attention on adjusted EBITDA. This figure excludes 7.5% of revenues that goes to stock comp. It also excludes large amounts of capitalized technology and content that is program specific and if iit were expensed, would be 8.2% of sales. Combined these represent a whopping 16% of revenue. Also excluded from EBITDA is corporate PP&E of course, which is running at 10% of revenue.

• I don’t need to see this to know what might be going on, but it does make me smile when I come across a comment like this from one of the former corporate accountants:

Accountant (Former Employee) –  Landover, MD – August 29, 2017
“CEO like to think he is a rockstar. Great business plan and execution but the inner cirlce pushes the ethical limits at times.”

Separately note that 2U has very high customer concentration with 53% of graduate program segment revenue with only four university clients and 82% of short course revenue from 2 customers in the first nine months of 2017. In 2017, about $50m of the company’s $14m in aEBITDA will come from just four programs at three universities that were launched from 2009-2011: USC, Georgetown and UNC. I don’t think these early successes are representative of future programs’ financial contribution.

Valuation

As mentioned, the stock is trading at 14x trailing and 10x forward revenue.

I spent way too much time modeling this business and looking at the valuation different ways. There are a lot of assumptions that go into it, but I think the stock is worth half its current value at best.

One way of looking at valuation is a DCF/NPV approach per launched program. Using a 10% discount rate, a longer than normal 15 years and even giving full terminal value credit (i.e. assuming program continues for a long time after the contract ends, which is probably unrealistic at the then current EBITDA) results in an NPV of $10.5m per newly launched program (note $10.5m was using excel’s NPV fn but when I calculated it I got $11.5m; conclusion is the same). Then subtract something for capitalized technology and content, which has been running at about $2/launched program, you can arrive at about $8.5m/program NPV. Note general PP&E has also been running at $2/launched program but this is not included in the NPV (should be but whatever).

These figures are based on management’s disclosures in transcripts, news articles, investor presentations and SEC filings.

Catalyst

disappointing margins