Aggregation Theory and the legacies of Paul LeBlanc and Chip Paucek

Is there a limit to the potential scale of educational institutions?

Hello!

I hope that many of you enjoyed Sunday’s edition of Weekend Reading. There will be one more Weekend Reading this weekend and Funding + M&A Update next week before I pivot to end-of-year thoughts and a holiday break.

With that, on to the update.

---

I normally use my second email of the week for Funding + M&A Updates. But, on Wednesday SNHU announced that President Paul LeBlanc was leaving. His retirement comes just a couple of weeks after 2U replaced founder and CEO Chip Paucek.

The circumstances of their departures are different, and the announcements had nothing to do with each other. But the timing is helpful to internet historians like me who like to think in terms of epochs.

Chip and Paul are two of a small circle of people who have driven the narrative in the education sector for the last two decades. Between them, they have conferred north of 250,000 college and graduate-level degrees (and 100K+ sub-degree credentials). Their organizations, and those affiliated with them, currently enroll 300,000+ students. They are two of the only people you know who (probably) have Google CEO Sundar Pichai’s personal phone number - SNHU and 2U spend $40M+/year on Google ads (and ~$200M/year in total advertising).¹

However, this essay is not a celebration (or denigration) of the achievements of Chip and Paul. Rather, it is an exploration of why their respective platforms are not bigger.

Aggregation Theory

The best explanation I have for why SNHU and 2U are not bigger is that education is hard. It defies Aggregation Theory.

Aggregation Theory is a perspective on how the internet economy works coined by technology analyst Ben Thompson. For our purposes, the core principle of Aggregation Theory is that the internet allows digital businesses to serve users anywhere in the world with effectively zero marginal cost.

To make this concept tangible, consider this newsletter. If I wanted to send a physical copy of this note to readers, I would have to pay to print, package, and ship it every week. Assuming the cost to do this is similar to the holiday cards my wife and I just ordered (a generously low assumption, because the service we are using is internet-enabled itself), I’d be looking at $2/week, or ~$80/year per subscriber.

If this were my cost of doing business, there would be no free version of this newsletter and, frankly, probably no newsletter at all.

Fortunately, I have Substack! The cost to Substack of me sending an additional email is so minimal that they give it to me, and thousands of other writers, for free.²

The internet, and internet services like Substack, change the strategic challenge of building my business from a supply-constrained problem - Is there a reasonably affordable way for me to get this content into readers’ hands? - to a demand-constrained problem - Is this content good enough to build a business on top of?

Applying Aggregation Theory

This paradigm shift is great for niche businesses like mine. It is even better for category-defining internet businesses, or aggregators, like the New York Times, Netflix, and Google.

Each of these companies uses the same principle as my Substack in their business models - they assume that the cost of serving X users is basically the same as serving X+1 users. They have baseline costs - a newsroom, the production of shows, and the scraping of the internet - but those costs don’t meaningfully change based on the number of users/subscribers they have. Once their baseline costs are met, revenue from each additional user can be used for growth to reinforce their position as the dominant player.

It is really hard to compete once this aggregator flywheel - using zero marginal cost revenue to accelerate growth - starts spinning. The aggregator will always be willing to spend more on customer acquisition than its less secure competitors. Until the market is no longer in competition, as happened with Google in the early 2000s. Then, the marginal revenue can be used for profits and/or self-driving cars.

Applying Aggregation Theory to Education…or not

The Internet economy favors niche businesses that couldn’t otherwise exist and aggregators that leverage zero (or low) marginal costs to substantiate heavy spending on customer acquisition.

This is true across industries and around the world. (Within reason, certain geographies and cultures - like India and China - seem to encourage local aggregators over global aggregators.)

Except in education.

Education, particularly degree-granting education, remains oddly un-aggregated. Paul LeBlanc’s SNHU is the largest university in the US by enrollment, with 170K students. This is, objectively, an enormous amount of students for one university. Yet it doesn’t even represent 1% of total university enrollments in the country.

SNHU is huge, and growing, but it is hard to argue that it is pushing out non-aggregator (or worse-aggregator) competitors in any meaningful way.

Is 2U/edX an aggregator? The company, alongside Coursera, has certainly aggregated a lot of users - edX has 73M people on its platform, Coursera 136M. 2U enrolled 45K degree-seeking students this quarter, and has enrolled as many as 60K students in previous quarters. Coursera enrolls ~20K degree-seekers.

This is, again, an enormous number of students. If 2U were a standalone university, it would rank in the top 10 by online enrollment. But it would also represent < .25% of total US enrollments.

So what gives?

Why does the median student from a public institution attend a school less than 20 miles from home? Shouldn’t they want to find the “best” school for them?

More practically, 2U and SNHU have a billion dollars in annual revenues, shouldn’t their spending power be able to crowd out smaller competitors?

I’m going to run through a list of common rebuttals to this quandry, but the punchline is no one knows.

But, but, but…

Pre-COVID, the most common objection to the premise that a university can use zero-marginal-cost design to become an aggregator, scaling infinitely(ish), was that online content was inferior to in-person. I think we can mostly put this one to bed at this point.

The second most common objection is teachers. How does a school scale while maintaining quality instruction?

This is a reasonable but dated objection. The reality is most universities - even selective universities - already treat instructors like widgets. 71% of US university faculty are now non-tenured, and half of these non-tenured faculty are part-time. The median salary for an adjunct professor is $24K. Furthermore, in WGU and Capella, we have examples of reputable, accredited, and national online universities whose use of competency-based pedagogy dramatically reduces their need for human instructors. These universities are, like SNHU, very big. But they are not pushing other universities to closure.

The last, and most reasonable, objection is time. We’ve been doing education a certain way for a long time and so adapting to a new, online modality will take more time. That objection is…fine? I certainly can’t disprove it.

But both Paul and Chip ran their respective (billion-dollar) organizations for 10+ years. In that same time horizon, both the New York Times and Netflix aggregated their respective industries. Uber and Airbnb, both founded after Paul took over SNHU and Chip founded 2U, aggregated other highly regulated industries in a shorter amount of time.

I could keep going, but hopefully you see the point. There are lots of reasons an individual might object to a specific online offering, but no fundamental reason education can’t be an aggregated, mass-market good. No reason that a “mega” university shouldn’t be able to grow into the mega university given proper finances.

And yet, here we are. With a cornucopia of universities surrounded by a constellation of vendors that support them. Teaching fairly similar content to fairly similar student demographics. With no substantial signs that change is coming to the market other than an exogenous population shift in the US.

The End

In sum, education is hard. Paul and Chip both managed to build substantial organizations during an unsteady epoch of education history.

Their departures are, again, not connected. All the same, I’m curious to see if we will look back on this as something of a changing of the guard. Longtime leaders like Chip and Paul (and Moodle CEO Martin Dougiamas) are leaving after steering their organizations through COVID. Pearson, Wiley, Anthology, and Discovery are also re-vamping their leadership (and, thus, their strategy) for a new world.

The optimist in me would like to believe that all this turnover is healthy, and will breed new ideas. The pessimist in me worries that new leaders bring new focus on short-term improvements to an industry that continues to hunt for a bigger vision.

1. SNHU breaks out their advertising spending on their 990 forms. Their 2022 990 lists just under $42M in marketing services from Google and just under $183M in marketing services in total. 2U’s earnings reports do not offer the same granularity, as the company combines marketing and sales expenditures in one line item. However, given that this line item was for $422M in 2022 (page 49), I feel reasonably confident in saying the company spends in the range of $40M directly with Google and $200M on total marketing expenses.

2. Yes, Substack is partially subsidized by venture funding. And yes, their free product is a loss-leaning means to an end - they charge me for offering a paid version of the newsletter. However, I still think we can agree that the cost to Substack of adding each additional subscriber is de minimis.