What the hell is going on with Lambda School?

The complicated business of selling social good

Sponsored By Axios HQ

Today’s newsletter is brought to you by Axios HQ, intuitive software from Axios that helps your team think more clearly, communicate more crisply, and send focused, effective updates.

Fortune 100s and startups alike use Axios HQ to craft smart updates that distill essential information in half the time — increasing transparency, boosting engagement, and building trust across the organization. Innovative teams know communication is the key to culture, connectedness, and growth, and use HQ’s intuitive platform to elevate company updates and get their teams smarter, faster.

Austen Allred is either an asshole or a savior—depending on who you ask. Neither one is his formal title, of course.  Officially, he is CEO of Bloom Institute of Technology, formerly known as Lambda School

Austen has been accused of being, at various points, a liar, a racist, and a man cosplaying at being homeless. For those who believe the savior narrative, though, Austen’s company is selling the ‘American-Dream-As-A-Service,’ and it's actually Lambda School's naysayers who are the real assholes!

Media coverage of his company is equally divided: negative articles call attention to Bloom’s allegedly exaggerated performance statistics, to its students' complaints about its cult-like atmosphere, and the fact that the entire executive team besides Austen has quit in the last year. In contrast, there is compelling and meaningful reporting about Bloom's successes—which, in my opinion, show how it can embody the best aspects of the American experiment.

This is all… rather a lot: it is a rare occurrence for a relatively small startup to attract so much attention, positive or negative. So for today, I’ll share with you an interview I conducted with Austen. Originally, in an effort to keep my life nice and easy, I was simply going to post the interview. But I couldn't help myself—I just had to do some analysis of all the claims and accusations being made by its stans and its critics. Why? Well, it strikes me that amidst all the hubbub, people seem to be forgetting to ask the key question: Is it a billion-dollar business or not?

The TL;DR  

  • Bloom is a for-profit institution that sells coding classes, but which has positioned itself as the answer to the structural inequalities in American society. 
  • It has thus far failed to fix classism in American society; but on the plus side, it has managed to build a business that is likely doing $10-15M a year in a competitive market, and which provides a service that a decent chunk of its customers find valuable. 
  • As far as Austen’s being an asshole, I do not feel qualified to comment (judge not lest ye be not judged, etc.) but he has managed to obtain the enviable position of fame/notoriety.
  • Whether Bloom is a billion-dollar business or not remains to be seen. My sense is that Bloom hasn't quite figured it out yet.

This piece is informed by my discussions with former students, industry experts, and an interview I had with Austen about Bloom’s rebrand. Before we dive into the problems, we need to separate hype from reality and analyze what the business actually is

A Quick Summary 

Bloom’s product is a 6-month coding bootcamp that claims to take computer novices and transform them into coding wizards. This, of course, is not unique in the slightest. There are over a hundred full-time, in-person bootcamps in the US and Canada, and over 120 universities offer somewhat comparable programs. Bloom also competes with more traditional computer science degrees offered at every major university, and it also competes with online courses like Udemy and Codecademy. Needless to say, most startups would simply not attack a market with hundreds of similar product offerings. 

Bloom’s original attempt at differentiation was to offer its students the ability to finance their education by entering into an Income Share Agreement (ISA), which allowed them to either pay $20K upfront, or to make $30K in gradual payments over time once they had graduated and had a job paying $50K or more. This was not a loan, per se—it didn’t need to be paid back if the student failed to get a job after graduation. 

The credit for the popularization of ISAs goes to Bloom. It means that schools are literally invested in the success of their students—in stark contrast to other financing options which demand that former students make loan repayments regardless of whether or not they had achieved success. Unfortunately, competitors caught on quickly to this trend: there are now 76+ bootcamps that offer ISAs. 

But still, it helps to be early. Bloom’s innovation allowed them to grow quickly, and they have had ~2K students per year since 2020. With roughly 25K bootcamp grads per year in the U.S., this gives Bloom an ~8% market share. Nothing crazy, but it’s respectable. Assuming an average payout per student of $30K, and a 5-year payback period, the business should be doing between $10M and $15M in revenue per year. 

With this context in mind, are these purported problems real? Are the complaints about Bloom all that dire?

Alleged Issue 1: Bloom exaggerates its student outcomes

Napkin Math Analysis—6/10 Truthiness Score

Somewhere on the website of every major coding bootcamp, there is some sort of placement rate claim: ‘99.7% of graduates get jobs,’ or, ‘Go to our school and your wife will never leave you.’ (Ok, that last one is a blatant falsehood that I made up, but I wouldn’t put it past coding bootcamps to try that out as a marketing claim). At any rate, Bloom is no different: its outcomes report proclaims a 74% job placement rate. 

However, the bad habit that exists across the entire industry is selective sampling: bootcamps will manipulate the ways they count their student population until the statistic looks good. For example, the total number of 2527 students originally enrolled at Bloom was whittled down to 1091 'job-seeking learners' for counting purposes. That shaving down took place with various categories like 787 students withdrawing from Bloom or 394 being classified as “Non-job seeking learners.”  The relative merits of these categories are up for debate but it is important to call out this is a universal trend by the industry.

Note: There have been accusations that these numbers are fraudulent. I am unable to comment on the veracity of those claims because I do not have access to the raw data.

The crux of the issue is that employment outcomes are not wholly predicated on technical skills. Instead, it has a lot to do with having the abilities, motivations, and resources necessary for an extensive job hunt. 

“One of the honest realities that we see now is [our] placement rate is very, very correlated with what a student does around looking for a job,” Allred said to me. “But you can't control if a student will look.”

As someone who has been unemployed an uncomfortably large number of times in his life, I can confirm that securing employment is not only a matter of skill sets, but more a matter of being able to hack an interview process. Learning to impress and/or manipulate hiring managers and recruiters is a skill that isn’t taught in coding classes or even in university classrooms, but instead is one that is learned in the hallways of privilege. (You know, bars that serve overpriced espresso martinis, expensive Barry’s Bootcamp classes, and places like that.) An individual's ability to cultivate relationships and learn particular patterns of speaking will unfairly streamline their passage through a job hunt much more than their ability to deliver a work product. 

The point is that even if Bloom had Alan Turing, Nathan Baschez, and Clippy from Microsoft Word among their coding instructors, it might make good programmers out of its students, but it wouldn’t be able to help them get jobs. This sucks, and it’s unfair, and it’s outside the capability of any single institution to fix. 

This problem wasn’t created by Bloom, but their ISA-based model forced them to walk right into the middle of it in a more direct way than other coding bootcamps with traditional payment models had. Here’s why: by using outcome-aligned financing vehicles, Bloom pushed the entire bootcamp industry to move from providing a service in the form of coding classes, to one that promises the outcome of employment. And unfortunately for the industry, successfully providing that outcome is predicated upon solving the litany of social problems that plague the hiring processes of tech companies.

Arguably the most cynically successful financial outcome for a company using ISAs would be achieved by one that leaned heavily into structural inequality. It could fill the school with college-degree-holding white dudes who like mechanical keyboards and the Matrix movies and maybe rock climbing (it’s very important to have these sorts of interests for cultural fit interviews). Their placement rate would almost assuredly be higher than a program filled with anyone who isn’t named John Smith.

In contrast to our fictional racist and successful school, Bloom’s program is roughly 20% Female, 80% Male with ~50% of students being white. Across all gender/race cohorts, 16.4% had a high school diploma or GED, with another 38.3% having completed only some college coursework. Bloom deserves credit for cultivating a student population more diverse than the average STEM program. 

I don’t find any coding bootcamps’ analysis of their outcomes to be totally intellectually honest because it’s pretty clear that they all practice the selective sampling discussed above. But I understand the rationale by which they justify their cuts to the numbers—they probably shouldn’t be penalized in their stats for having students who decide not to seek employment! But to what degree their lack of job hunting is the result of insufficient knowledge, inadequate resources, or changed desires is unclear.

And really, the bootcamps are just doing the same thing that all startups do: product marketers' favorite move is to put customer case studies on the front page of their website proclaiming some value prop. Bootcamps are just doing the same thing! The difference is that while B2B SaaS vendors saddle their customers with confusing software and semi-painful annual contracts, coding bootcamps often saddle lower-income individuals with lifelong debt that can be compounding and difficult to escape. There is merit to the argument that these schools should be held to much higher standards than a typical startup. Individuals deserve more care than institutions. 

In my conversations with bootcamp students, whether at Bloom or in other major programs, they all said that the bootcamps only partially prepared them for software engineering roles. It usually took months of work after graduation to secure their first position. That typically meant those who were successful in the job hunt were those who already had a technical background or the resources to spend months job hunting. 

Alleged Issue 2: The Payments are Incredibly Expensive

NM Analysis—5/10 Truthiness Score

While Bloom started with ISAs, they have recently launched the ‘Outcome Based Loan’ that is supposed to be even better. Some argue that these are both wildly overpriced. 

For the ISA option, Allred says that payments can sometimes reach $1,200+ a month, with Bloom capturing 14% of graduates’ monthly income. For the 48-month payment option, a student’s all-in cost tops off at $42,950. This is in contrast to the average cost of $13,579 for other bootcamps. That being said, Harvard is more expensive than community college—the schools with the best brand in any field will charge extra, and there are those who would argue that Bloom is a premium program charging premium prices. Most of the largest programs all charge $20K or more in upfront cost. Additionally, these ISAs can last for up to five years, meaning that if you get a tech job making more than $50K three years after graduating from Bloom, they still get their pound of flesh. However, my research on other schools' ISAs showed similar terms, with costs ranging from $20-40K, and income shares going from 10-15%.

The new financing vehicle that accompanied their rebrand, the ‘Outcome Based Loan,’ seems equally crazy. With an interest rate of 12.5% (!!!!!) which starts accruing on day 17 of the program, the $21K tuition cost is not cheap. Bloom justifies this by saying that if you don’t find a job within 365 days of graduating they will give you 110% loan forgiveness. All you have to do for 46 out of those 52 post-graduation weeks is:

  • Apply to 10 jobs a week:(This can be done locally or remotely.) 
  • Network with 10 people per week: (Bloom has tools to automate this outreach.)
  • Post at least 5 Github contributions a week: (Github contributions are not monitored for quality.)

To me, this seems very tough for someone to accomplish while also having a full-time job. But Allred disagreed via email, saying, 

“These are very simple to do—you could do all of this in a couple of hours a week. We’re really optimizing for the 0 vs. 1. ‘Are you looking for a job or not?’ Our data shows that if a student is looking for a job in any reasonable way they get hired, and we’re willing to stand behind that. What we can’t do is automatically refund everyone who doesn’t care to look for a job. So these are actually very lightweight.”

However, just because something is expensive doesn’t mean it’s a ripoff. To make a clear decision in this case, you have to put it in context with all the other higher education options available to students. The average income of a college graduate is $55K, with a typical debt load of $31K at 4.6% interest rate. Around 43M Americans currently have some level of student debt. Thus, Lambda’s costs are roughly in line with those of a standard U.S. college degree. However, only programs that offer ISAs care if you are employed upon graduation. The typical college program will saddle you with that debt upon graduation which typically takes 53 months (4.5 years). Bloom’s slogan could arguably be, ‘Get here, get in debt, and get out.’ Even if you did the full program and took a year to find a job, you’d still be 35 months ahead of the average college student. This is a bigger deal than it may seem—the time cost of money in your early 20s is incredibly high. The sooner you can get out and get working the better off you will be. 

The bottom line: Bloom may be among the most expensive coding schools, but the debt load its graduates carry is is similar to that of graduates of an average school, and you can be employed in half the time. Sure, you’ll miss out on college football and a social life, but you’ll gain back time. You could learn the same subject matter online for free via Youtube, but if you need structure, Bloom offers a $40K way to get it. Whether any bootcamp is worth it, though, is still up for debate. 

Alleged Issue 3: The Rebrand Was Done to Hide Icky Stuff

NM Analysis2/10 Truthiness

In my favorite reality television program, Love Island UK, the most dreaded pronouncement a contestant can receive from a potential love interest is ‘You’ve given me the ick.’ This is when, for some inscrutable reason beyond their ability to adequately explain, someone loses interest in you. Note: One of the highlights from this year was when a girl lost interest in someone because he didn’t like cheese boards.

There is something about Bloom (and Austen too) that gives some people the ick. Whether it is because of the hustle culture so proudly displayed on their website, or due to Austen’s sometimes divisive Twitter account with ~193K followers, there is a sizable contingent of folks who truly dislike the existence of the business. Whatever the reason, in my entire technology career, I have never seen a business so tiny covered to this extent. Most of that coverage has been negative, with posts from places like Business Insider, The Verge, and NY Magazine writing exposés. 

Ironically, as Allred points out, “It doesn’t look like it when you look at the articles, but they’re net positive for the business. Every time there’s a negative articles about us, our enrollment grows and more hiring partners reach out. I think there’s truth to the axiom that there’s no such thing as bad press.”

Note: This makes me want to get involved in some kind of culture war. I’m willing to wager that my subscribers would bump up significantly: Working lots of hours is helpful to building a successful career! Calendly is great! Marx had some useful ideas! Our overreliance on Chinese manufacturing means we all are second-hand participants in modern day genocide! Clubhouse could still succeed! NFTs are just MLMs for tech elites! Web3 could be cool if it wasn’t primarily advocated for by a bunch of weird nerds! Etc., etc. Hopefully one of those will take and someone will hate-tweet me and I’ll make lots of money.

So if the negative coverage has been net positive for the business, why the name change? The answer is fairly simple—they were being sued over it by another company called Lambda Labs. While it could be that they are using the lawsuit as a convenient excuse for a rebrand, I’m not sure why you would willingly rebrand when you have one of the biggest names in your space—and it seems unlikely to me that a rebrand would fool anyone into thinking it was a different company. It seems clear and credible to me that this rebrand is because of a lawsuit.

While it is tempting to try to make Bloom into a clear-cut hero or villain, that just isn’t the way the world works. There are some incredibly admirable things about the business, and there is some stuff that just feels weird—most of which is endemic to the industry it operates in. 

I want it to work

In the strategy classic 7 Powers, Hamilton Helmer argues that one of the primary sources of a business’s power is what he calls ‘Process Power’. That is when a company builds superior long-term profits not through novel technology or competitive positioning, but by doing a unique combination of activities that are incredibly difficult to replicate. Think of how Tesla has given away many of its patents, but other companies still can’t seem to match Tesla’s electric vehicle capabilities. It is the unique combination of thousands of activities that allows Tesla to build a killer company. 

Bloom is very similar to that: There is nothing unique in the content they are teaching and nothing special about their financing options. Yet, if they can get every aspect of it working—from student acquisition to student placement—it is a powerful business that is incredibly difficult to replicate. Financing is the only portion of the business that Bloom has done differently, and it gave them 8% market share. What happens if they can get the rest of the machine working better? 

Our education system is broken. We can all agree on that. A school that only gets paid when its students get well-paying careers is a good thing. While the university experience at its best should produce elevated, thoughtful humans, at the very least it should produce graduates who can earn a comfortable living.

Without access to Bloom’s P&L, I can’t definitely build out the model to say that this company is good. But what I can say is that their idea is good. The more closely you can align incentives between companies’ success and customers' success, the quicker the flywheel turns, and the more meaningful the impact of the business. Bloom’s model is closer to alignment than most educational institutions’. 

It could be that coding bootcamps are just an example of those types of businesses that don't capture a ton of the value they create. There are many businesses whose value to society far outstrips the economic value they capture—book publishing, for example, or air travel, or even religion, depending upon your perspective. But it's clear that all of the leadership changes, the anemic growth, and the layoffs point to the fact that things simply aren't working quite right for Bloom as a business. 

So, with regard to the burning question of whether Austen Allred is a savior or an asshole, it’s clearly way too early to tell whether he’s a savior. As for whether he’s a bad guy, personally, I don’t feel qualified to determine his level of assholery. I can tell you one thing, though: he is very good at being very loud. On a podcast with Y Combinator, he said, “If there’s one thing I’m good at in life, it’s growing something quickly, building hype for something quickly. That’s kind of my superpower.” He’s right! And that is a lot of the job of an early-stage CEO. Really, the question of Austen’s persona isn’t overly crucial. If he can build a Bloom into an institution that lives up to the promises of its ideas, then perhaps much will be forgiven.


If you enjoyed the post, you can follow along for more analysis of technology businesses by clicking the button below. Your support means I am able to continue to publish high-quality, independent research.

Like this?
Become a subscriber.

Subscribe →

Or, learn more.

Thanks to our Sponsor: Axios HQ

Thanks again to our sponsor Axios HQ for today’s newsletter. Click here to see a sample newsletter written using Axios HQ and learn how your team can start communicating more effectively.

Read this next:

Napkin Math

The Case Against Sam Altman

The board is bad at their job, but they may not be crazy

2 Nov 21, 2023 by Evan Armstrong

Napkin Math

How I Use ChatGPT (As A Reasonable Person)

My AI awakening

8 Nov 30, 2023 by Evan Armstrong

Napkin Math

Live Fuller, Not Bigger

You must escape the shallow joy of more

10 Oct 26, 2023 by Evan Armstrong

Chain of Thought

🎧 ChatGPT for Radical Self-betterment

Clinical psychologist Dr. Gena Gorlin’s AI-powered annual review and goal-setting session

Jan 31, 2024 by Dan Shipper

Napkin Math

Profit, Power, and the Vision Pro

Will Apple’s new headset change everything?

5 Feb 6, 2024 by Evan Armstrong

Thanks for rating this post—join the conversation by commenting below.

Comments

You need to login before you can comment.
Don't have an account? Sign up!
@Erin.rose over 2 years ago

It's disingenuous to say that the ISA was the first bootcamp funding model to align incentives between schools and students. A number of schools had tuition reimbursement guarantees that refunded tuition in full if someone didn't land a job within a certain timeframe (usually 6 months) long before Lambda enter the game with the ISA model. You also didn't mention that Lambda sold some of their ISAs (which they had promised not to do) which means they actually weren't as aligned as they implied in their PR and marketing.

Evan Armstrong over 2 years ago

@Erin.rose Hey Erin! I think disingenuous is a little strong. Technically speaking, Purdue invented ISAs but I think it is fair to say that Bloom popularized them. You could argue that I'm giving them too much credit but the volume of schools that do it now versus didn't do it before bloom started makes me feel my assertion is fair.

I debated about including the ISA component but decided ultimately it wasn't the most important part of the story. The thing that matters is whether the students learn to code and get SWE jobs. which bank account those tuition payments land in is less crucial. However, you are totally right that contradicting public statements and private actions are very troubling.

Every smart person you know is reading this newsletter

Get one actionable essay a day on AI, tech, and personal development

Subscribe

Already a subscriber? Login