How to deliver a 95% placement rate with no fees and no teachers — top lessons from Microverse

Ariel Camus is the founder and CEO of Microverse, an online school for remote software developers. Launched in 2018, Microverse connects learners from more than 100 countries around the world through an intensive 10-month training program, scaling high-quality teaching using fully online peer-to-peer learning. Students only pay once they are employed, and 95% of graduates secure a job within 6 months of completing the course.

Here is what founders will get out of this conversation with Ariel:

• Microverse’s product playbook for delivering a highly scalable and impactful learning experience
• The methodology behind peer-to-peer learning and how to make this work in a fully online environment
• The practical implications of delivering education in emerging markets
• Ariel’s biggest leadership learnings as a founder and CEO

What’s the story behind founding Microverse?

There are so many challenges in the world right now — climate change, Covid, cancer, space exploration — that we need every single human being working to solve them. I was thinking about the people that I left behind in Argentina when I was a kid, and all the places I’ve travelled, and I realised how it just doesn’t make any sense that we waste so much human potential in the world with the massive challenges that we have to solve. What we’re trying to do with Microverse is come up with a model that enables all that wasted human potential at scale, so not only can we get them technical training but the training allows them to work internationally. We mostly focus on emerging and developing countries, and we focus on training people for remote jobs, where they can avoid the limitations of their local markets. And we are the only school today that offers an income share agreement available worldwide. You don’t pay anything upfront — you pay once you get a job based on a percentage of your salary.

What differentiates the learning experience at Microverse from other bootcamps?

For me, it’s not about teaching — teaching doesn’t matter. What matters is learning, and learning outcomes. Are the people who graduate getting jobs that they’re happy about? Are they getting paid what they deserve? Can they keep those jobs? Can they do awesome work?

Two key things differentiate us:

• Graduate outcomes: In the last quarter, we had a 95% employment rate within six months of finishing. And what’s most important, 70% of graduates are working internationally and 95% are remote.
• Peer-to-peer learning: In a job you’ll be expected to continue learning every single day and you’re not going to have a teacher — you’re going to have peers, colleagues, performance reviews. We believe that learning should feel like working and so we design the learning experience to look, feel and work like a job.

Can you quantify your impact so far, in terms of reach and accessibility?

We receive between 8,000 and 15,000 applicants per month, out of which we accept 60, which is less than 1%. We are looking for people who are already good communicators or team players who have an exceptional passion for making software engineering a career. You need a good level of English. You need to have stable internet access, stable power supply, good hardware — those are things that in the developed world we take for granted, but when you work with people in Latin America, in Africa and Southeast Asia, I realised how privileged I am to take those things for granted.

We have a 7-month long program, plus 3 months on average to get a job. Around 250 people have completed the program and got jobs so far. The average instruction score is 8.5 out of 10. 96% retention after six months of those jobs, with salary increases of 40% with their second jobs or second-year raises. The average salary increase from previous jobs before Microverse compared to after Microverse is 2.4x. The net promoter score constantly is around 70, but with the new program that we’ve launched, Microverse 2.0, we’re actually getting to 80.

The structure of your peer-to-peer learning program has lower pedagogical costs, so how does your gross margin compare to other bootcamps at your scale, and how does that enable you to improve impact for your end users?

To me, the margin is not about the business, it’s about the customer. We had to design a solution that works for people in emerging and developing countries. The fact that our program uses a peer-to-peer learning model with high margins is a consequence of knowing who we wanted to serve and who our customers were. It’s not just that we can help more people, which is part of the equation. It’s also about widening access to world-class education. If you need to charge people upfront, you’re leaving out so many people. The income share agreement and peer-to-peer learning allows us to get to more people and allows us to get the margins so that we can invest in any country or in training for underrepresented groups such as women.

Our margins are in the range of 70 to 75%. That includes customer acquisition costs. Some of the leading organisations in this space achieve around 25%. The main thing is that every time we grow to support more students, we don’t need to hire more teachers and we don’t need to create content, so this is way faster to scale and get into new types of fields. It is those margins that allow us to have the flexibility to take the risk to invest in people that we care about.

Can you explain your learning methodology in a bit more detail?

We don’t teach, we enable. There is no teaching in the same way there is no teaching in a job — there’s learning in a job. A teacher has three main roles, in our opinion:

• Communicating knowledge
• Accountability and motivation
• Guidance, to help you know if it’s time to ask for help and find out how to get it

Those are all things that we are able to replicate using peer-to-peer interaction, where students learn from and with each other.

The program is 8 to 10 hours a day of deep focus. All the deadlines are designed so that you can only meet them if you’re working full time, so it’s a very intense program. At the same time, it’s not completely self-paced learning. You have a morning session with the team, and then you get assigned a project. You need to build something like an internal product for customers or clients, where you get requirements from a product manager and specs from a designer. That’s how it works in a company and that’s exactly what you get at Microverse. Our project-based learning is not exactly the same as a project at an IT company. It’s a project that we have designed to scaffold learning through a series of goals and objectives. So you might start with HTML and CSS, and we’ll say: “Hey, this client has asked you to build this website. Here are the specs and requirements, and here are some resources that we have curated for you to guide your learning”. Sometimes it’s a solo project. Sometimes it’s collaborative, where you might do pair programming with another person or split the work. For the first project you work with another person and then for the second with four people, so that you learn how to work with larger and larger teams remotely. We have designed milestones, like company reviews, where someone will review your code and leave feedback, and you then have to introduce changes to the code and submit a new code for review. And they keep doing that until you get a thumbs up, so that you can move to the next project.

The thing is that a majority of people who do the code reviews are other students who are more advanced in the program. We only select around 10% of those who apply for this; they go through training, get paid, and have rubrics to follow to review material. They do quality assurance and performance reviews of each other, to make sure that the work they are all doing is always exceptional. And then for capstone projects at the end of the modules, you will often have more than one person reviewing your code, so there will be more statistical significance to the feedback you’re getting.

Around that, you’ll have your standup with the team at the end of the day, for more accountability, who also help you set SMART goals for the following day. Then you have your onboarding mentor when you join the school, just as you would in a company. Then there’s the part that we don’t do with peers where, towards the end of the program, we hire professional career coaches so you can get ready for interviews. But even there, the first reviews of your resume or your portfolio are done by peers who already have jobs or are further ahead in the program.

What are some of the pedagogical tactics that you’ve developed as an alternative to teachers?

I don’t know many places in the world where you have a teacher available to you all the time — the average time for our students to get an answer to a question is around 60 seconds. We accomplished that by building community tools, so that it is not just one person who can answer the question — rather everyone can contribute. It’s all about the culture and setting the right expectations among students.

We also focus on the skills to learn how to learn things for the rest of your life, in your career, fast. If a teacher teaches you things, you’re not going to continue learning in the best way possible, because you’re not used to independent study. We’re optimising the model to teach people how to learn, to empower people to learn how to work, so they can continue doing that for the rest of their lives.

The peer-to-peer aspect of your approach is obviously key. Could you touch on your approach to matching people?

There are cases where you could be pairing up with people with very different levels of experience or understanding, and that can be a wonderful experience for both of them. It depends on whether the person with more experience enjoys teaching other people, and if they perceive value in teaching other people. The kind of person who will join that is the person who understands that they learn the most when they have to teach other people. Now, you can get almost everyone, if you select the people well, to understand the value of doing that.

It gets tricky when you factor in the fact that, especially working from emerging and developing countries, they have limited time to finish the program often due to financial constraints. So if you make teaching others or empowering other people to learn optional, they will often not do it because they’re not going to graduate as fast as possible, which means they might run out of money before they get to the outcome they want. If you really want to empower people to help other people, you need to design the experience so that helping other people is part of it. You need to create activities where there are learning goals — such as transmitting and communicating to other students — that must be accomplished.

We get feedback from employers about this point. I was talking to an employer and they told me we only hire Microverse graduates — they have hired like five or six so far — who have been code reviewers, because they are really good at working with other people at all different levels and that has shown to make them work much better in a remote setting.

There’s a lot to plan for that to go well. With peer-to-peer learning, the quality of the experience is determined by the quality of the peers, that is the most important thing, so we’re extremely selective on the admissions side. It’s also, no matter how determined or smart this person is, if they don’t have a stable internet connection, that affects the capacity of another peer to complete a project. So you can’t compromise on who you let into the school. But even once someone is inside the school, you need to constantly monitor whether they’ll be able to perform at the level of commitment that is equal with the level of commitment expectations of the other peers. At times, that may not happen because as a person they might be facing financial issues, mental health issues, learning difficulties, family struggles, and all these things that break our heart to hear about. We accept that you have been doing extremely well to work through this. Today, however, we clearly have a problem. That’s okay. We ask you to leave school, take care of those things in your life, focus on what’s important, and once you’re ready to come back with full commitment and dedication we’ll accept that. But managing that, doing that well, and reacting to changes in that quality of the peers very quickly — that is the key to great peer-to-peer experience.

You’ve made a strategic decision not to invest in your own content, but to base everything you’re doing on open source. Has this been limiting in any way?

First of all, not creating content, or trying to avoid creating content, doesn’t mean that you don’t create a curriculum or you don’t do instructional design. We are very intentional in the instructional design process to create our curriculum. It’s a detailed design process, with insights from employers and market data that we use to build module by module. And the learning goals that are requirements are things that we filter for in admissions. We even designed learning as part of the admissions process, so that we can help people meet those goals.

The second thing is that we believe there are like 15 years of innovation when it comes to creating amazing content. You have Coursera, Udacity and MIT making amazing content available for free or very cheaply. So where we can, for a given learning goal for a given project, we will always look for the best possible existing content. It is incredibly unlikely that we would hire the best teacher for every single learning goal that we teach, but we can look for the best piece of content for that specific learning goal and integrate that into our curriculum. So for a lesson about positioning elements with HTML, for this project we are currently building websites for clients — we will need to identify the learning goal behind this and then link to an article or video that we have curated for you. Some people are visual learners, some prefer to listen, some prefer to read: we can accommodate that because we’re not limited by the crazy amount of work that it takes to create content and keep it up to date.

Not only that, the curriculum and all the references to external content are open source to all the students so they are constantly providing new material that they think is better. If you find a resource that you, the learner, think is awesome, you can do a pull request to suggest that we add that change to the curriculum— our learning team will then recognise that they think it is a better piece of content.

Are there any mistakes you have learned from building this product?

The first one is trying to help everyone. For a mission-driven company like ours, it’s really hard but you can’t try to help everyone. That is what many people call “idiot compassion”: you’re trying to be compassionate, but you end up hurting more people than if you had been honest from the beginning.

Mistake number two was being too distant from the reality of our customers. Coming from a developing country, Argentina, and having travelled to 80 countries around the world, I thought I understood the reality. But I’ve never had to walk one mile every day to get gasoline in these big containers and walk back home with the gas every day to keep generators running in order to stay in the program. I’ve had students dying from malaria while going through a program. It took me a while to really accept that I couldn’t change those things. And a big mistake was assuming that these people have predictability in their lives. Predictability is a massive privilege. If you know that you’re going to have a house and be healthy a year from today, you are a lucky person. When we first designed our program, we said we believe in the concept of mastery learning. Well, what happens when you give people time is people will take all that time, so people were taking 13, 14 months to complete the program instead of 9. A lot of those people couldn’t make it to the end of the program because they didn’t have the privilege of resources to support full time dedication for 13 or 14 months. So not understanding that was a key mistake that we are fixing with the new version of the product that was launched on the 31st of May and that we have been piloting for the past six months, which is working much better. We’ve added a lot of rules like self-paced learning with mechanisms to control avoiding them getting into rabbit holes and procrastinating and all those things.

The third mistake is underestimating how hard it is to accept and to deal with the long validation cycles of education, especially when you have a model based on deferred income. As an early-stage company, one of your advantages is that you can move fast. When your validation is based on outcomes, so whether people get jobs and keep them, it happens a year or a year and a half later — you can’t just rely on the outcomes to make decisions, it takes a lot of time. And also understanding that you’re not just dealing with information around code and design, you’re dealing with people, their lives, their goals, their feelings — that makes it even harder to achieve fast returns. Building a culture that is intentional will allow you to iterate faster in this complex environment.

To access more practical insights from founders like Ariel, sign up to our newsletter here.