Introducing Coding with CodeDay

Computers are our future, but most people still don't want to learn to code. Even in places which require CS classes, the majority of CS graduates are middle-class, white, and male.

CodeDay is a series of online and in-person events which target those who think "I don't want to learn to code."  They attend because we target existing interests ("I want to make art for games"), form teams, and pitch their own projects. Volunteers teach using one-on-one, just-in-time CS lessons. In the end, ~70% keep coding.

In the past this has not been possible because of the significant personalization required, but we've scaled nationally using custom apps to empower volunteers based on individual participant interest and real-time feedback.

Millions of already under-served and marginalized members of our society
are at risk of losing their ability to have a say in the future. CodeDay has the potential to change that.

Society has made great strides in providing access to Computer Science education. One example: since 2013, 39 US states have started requiring schools to offer CS classes. Many CS education nonprofits have also emerged (some have becoming household names).

Despite this, CS equity has fallen. The field is still overwhelmingly dominated by white men from upper-middle-class backgrounds. The percent of CS degrees granted to women has fallen by half a percentage point over the last decade (and by 7 points since a peak in 2003). For persons of color, recent trends are worse. The percentage of CS degrees to Black students, for example, fell by 5 points from 2005-2015.

The future of technology is the future of our world, but many people don't have a voice.

The reason most interventions have not made a difference so far is obvious to anyone who's been in a classroom: most people do not want to learn to code or see themselves in a tech career. CS classes and non-profit workshops haven't made a difference because most people simply aren't enrolling in them.

Changing perceptions requires new techniques, and the infrastructure to scale. This is the sole problem CodeDay is designed to solve.

CodeDay is a program of events designed to do two things: (1) get those with no interest in pursuing Computer Science to attend, and (2) change their mind.

CodeDay events are run multiple times a year in over 50 cities around the world, and online. People attend a CodeDay because it's marketed to a related interest, in a way which helps them imagine how they can contribute, currently:

• Artists: make art for games
• Musicians: score games
• Actors: do voiceovers for games
• Writers: write stories for games
• Designers: design an app UI

We promote the events through partnerships with over 160 schools, and target promotion to teachers of art, music, etc.

Once at the events, participants pitch ideas for apps or games of their own design, then make them in a weekend. We use "just-in-time learning," where volunteers help participants learn CS concepts as they're needed, so participants don't realize they're learning at all.

This would be difficult to scale with traditional resources, but by using individual-level participant interest and real-time feedback collected through our app, we're able to prepare volunteers for their unique audience and provide insights about when and how to engage.

3mo later, ~70% are still coding!

CodeDay serves those who do not traditionally pursue Computer Science education or tech jobs.

A majority (60-70%) of CodeDay participants identify as Black or Latinx, Women, and/or LGBTQIA. (By comparison, these groups make up a small minority of students pursuing a CS education, and a smaller minority of tech industry workers.)

30-40% come from a low-income household. (Many don't have a laptop! Through a generous network of alums and supporters who work in the tech industry, they can take a laptop home with them!)

Most participants are from outside of major tech cities. While in-person CodeDay events do happen in San Francisco and Seattle, our largest events are in cities like Phoenix, or Normal IL.

We've run 444 in-person CodeDays (and several virtual ones) for over 50,000 participants to date. For nearly that entire time, we've collected real-time and after-the-fact data about perceptions of coding, their experience at the event, whether they're getting the help they need. To measure outcomes, we follow up three months later to see whether they're still coding. We track this data across A/B tests.

Many alums of the program have become volunteers, helping us keep close to the communities we serve.

This challenge is focused, in part, on providing equitable access to good jobs and, quite honestly, the future of many jobs will involve technology. ("Software is eating the world" as one prominent venture capitalist put it.)

The software industry is not anything close to equitable, and there are no signs that it's going to change without radical action.

This poses two problems to the "Good Jobs" focus: (1) tech jobs are, by in large, not available to most of the population and (2) the jobs new technology creates are not accountable to most of the population (think: Airbnb).

Most education non-profits are focused on teaching how to code, but most people do not want to learn. Existing work in this area has mostly focused on small, incremental changes to curriculum to increase student retention, or different ways to promote CS workshops, clubs, and classes. We don't care what specifically CodeDay participants learn.

We've found that an extremely effective way to get people to pursue technology is for them to create a project of their own design, from start to finish, without anything that feels like "education".

To do that, we developed a new way of thinking about teaching which we call "just-in-time education:" identify which topics each individual needs to learn in order to progress, in real time, and teach it to them, in a way they'll easily understand.

Our program is powered by what we call "just-in-time education," a model where participants receive individual instruction as they need it, in a way they can understand.

Sound familiar? Tutors have been doing this for decades! The problem is that it's easy to do this in the small-scale, but until now, no one has been able to scale it.

We built a suite of custom tools to make this possible, centered around a custom event management system called Clear which we've been iterating on for 6 years.

Clear interoperates with tools like "Companion" -- our custom iOS and Android app -- let attendees share real-time feedback with Clear, websites like "Present" -- which customizes slide decks for local interests and opportunities -- and "John Peter" -- which checks in with attendees in our chat community. In fact, we've open-sourced over 100 projects which help power our events and unique system of learning, and we have ~50 "microservices" running at any time!

Nothing about any individual piece of technology is groundbreaking, but the suite we've created has enabled us to scale up to 10-12,000 participants a year with an average ratio of 1 volunteer to 30 participants (which is better than a teacher in a traditional classroom), and with dropping per-participant costs.

All of our technology is open source at https://github.com/srnd/ (our old name) or https://github.com/codeday/

We currently serve between 10,000-12,000 participants a year.

In one year, we expect to serve 25,000 participants a year.

In five years, we expect to serve 500,000 participants a year.

These numbers represent the number of participants we directly serve. They do not include the very real, but very hard to measure, follow-on effects in a community with a large number of participants who suddenly become interested in CS.

Within the next year, we'd like to leverage the international support we built up in running online CodeDays during COVID-19 to expand into Europe and extend our reach in Africa. We'd also like to expand our North America partnerships in several focus areas where we've seen past results -- rural communities, and organizations serving the children of migrant workers -- with an eye toward growing similar partnerships internationally.

Within the next 5 years, we'd like to continue to grow our presence, with a particular focus on under-served communities in Africa, Latin America, and India.

There are two major barriers we have identified:

1. While we have been successful in bootstrapping a sustainable business model, we have never received an individual large grant to invest in CodeDay's growth. This has made our growth significantly slower than we'd like, and to some extent our growth is coming from reducing costs with better technology, rather than increasing our funding.
2. Our team is largely from the US, and most of our volunteers are from the
   US and Canada, with only a small amount of overseas support in Kenya. We need to gain experience with different cultures to understand how our model can adapt internationally.

1. We've started applying for grants which are specifically targeted to one-time, start-up funding instead of sustainable funding to support a community.
2. We started running online programs as a result of COVID-19, which have brought in participants in 11 new countries, many of whom have expressed interest in helping start up local events. We're also looking to build up partnerships with local organizations who can assist in tailoring our programs, outreach strategy, teaching, applications, and algorithms.

Full-Time: 2
Part-Time: 8
Contractors: 5
Year-round volunteers: between 160-200

Tyler has a background in the world of technology, as the former founder of a Y Combinator and VC-backed technology company, as well as 7 years of experience in education as a teacher (through Microsoft's TEALS program) in Seattle.

As a whole, CodeDay has proven we can scale. As a bootstrapped nonprofit, we've grown from 1 city to more than 50. We've developed partnerships with schools, states, and other organizations to recruit participants without prior CS interest, and regional and international organizations in order to grow to new areas.

We've also proven we can execute: more than 50,000 participants have physically left their homes and spent a weekend making apps and games with us (to say nothing of online participants). Along the way, we've maintained our high outcome metrics. (And a 70+ Net Promoter Score!)

Because we generate a stream of people with a newfound interest in Computer Science, but don't provide the education follow-up, we partner with most national tech education nonprofits (such as Code.org, Hack Club, and TEALS). We also partner with many regional organizations and schools with an interest in advancing computer science (such as CodeVA or UAT).

Every company is now a technology company, which means tens of thousands of companies, from SMBs to enterprises, are competing for tech talent in traditionally small tech markets.

These companies fund CodeDay as a relatively low-cost way to build their local talent pipelines, and are our primary "customer" from a business model perspective.

Although many non-profits rely on purely charitable giving and individual contributions for funding, we don't believe this is a sustainable business model.

We are currently sustainable through the revenue generated from recurring partnerships with local SMBs and enterprises across the US. Developing local tech talent is a service which these companies find incredibly valuable, so we consider them our customers (although whether funding is legally considered a grant or revenue depends on the level to which we provide direct benefits).

We identified two challenges we'd need to meet in order to scale our programs up further:

• A need for a grant to invest in growth, rather than a focus on individual regions.
• A need to build partnerships to help us grow the market of students we're reaching.

MIT Solve provides an opportunity for funding and, more importantly, provides the opportunity to network with partners who already work with students around the world who could benefit from tech jobs (either for the economic mobility they provide, or the need to have their voices heard in tech).

Currently, we largely market to those with interests in the arts or literature, but we'd like to expand our partnerships to other organizations whose members are traditionally excluded from jobs in technology.

We're seeking partnerships from organizations and funders who care about marginalized students having a say in the future of technology. We are particularly interested in partnership with organizations who reach students without a prior interest in technology, in Africa, Latin America, or India. Organizations focused on general education accessibility, the arts, or community development would be great fits.

Providing equitable representation in high-tech jobs is more than a matter of economics, it's about giving everyone an equal say in our future. Right now, the voices of women are not being heard.

The people who build the products of the future are mostly male (among leading companies, between 70% to 80% of tech workers are male) and the needs of women are not being met:

• While 83% of women (18-29) report online harassment is a major problem, leading social networks failed to take even preliminary steps to reduce this until recently. Why? 60% more women than men consider this a problem.
• While two-thirds of women believe it's important to feel safe online, a minority of men do, so content policies are often comically unfavorable to women. One social network with notable harassment and hate speech problems was devoting its resources to removing photos containing female nipples.

Although women are over half the world's population, the technology industry is so male that products designed with the needs of women in mind have a special name: "femtech"

Unfortunately, education statistics show there's no big change coming; the number of women pursuing CS degrees as a percent has been declining for a decade.

CodeDay has proven to have the ability to change that! 42% of participants are first-time female coders, and 70% stick with coding after attending. Support from the Innovation for Women Prize would help reach millions of new woman coders through support new marketing and partnerships.

As industries continue to modernize, jobs increasingly require not only knowledge of how to use technology systems, but how to change them.

Knowing how to code no longer implies you work for Microsoft. In 2020 it might mean you manage inventory systems for a warehouse, optimize planograms for a supermarket, or customize accounting reports.

The digital literacy of tomorrow means learning to code.

Unfortunately, while plenty of classes are available, most people still do not want to learn. Why? Programs are marketed to those who already know they want to code, and do hope to work for Microsoft.

That excludes most people, and presents equity issues: right now, most people who know how to code are white, male, upper-middle class, and live in San Francisco or NYC.

CodeDay is a powerful solution to this problem. Our events aren't marketed to those who want to learn to code, they're marketed to those who don't, on the basis of other interests, like art, music, or storytelling. We work with students on an individual level to introduce coding concepts gradually, in a way that's relevant to their interests and a project they want to create.

Although participants say they're not interested in coding when they arrive, most keep coding! ~70% are still coding 3 months later.

We hope to grow CodeDay internationally: in the next year, we'd like to build an initial program in Europe and Africa and double our reach. The Gulbenkian Foundation's support would fund that expansion, including into Portugal.