Chapter 1: Engineering a Problem
The University of Waterloo is a funny place. Visually unremarkable, it sits in a quiet corner of Southwestern Ontario known as the Saint Lawrence Lowlands¹. The region might best be known for its delicious produce (a product of remarkably fertile soil) were it not for a handful of surprisingly well-known universities. The graduates of these universities will be so traumatized by bitter Waterloo winters that, immediately upon graduation they will flock south for the winter, clutching their technical degrees². It is fitting then, that the official bird and mascot of the University of Waterloo is the migratory Canada Goose³.
Everyone has their preferred metaphor for university; the administrators, a womb; the students, a pressure cooker; the skeptics, a printing press. I prefer to think of it as a funnel with holes in it. It’s probably better than nothing at all, but it’s still a terrible funnel⁴.
One of the universally defining features of an engineering curriculum is the explicit instruction to solve problems. Personifying the post enlightenment compulsion for technological progress, engineering as a profession prefers to expand outwards in the quest for new problems over modest inward-looking improvements. Driven by the lump of labour fallacy, each new generation of engineer will boldly claim (and then succeed) at curing the ills of their forebears. This autonomous, manic, and decidedly canine obsession to pursue their generations ripest problems is simultaneously the engineer’s most endearing and terrifying quality.
The tool of choice for engineers is a framework known as the “Engineering Design Process”. The importance of this framework in forming the minds of young engineers cannot be overstated, both in what it contains and what it lacks. The process is quite common sense: given a problem, define an improvement, and then iterate until the success criteria are reached. The framework provides little input on its far more interesting prerequisite: how to choose a problem in the first place. As it turns out, this is the only thing that matters.
The best sorts of problems are the ones you experience yourself.
This is common advice when searching for a problem. Herein lies the issue — although the student life is hardly gaudy, most Canadian undergraduates live a fairly privileged and refined lifestyle. As might be expected, the engineering projects these students choose to work on are often whimsical and extraneous. So what was the problem that we would choose to to spend six months and as many thousands of dollars developing a solution for? The Booster Juice line in the student centre seemed a bit long.
Chapter 2: Paradoxes in Impact Entrepreneurship
It was early 2019 when I became seriously involved in a pitch competition known as the Hult Prize. The Hult Prize is a global social entrepreneurship competition with the mission to build businesses which benefit both society and shareholders. The social entrepreneurship story is a compelling one: over the next twenty five years some $50 trillion in wealth will be transferred from baby boomers to Generation X, and this socially conscious generation will only purchase from businesses which also benefit society. As I would soon learn, there are many layers to the social entrepreneurship onion, some more palatable than others.
There are many flavours of social entrepreneurship businesses model, each varying in their generosity and effectiveness. Some operate essentially as charities, with their social enterprise branding being the only competitive advantage. These businesses are destined to fail, as their executives are forced to walk the line between impact and profitability and can never push unidirectionally for excellence. Of course there are exceptions (Patagonia) but these businesses are usually extraorindary in other dimensions.
This is the inherent paradox in social enterprise: businesses are economic vehicles, and there are economic tradeoffs between profitability and social impact. Polluting the river is cheaper than recycling chemical waste. Beyond this, there seems to be an entropy driven hysteresis in many social and environmental problems. Cleaning up the river is more costly than the profit derived from polluting it. Exploiting a people is more profitable than uplifting them.
There are certain complementary pairs of problems and business models for which this tension does not exist. One such model is marketplaces, where the health of the marketplace (measured in size and liquidity) is generally correlated with buyer upside. Moreover, efficiencies of scale can offer vast performance improvements to marketplace matching problem.
The Hult Prize has an annual challenge (which I was thankful for, having recently exhausted my idea reservoir with the smoothie bar masterpiece). The 2019 challenge surrounded youth unemployment and tasked each startup with the audacious goal of employing 10,000 youth in ten years. To put this into perspective, AirBnb (one of the fastest growing technology companies ever) has grown to 6,000 employees in a decade. Our interpretation involved a loophole in the definition of “employ”, since generating a unit of employment with another business offers the same benefit to society as generating that unit yourself. Using this looser definition, the AirBnb marketplace “employs” 650,000 hosts. Even then, generating employment is a long and difficult process. No matter the company, candidates must apply, be selected, interview, get hired, and then ramp up. But people are fired all the time from work, very quickly, and for a variety of dumb reasons⁵. The feedback loop between employed and unemployed creates a bistable system, tilting heavily towards unemployed. The problem then becomes preventing 10,000 units of employment from disappearing, a distant but now visible target. And one of the biggest drivers of unnecessary employment loss in the United States? People stuck in jail because they can’t afford their cash bail.
Chapter 3: Cash Bail and Broken Models
The idea for money bail dates back at least 1500 years in England, with commercial bondsmen popping up shortly after. Interestingly, England identified bail and sureties as ineffective and started moving to other forms of ensuring appearance as early as 1900, whereas the United States would experience an explosion of bail bond shops around the same time. It should be mentioned that there is nothing inherently predatory about bail bonds. They are a financial product, a specific form of financing for those who can’t afford to cover the entire bail upfront. Like all financial products, they can be designed to be predatory or fair depending on the lender⁶. In practice, however, the bail bond industry is extraordinary immoral, extracting massive profits from a system propped up by aggressive lobbying against criminal justice reform. Nowadays, bail bondsmen don’t control much of anything. The rates (10–20%) are often fixed by the state, removing any competition from driving down rates, and massive insurance companies underwrite every bond, leaving little profit for the bond shops themselves. Approximately $20 billion in bail bonds are written annually in the United States, and bail bonds are one of the largest taxes on poor communities in the country.
Before designing a business, we first made the ever-important naming: Better Bail For America. BB4A was crowdfunding platform and marketplace for bail bond financing. The idea was to provide an online platform where bail money could be easily crowdfunded for court payments, with the crowdfunding signals used in risk assessment to offer more affordable financing. Our hypothesis was simple: individuals who can get numerous people to donate to a bail campaign are more likely to appear in court. There are a number of problems with this business model, both practical (court payments usually have to be made in person) and ethical (risk assessment infringes on the presumption of innocence), but like any strong-headed startup we were plunged ahead, shielded by a vivid reality distortion field.
Chapter 4: Jails to Castles and other Pivots
In August of 2019 we ended up in a castle with forty teams from around the world. Guatemala, Mexico, Palestine, Jordan, UAE, Kazakstan, Kenya, Nigeria, Indonesia, Japan, Australia represents a shortlist of a just few of the countries present. There aren’t adjectives expressive enough to do this massive exposure of new people and ideas justice. I am permanently left with a deeper appreciation for the multitude of perspectives through which people view the world. To put it simply, why think outside the box when you can just ask someone what’s inside their box?
Most startups have an elephant in the room. The key is to avoid this topic when discussing externally, but obsess over it internally. BB4A had one such dirty secret: the business model didn’t work.
There was one evening where we wrote every single problem with the business on a whiteboard, and labelled them either a one, two, or three. Category three problems were real, but manageable. Solving a P3 would usually take one week. Category two problems were more serious, and could take months to solve. A business could usually only have one or two P2 problems. A category one problem guarantees the death of the business if not resolved, which could take years. BB4A had six things in this first category.
So we pivoted.
The Hult Prize Accelerator is a five week program, with a mini pitch competition at the end of every week. The pitch competitions start Friday morning. We decided to drop everything and start over Thursday at 4PM.
Our new idea was Phonic Research, a platform where youth conducted face-to-face interviews on behalf of companies looking for high quality and cost effective market research. The model combined the inexpensive, distributed nature of quantitative research with the emotional, open-ended nature of qualitative research. What’s more, audio surveys capture significantly more information than plain text surveys. As anyone who has spoken with a robo-caller knows, tone of voice is a much more information dense communication mode than plain text.
Startup investing is highly risky and speculative, which is why trajectory is more important than cumulative work. This is the only explanation for what happened next. In just one week of work, we built an MVP, landed several demos with large consumer brands and ad agencies, and had our first LOIs. This was farther than BB4A had ever progressed, and was enough traction for us to gain the privilege of pitching at the United Nations.
Despite a teasing speech by former President Bill Clinton, in which he alluded to tech and AI as one of the most significant challenges to the next generation, Phonic did not win the $1 million dollar prize⁷.
Chapter 5: Building a Real Business
Just as the chaos was subsiding, 2019 ended with another surprise — an acceptance into another accelerator⁸. I’ll write about this experience at length, but in the near term it meant more hard work and the opportunity to scale Phonic.
Utopian maxims aside, technology businesses are about one thing: compound growth. If the theme of 2019 was exposure, then 2020 is the year of growth.
Let’s get after it.
 Waterloo is actually not in a corner, quite the opposite, sitting centrally in the southern tail of Ontario. There is something irresistibly quaint about corners.
 Mark Twain is attributed with saying “The coldest winter I ever spent was a summer in San Francisco.” Waterloo graduates would say that Mark Twain is full of shit.
 It is not the official mascot.
 This is the kind of imagery only a dissatisfied alum could conceive of.
 It was important to focus on the unnecessary reasons people lose their jobs, since this was fulfilled the social mission of the company.
 Bail abolitionists would argue that cash bail is a violation of the presumption of innocence, and that any form of money bail is a regressive tax on the poor. I tend to agree with them.
 The award went to Rutopia, an indigenous tourism business in Central America. The team could not be more deserving.
 I will hold off on disclosing the accelerator and funding until Phonic publicly launches our new platform.