When I returned to Canada in late April, the roads were muddy from the winter thaw and the chilly air would numb the hands while riding. Now, where you could see the hillside through the bare tree branches, your eyes are met with a thick green mass of leaves, flowers, and new buds. It’s amazing how much life was dormant and hibernating in the crisp, quiet winter, just waiting to spring forth when the season changes. Even better than the warm weather is a wave of vaccines, allowing the humans of Canada to cautiously step out of their shelters and flock to patios and stores.
Meaford, ON is my Walden Pond, where I’ve now completed two reflective quarantines ensconced in nature. This year I’ve been in Canada for around four months, which is the cumulative amount of time I would have spent here in 8 years on my pre-pandemic cadence. I’m grateful to have the quality time with family and to help out as much as I can while I’m here, though at times it has been tough. It has tested my ability to set boundaries, mediate heated situations, and let stress and anger flow through me (and given practice coaching my mom to do the same), while also having compassion for why and how these types of impulses and outbursts can occur. Things have improved since the winter, though they’re not yet (and might never be) in a sustainable place. Now that both my mom and brother have had their second shot I’m hopeful we’ll figure out how they can each have their own space.
I’ve also been reflecting on how a rather tumultuous environment might have impacted me growing up. I think it gave me better resilience against chaotic situations, because whatever happens in “normal” life feels pretty calm by comparison. I’m glad to be calm and stoic, but it might have also compressed the range of my emotional expressiveness. I actually think I had a good blend of independence, love, and stress when growing up, and I think I came out of the situation pretty well, (hopefully?) not continuing the cycle of trauma. I think I might be oversensitive and harsh against people who I perceive to have poor emotional awareness, and against people who lash out at others to cover up their own insecurities. In any case, I like how I turned out, and acknowledge that both the good and the bad played a role in making me who I am.
Aside from that, I’ve been riding about 80 miles/week and love the rolling country roads out here. I’ve taken countless photos of farm fields with some combination of dramatic machinery, clouds, and cyclists. I’ve even met some people in town! I got coffee with a new friend this morning, he’s 77, plays golf 4x/week, and rode across Canada for his 70th birthday. I’m also getting to know the nice people at the coffee, cheese, and ice cream shops. There is a bakery here with aged cheddar croissants that are to die for (yes, even by SF standards). Housemate recruiting has been a fair amount of work - we have lots of good candidates but I’m convinced we might have one of the hardest set of criteria around, though that’s a whole different topic to unpack. I miss the house and am excited to be back in my new room just down the hall from my old one.
Riding in the Blue Mountains, Ontario.
The startup journey has been going well. To use my previous terminology, I’m now in the “Running” phase; I’ve committed to a cofounder, settled on a domain and angle of attack, and am hustling to get users and test MVPs. It has been a while since I posted, so I thought I’d share the latest.
Cofounder dating generally means figuring out if you want to start a company together, and involves working on projects, talking about life, and generally figuring out if you’d be a good match in disposition, working style, skills, and goals. It definitely took some time. It’s scary, because having a cofounder massively improves the chances of success for the company, but getting it wrong is also one of the biggest company killers of early stage startups. You have to ignore your self-imposed deadlines to some extent, since it would be crazy to say yes to whoever you happen to be working with just because it’s late April and you said you’d have picked someone by May. Aside from trying to see if any of my friends were interested, I thought I’d create additional serendipity by joining some organizations where people are looking for cofounders.
Community 1: On Deck Founders
On Deck Founder Fellowship was okay, I’d give it a 5/10. I feel like I’m in the minority here since the fans are quite vocal on Twitter, but the ROI on my time felt very low. Part of this was because I was focusing on healthcare, and most on deck people were focused on B2B SaaS, creator economy or crypto. Another part was that there were many people hoping to break into tech from sales, finance, or consulting, and there were relatively few technical people to build with. This meant the few engineers were pretty exhausted from being pitched all the time, while the non-technical people would pitch you a (often underdeveloped) idea hoping you’d want to be the technical co-founder. More balance in this department would have really helped. Finally, I found myself craving people who would poke holes in my thinking, reframe the problems I was grappling with, etc., as would happen at Forward, but I couldn’t really find that kind of intellectual environment. I did officially “cofounder date” one person in this period, and we even applied to YC together, but I decided not to proceed after working together for a few weeks.
Community 2: On Deck Health
I signed up for “On Deck Health” afterwards, hoping to find a tribe grappling with similar problems. I’d give it a 7/10. The best part was having better access to payers, clinicians, providers, and others trying to build in the space. At Forward I was so accustomed to being a maverick outsider that I really didn’t know how some parts of the system worked, and I did learn a lot. There were surprising perspectives like, “Epic isn’t that bad and is a leader in interoperability” or “value-based care might actually be more expensive than fee for service”, and some unsurprising ones, like most of the industry has bad incentives, is extremely inefficient, and that everyone’s excited about care moving into the home. More of the advice was about how to navigate the bureaucracy of the healthcare system rather than how to structure the problem and dramatically improve things with technology. There were a few people I might have liked to work with, but I didn’t end up cofounder dating anyone from here.
Epic isn’t that bad and is a leader in interoperability
- Someone from On Deck Health
Community 3: South Park Commons
I prefer South Park Commons to On Deck, and I’d overall give it a 9/10. It’s a smaller and tighter community. With ~100 active people, it’s more likely that people will respond on Slack and want to help. On average the members are more experienced, and usually have a bunch of great advice, feedback to share, and intros that they’d be happy to make. It has ended up feeling like a second home (in the cloud), and it has been rewarding to contribute and be a part of the community.
Zain, one of the members who came in through the SPC fellowship, was dabbling in similar problem areas, and we slowly went from occasional conversations to cofounder dating. After about six weeks of hacking, sharing values and goals for the company, and getting to know each other, we decided we made a great team and and committed to working together. I feel incredibly lucky to be working with him, and I’m excited to keep building together in the years to come. I’m happy to say the work visa is now in progress too, and with luck I’ll be back in SF again later this summer.
Previous Focus: Trial Matching
The main focus for April and May was figuring out how to help clinical trials recruit more efficiently. Clinical trials take a lot of time (years) and expense (100s of millions of dollars down the line), and are a key bottleneck for getting life-saving medications on the market. Finding patients is a major rate-limiting step, and we discovered that a big part of the problem is a lack of awareness. Patients generally don’t know what trials are available, and doctors generally don’t have time to stay on top of the latest, and have a major disincentive for referring to trials (they’ll lose the patient and the associated revenue for their hospital system). The most common way for people to discover trials today is if they happen to talk to a specialist who’s also an investigator in a particular trial, which is like two needles finding each other in a haystack.
Trials need people, and people want to know their full set of treatment options - maybe we could bridge the gap by going directly to patients directly? There are many companies working on this because it’s such a big problem, and the reward for finding even one eligible patient can exceed $10000. We thought the existing solutions were too focused on patients who were already looking for trials; our approach was to target people earlier in the patient journey, and guide them from a recent diagnosis to a full understanding of their options (which would include trials alongside other, more conventional options).
We built a basic site to test this hypothesis, which allowed people to enter their diagnosis, see a list of therapies, and assess safety, efficacy, and risk for each. If one of the therapies seemed like a good option, they could get in touch and we’d help connect them with the trial. We bought some ads to drive traffic. We talked to about two dozen users, with diseases ranging from cancer to IBS to heart disease, and while they were keen to play a more active role in their treatment, it was hard for them to grok the pros and cons of new drugs. It was also extremely humbling to hear about each of their struggles, and motivating to build something to help them. Healthcare has massive disparities between coastal cities and rural settings - there are 10 counties in the US that have lifespan below 70 - and I really think technology will drive down the cost of much needed care and help people across the country (and world).
The landing page for our trial matching site.
After driving some traffic to the site via Facebook and Google ads, we found that conversion rate from visitors → people who were interested in a specific drug was very low, and of those who were interested in a trial, we weren’t able to get any signed up. Trials are meant for getting clean data for researchers, not to give people access to new, lifesaving medications. As a result, they have restricted enrollment windows, strict enrollment criteria, and availability in a very limited set of hospitals. One person was interested in a trial investigating weight loss for diabetics, but their BMI was two points too high. Another potential patient was located in Pittsburgh, were there weren’t any trial sites for the trials they wanted.
I’m sure we could have optimized the funnel and solved the education piece with a higher touch “sales” process, but in the end the trial supply just wasn’t there. We couldn’t convince ourselves we could add value to the trial matching problem, so we decided to take a step back.
Current Focus: D2C Self-Experimentation Platform
It’s clear that trials are going to look completely different in 20 years from now: the trial will naturally integrate with the patients’ lives, the trial will be a great experience that patients want to do, and the entire process will be automated. A global patient population will help generate evidence through trials that are 100x faster and 1000x cheaper than today. I don’t believe that the existing players will get us there, because they sell to pharma, which is an extremely risk-averse customer. If you’re spending $100M on a trial, saving $10M isn’t very tempting if you think it might end up getting the results of the whole trial thrown out!
A disruptive change is more likely to come from the outside. Our newest north star is to build the future of clinical trials today, in a less regulated space, with a business model that allows us room to experiment and iterate quickly. This means investigating less regulated, low-downside interventions like diet, exercise, and supplements, solving this problem in a way that’s tangible and useful for consumers, and having them pay us directly so we have the incentive to build a system that’s well-designed and easy to use, provides them with significant value, and low cost. If we can build something valuable for one person, maybe we can scale it up and provide value for millions (billions?).
Bottom-up, many people we’ve spoken with have clearly expressed that they want to know how something they’re doing is impacting their health, but it’s hard to do in a way that provides conclusive answers. This is the atomic unit of systemic health improvement; test something, keep doing it if it works, and discard it if it doesn’t. Your progress will snowball. Why do we need to test — can’t we just eat our veggies and exercise more and be done with it? To some extent, yes, but health is personal, and what works for one person won’t necessarily work for someone else. Drugs are approved for populations, not individuals. Diets are personal. Autoimmune diseases resist generic solutions. And there are many new things to try, from nootropics and new supplements to new diets, that don’t have consensus support, but might have a lot of upside for you. Plus, measuring stuff and seeing progress is fun and helps motivate you to keep going.
With a bit of structure, I think we can guide individuals to run a fairly rigorous “n of 1” trial on themselves. You just need a few key ingredients:
Outcome metrics that meaningfully approximate the desired outcome, and can be measured frequently enough to provide trends and statistical power throughout the experiment
A protocol whereby only the target intervention is changed over time (and confounders are logged to help control for their relative effect)
An analysis that tries to answer the question “how much did x cause a change in y?”
Many people will log data and hope for the best, but lack the tools to produce the right data, and to analyze it once they have it. These experiments can actually produce really good data - they’re interventional, rather than observational, and if you set up the experiment well (and the person sticks to it), you can answer a specific hypothesis with some authority. Our knowledge will only improve as more people use the platform too, and we can provide stronger data with more personalization as we scale.
We’ve been testing the impact of melatonin on sleep, intermittent fasting on productivity, and coffee on focus. Here are the results from an experiment I ran on myself looking at the impact of coffee, where I actually found very little benefit in terms of focus and productivity. Good to know! We already have one paying customer(!), we’ve had one customer referral, and people seem to be sticking around, all with a somewhat ad-hoc system I built in Coda. Yes, we’re also working on building an app to make things more robust and usable.
Some results from my “Impact of Coffee on Productivity” self-experiment.
Please let me know if you’d be interested in improving your health - all you need is some desire to sleep better, feel better, be more productive, or improve some chronic symptoms you’re feeling, and I bet we can help by nudging your lifestyle in a better direction, and proving to you whether the changes are working or not. We’re going to keep building, test a few more initial niches, and see if people retain and tell their friends about it. Results are promising so far.
We’re also going to start sending out more regular updates about company progress and things we’ve learned. It’ll be on a different mailing list from this one, so if you’d like to be added please reply to this email.
As always, thanks for reading!