Quoting Sam Altman: “You’ll get more support on a hard, important project, than a derivative one”
Rx Assist was initially built to help Canadian pharmacists react to drug shortages. During my time working as a part-time pharmacy assistant in high school, I noticed that every pharmacist I’ve worked with struggled with this one thing daily.
Since 99% of pharmacists nationwide could point at this problem and say, “Oh, that’s definitely an issue,” I received an immense amount of support. I’ve had colleges/associations discuss my questions in internal forums, grants to fund my startup, pharmacists who voluntarily built algorithms on my platform, and more.
I think it was the support that made my work feel unexplainably meaningful and real.
It is very hard to change user behaviour
My COO and I, at first, had high aspirations to have pharmacists perform more therapeutic substitutions than they currently already are – only 50% of the possible substitutions are being performed today.
Yet, after releasing my first MVP[1] to the hands of 15 pharmacists, I quickly realized that it is very hard to change user behaviour. I’d say this fundamentally applies to most startups, but what made it even more obvious, for us particularly, was the fact that pharmacists are not incentivized to increase the number of substitutions they make – or less incentivized than I expected at least.
Since staff pharmacists are usually the ones who handle substitutions, they weren’t incentivized by the increased revenue that more substitutions would bring.
Further, replacing the other ways of reacting to a drug shortage – ordering more stock, calling other pharmacies, contacting the original prescriber for changes – would be a behaviour that’s even more difficult to change, especially if that’s what the pharmacist is already used to.
The two best ways of knowing if you’re creating something people actually want
1. Getting a prototype in front of them and seeing their reactions
2. Asking if they would pay for it
It’s much better to have a small number of people who LOVE your product than millions who sorta like it.
I think it was Brian Chesky who said this, but when you do things that don't scale, create a product that a small number of people absolutely love, maybe they’ll tell everyone about it, and you might end up with millions of people who are your ideal customers.
I was a lot more mission-oriented than I thought
Because I had a more personal connection to the problem I was trying to solve, I didn’t realize how much my passion would be affected by making a pivot to solve a different problem for a different group of people. After making a pivot [2], I realized that I was working on a problem that I wouldn’t be morally content with, even if I did solve it.
Momentum is everything
The only thing I regret from this journey is not making the decision to stop working on the startup quickly enough. There were many early hints dating back to December, but due to my lack of passion, my work ethic faltered, causing me to reach this conclusion more slowly. Something I do wonder to this day is whether I would have had more momentum if I had a cofounder at the time.
Closing thoughts
Starting a startup is less scary than I expected it to be. At the end of the day, it comes down to building something people love and spending less than you make. Plus, it was kind of fun driving around to 70 different pharmacies and chatting with as many pharmacists as possible.
The worst case scenario of working on a startup is that it fails and you get a job using what you’ve learnt. I’d say only 10% of the total knowledge that I have on the pharmaceutical industry today is from before working on this. (I tried my best to keep my more technical learnings out of this essay, which I’ll post another day)
Notes
[1] The first MVP was a clinical decision-support tool that held decision trees for therapeutic substitutions, lists of current ongoing/anticipated drug shortages, and a documentation tool.
[2] The pivot was drug shortage prediction using machine learning.