Time to read: 5 minutes

I gave myself 4 hours. Could I take an online resource and transform it into a dynamic report that tells you how risky your startup is? Yes.

I have taken some time off between startups and as I’ve written about before, I was super burnt out at the end of 2018. I’ve been getting back in shape, which you can follow along here and thinking a lot.

I haven’t come up with anything new really. Just small little projects to build to learn new skills and keep my chops on problem => solution process sharp.

I subscribe to a few beliefs in this entire business and startup world.

  • Only b2b
    • I know nothing about consumer and don’t pretend too.
  • Problem must be frequent
    • Willing to ignore some risks if daily
  • Problem must be painful
    • Willing to ignore some risks if so painful it MUST be solved
  • Willingness to pay must be at least $15/user/month
    • Ideally, $30/user/month or more
    • This can be a tough test depending on the market
  • No marginal cost per user
    • This rules things out like services businesses and youtube like products

In thinking about all of these rules and deciding what to do next I kept coming back to Leo Polovets article about de-risking your startup, read it here. This resource has been helpful in thinking about products worth building and iterating through a series of dumb ideas.

One disconnect I had between reading and thinking about the article he wrote was transferring that to a log I could come back too. I struggle to stay organized and often find myself with some halfway thought out ideas in a notebook, some in my iPhone and some in Asana. I wanted an easy way to just hit a url and answer questions and have it stored for me. This might sound inconsequential to others but it was worthy of a few hours.

So, I sat down and thought about how to build. I decided on my standard tech stack because I only had 4 hours and bought the domain.

The site itself looks super simple. Just fill out a form and get a report. Behind the scenes there is a hacked together system indicative of a true MVP.

I grabbed a jekyll template from online and made a few changes and typed website to deploy onto github pages ==> read more about this here: my custom deploy script

First step complete.

Next, I hooked up zapier to the google sheet the responses came in too. I used google sheets and zapier’s code option. This let me set a url to send data from the sheet too. I would eventually set up a formal api using AWS Route 53 + Elastic Load Balancer + Elastic Container Service but for local development I just spun up an ngrok tunnel.

Available via my bash command. Read more about this here


At this stage (about 2 hours in - there was sometime spent trying to figure out how to do all of this) I initially thought I’d use a google script and that turned out to be a massive pain in the ass - so I’d highly recommend zapier.

The next step was to write the service that would make all of the magic happen. I didn’t just want to know what answers I had clicked on the survey. I wanted to know things like:

  • How risky was my sales vs my funding?
  • How risky was my team vs product quality?
  • How risky was the entire venture based on what I thought I could do, current progress, etc?

So, to do this, I ended up writing an API that takes in all of the survey data results and scores them.

This takes an answer like:

  • You have unpaid pilots
  • You previously led a team that built a great product.
  • You have prior recruiting and management experience.

And scores each as a 3 in their respective categories. As in the blog post anything below a 3 is a concern. You should consider addressing via either a cofounder, investor or advisor. Anything 4+ was a major strength and has been significantly de-risked already. You should focus on other areas of the business.

After writing the logic of this (major headache but worth it) I now had the ability to take all of this insight and objectively score. Again, this is solely meant for me to think about side projects I’ve been working on and what risks I might be underestimating.

At this stage I had burned another 90 minutes. So I had 30 left. I have used sendgrid before and enjoyed the experience so I grabbed their sdk and imported it. I wrote up a super simple .format statement to pass in all my variables needed (in the most ugly and unformatted way ever) and had a way to deliver an annotated report.

This took it a step further from manually looking through and thinking to a true print out (if you like to hold things in your hands like me) where I could get an email with the things I needed to consider!

I had burned my 30 minutes left at this stage. So, I fibbed a tad and spent another 30 minutes getting everything set up on AWS and called it a day. I have a standard template I use that makes all of this super easy (docker, elb, route53 and ecs). It would have been impossible if I hadn’t done this with hundreds of similar small web services.

I also went back and added sentry to log any errors. I intend on sharing this publicly once Leo says it’s okay and wanted to know when all hell broke loose and reports didn’t get generated!

All in: About 4.5 hours. And a resource that (in theory) should be available 24/7 to help those thinking about new ventures or de-risking their current one.