Indie hacking a nonprofit
Can you indie hack a nonprofit by shipping software products like madman? This is a story of how Herizon ships software MVPs.
I'm a startup hustler who dreamed about a nonprofit but didn't realize how draining it is to depend on public sector funding. The biggest problem is, that there's no instrument to fund "€1000 for one employed immigrant", so you're forever stuck in figuring out projects that rarely are measured by employment.
Therefore, in 2024 we're trying out something else. Can you indie hack a nonprofit by shipping software products like madman?
This is a story of how Herizon ships software MVPs.
One of our latest software products is Job Application Assistant.
You can try it out by investing an affordable €5 for one month, but you can also access screenshots by reading our blog post about Job Application Assistant.
For Behind the Scenes today, I'll show you how these ideas become reality at Herizon.
Problem discovery
At this point, we understand quite well the fundamental problems of immigrant employment. One part of it is too generic, too long, or irrelevant job applications no one even reads - and to be clear, this isn't only an immigrant issue, but a general with people searching for a job. Skills to improve one's employment are not being taught in any school.
Currently, job application improvement is mainly done by unemployment offices and by manual labor, so why not test scaling it with a software solution?
MVP development
Job Application Assistant was discussed internally in April after I posted a post on LinkedIn that we're searching for an Account Manager. I didn't receive any relevant applications, so we decided to build a lightweight MVP for 2 usual use cases: a formal job ad and a less formal job ad.
How we usually work, is that our CTO discovers a problem-solution from the general discussions on Slack or our weekly. Then he builds a quick Figma prototype and finally jumps in a huddle with a person who has the domain expertise in the field - in this case, it was our Head of Talent.
It took a week to have a working demo for internal testing for this very simple tool.
Our stack is very simple. We use Figma for prototyping, Heroku for hosting, and Github Copilot to be able to build products for a full software development team by one CTO. The cost of our stack is $400 /month + one full-stack developer.
Monetization
Useful software needs to make money. Otherwise, it's not useful and one shall kill it, so a high priority of ours is to monetize our software products immediately.
We tried out 2 methods: B2G and B2C.
B2G learnings: Cities are relatively interested, but the purchase decision is currently on hold due to OpenAI. The optimal case for the city is that we could somehow anonymize user's data, so the platform doesn't send anything sensitive to a US-based company.
We tried to do this and technically it's possible to train a European LLM model to anonymize the user's input before it's submitted to OpenAI's LLM model, but the value of the deal is €1000 /year, so...
We're currently focusing on finding a more cost-efficient workaround or waiting for cities to implement an AI policy so the decision-makers can make decisions like this.
B2C learnings: Immediately when we identified B2G will having a major purchase barrier related to AI, we decided to publish a B2C version. We simply implemented a no-code purchase page of Stripe and there it was - a fully monetized software product.
Even Job Application Assistant is a super simple product without a proper landing page, we started getting conversions from our community.
Marketing and scaling
So, with our Job Application Assistant, we have:
- Identified a problem and solution
- Built an MVP for the solution in weeks
- Launched and monetized it
The next stage will be marketing and scaling, so I built a landing page for marketing:
The volume of traffic is still relatively low, but steadily increasing and gaining some organic traction. We've limited resources for marketing, so our game is slow and steady, but the team is here for the long term, so luckily we've time.
Until next time! Let us know, did you liked the post. 🙏
