Hey, I’m Colin! I help PMs and business leaders improve their technical skills through real-world case studies. For more, check out my live courses, Technical Foundations and AI Prototyping.
In early 2022, a friend of mine (we’ll call him Alex) approached me to work on a new product together. After running more than a dozen Airbnb’s himself, Alex had recently moved to a firm that professionally purchased and managed short term rentals. His role was to identify prospective properties, analyze potential cashflow, and ultimately determine which properties to purchase.
Alex had built a methodology for selecting properties and was creating free content on Youtube teaching others to do the same.
Our idea was simple: automate the data processing and provide clear step-by-step training.
The Plan
Here was our plan:
Our target persona was a multiple property owner who was looking to expand out-of-state. Ideally, they had experienced some of the challenges of property selection and understood that making the right decision could mean an additional $20k+ per year in profit.
We intentionally did not want to automatically recommend properties to purchase. Instead, we wanted to build a system that provided hosts the data and training they need to quickly assess a market. Although there were many competitors in the Airbnb data space, they all approached the problem from a pure SaaS perspective. They provided complex tools to manage data within their platform, but this required expertise and time from hosts.
Our product would be closer to a tech-enabled service. We would provide an extremely simple interface – select your city, generate the report, and get an export to Google Sheets. We would also provide detailed training and expert support from our team.
Distribution would be primarily through Youtube. Alex had a small following (~3K subscribers) in a very specific niche, so we felt confident we could convert a decent number of users. We priced each report at $1000 initially, with multiple reports bundled at a discounted price.

We expected to reach the $25k mark fairly quickly as we converted existing demand for Alex’s system, and saw an opportunity to scale well past $100k in 12 months time.
Our primary risk was the data itself. Airbnb does not make their data publicly available, so we were reliant on a 3rd party who had been collecting and reselling data.
The Build
As the technical partner on this project, it was up to me to build out the tooling and infrastructure needed. To start, we decided to only support the United States, as the vast majority of our demand was concentrated there.
Our data supplier provided us with 1.5M+ records every month with 65+ data elements per listing. We also had in-house data cleaning, validation, and aggregation steps to run per listing.
This was going to be a complex project.
Here’s how it worked:
Each month, our data provided dropped a 10GB CSV file in object storage (S3)
An automated data pipeline would pick up the file, clean it, and pre-compute values. Then, it would insert the records into a MongoDB database
Mongo would store all of proprietary property data, and a separate relational database would store all other data (users, accounts, reports they created, etc)
The user would navigate to the site, select a location, and generate a report. The application server would then grab the relevant data from Mongo, create Google Sheet, and insert the data. Then, we’d create a gSheet pivot table automatically to prepare the report.
Finally, we automatically added the user as an editor on the gSheet, which triggers an email to notify them the report was ready.
This backend of this product was fairly complex, but it was critical that we kept the User Experience simple.
We decided on a single page that linked out to your existing reports and allowed you to request new locations.
Scaling and security were fairly straightforward. We used GCP, Google’s cloud services. Our app server was hosted on App Engine, which allowed us to automatically scale servers based on demand. We ensured that all requests to our server had to be authenticated and from known URLs using whitelisting. Our Mongo environment was also hosted in GCP and was secured behind a private network – no one could access its contents directly.
Our S3 bucket was secured with a simple username and password, and we archived files monthly after processing.
How It Went
After spending ~4 months building, we launched an early access version to a few select customers who we thought would benefit most. We iterated on the product for about a month, then launched to a broader audience with a limited number of reports available.
Our sales far exceed our expectations, passing $100k gross in 2 months! 🎉
During this time, I was putting out a ton of technical fires (while working full time). I was working 12+ hours a day, fixing bugs between meetings, and generally just trying to keep the ship afloat. In case it’s not obvious, I’m not a professional engineer and built 100% of this product solo, so I constantly felt like I was making it up as I went.
Unfortunately, our success quickly dried up as we ran into a new problem – access to data. Our data provider was unable to continue working with us and other providers charged upwards of $240k per year for the same product. Changing providers would also require us to rework our data pipelines and transformations, creating weeks of new work.
We were also unsure if we’d be able to scale our go-to-market enough to maintain margins, given our new (very high) fixed costs.
Ultimately, we decided to shut down the product and Alex continued offering his product as service. At the same time, my live courses on Maven started to pick up and I decided to prioritize providing the best product I could to help other PMs become more technical.
What I’d Do Differently
Overall, building this product was one of the most fun experiences I’ve had in my career. Taking on the lead technical role was definitely a stretch for me, but we were able to make it work.
Having built 12+ products solo, here’s the 3 reasons why I think this one worked:
Clear distribution strategy - Alex had already spent 2+ years building trust and expertise in a very niche topic
Clear pain point - Multi-home hosts had experienced the pain of unprofitable properties and desperately wanted to avoid doing the same thing again
Simple product - We were able to deliver a clear, simple process that actually worked – something that was difficult to find on the market
If I was to do this again, I’d probably take the leap and purchase the data for $240k. I think we underestimated the size of this market and could have easily scaled this to $2-3M in revenue through channel partnerships with larger influencers in the space.
But, there’s always next time!
Did you enjoy this post? Make sure to hit subscribe for more!