Shovels May 2023 Newsletter
Here are a few more updates since April.
- Our final fundraising total is $1,071,939.39. I'm happy to connect with any additional interest but we'll ultimately push it to the next round.
- We hired a full-time data engineer and two summer interns.
- One of our interns made a chat AI interface to the Shovels API. We're going to incorporate it into a new website we're building this summer.
- We're thinking a lot about building and zoning laws. Aren't you?
- National data coverage is right around the corner.
That's a wrap! We set out to to raise $1M, pared it back to $500K, then raised the target to $750K and then back again to $1M. Fundraising is hard. It was hard for us too, but I think we did a few things right and caught a bit of good fortune.
I'll share a few learnings here about the fundraising process.
- I was told by a climate tech CEO who recently raised several million dollars that I should expect to pitch 100 VCs. She generously shared her VC list with me. I ended up pitching about 20 before finding a lead investor, and then about 20 more to fill out the round to $1M.
- Nothing beats customer interviews. I didn't know this at the time, but I refined my investor pitch each time I spoke to a prospect interested in building permit data. It helped me choose the right language, define the right need, and formulate a solution that makes sense.
- What makes sense to customers also makes sense to investors. It makes sense that these pitches should be quite similar.
- It helps to be unique. Investors don't hear a lot of pitches about building permit data, so Shovels doesn't sound like anything else. I can use that to capture their imaginations, and I've learned from customer interviews about the many uses that permit data can have in the pre- and post-construction phases. I can make Shovels sound big because it's true. This can be a VERY big business.
- Investors and customers love the Shovels brand. We lucked into that bit of inspiration (the first domain I bought for this project was ChooseAGC -- "choose a general contractor" -- not nearly as cool).
- And we got lucky that our lead investor bought in right away and has stood by us the rest of the fundraising journey.
I want to share a bit of high-level strategy about how we operate Shovels, which also worked well when fundraising. For anyone who knows me and reads my stuff, this strategy below should not be surprising.
There are two central operating principles in our 2023 budget:
- Get to breakeven as quickly as possible. This means staying lean for as long as possible and offsetting expenses with related revenue wherever possible.
- Convert investment into ARR. This ensures a good return for investors as well as founders. It’s our equity we’re selling, after all! If we raise $1M, we want to be at a $1M annual revenue run rate when we start fundraising the next round.
With these principles in mind, our financial projections are modeled with the following results:
- Total burn of about $500K. We want to spend no more than half of funds raised before reaching breakeven, so we’ll still have powder left for growth experiments.
- Peak burn of about $70K/mo. High fixed costs are salaries and data acquisition.
- We will be at breakeven by summer 2024.
Some of the assumptions built into this model are:
- We can limit churn to 1% per month and upsell at 0.5% per month
- Luka can get by with 1 data engineer to breakeven
- Ryan can get by with 2 junior sales or 1 senior sales person to breakeven
- Our ACV is at least $1,000/mo
- Our product works and excites our customers
- We can reach profitability by selling just our API products
I plan to leverage my background with Xenon Partners to execute on revenue while being sensitive to cash flow by utilizing SEO and high impact content on target social media networks for market development. I also expect that our product will be great and we’ll continue to get word-of-mouth referrals because my co-founder is a badass CTO.
Most investors liked, if not loved, this message. The ones that weren't impressed, I wouldn't want to work with anyway. It's an important filter for me so I was very open about it. Plus, it's true.
I did hire two summer interns. Luka got his data engineer (Petra is fantastic!) and I'm still experimenting with some sales strategies before hiring anyone full-time. In sum these are not insignificant expenses, but I can justify them 100% knowing that everyone at Shovels is working towards one goal right now: get revenue as fast as possible.
More thinking about AI
Continuing from last month's newsletter, I have a few updates to share.
- This summer we're looking at all the large language model (LLM) options and figuring out the right tech stack to use to process millions of permit records through it for permit classification and entity extraction (appliances, rooms, quantities, etc.)
- We found another niche in municipal building and planning codes. Someone (might as well be us!) needs to load them into a LLM and use the LLM to figure out what the building rules are: setbacks, height limits, density requirements. These rules can completely disrupt the economics of a large project and they're very difficult to find and update. AI to the rescue! I know we're not the only ones with our eyes on this problem. Solving it is a stretch goal for the summer.
- In general, I see good data as our competitive advantage in the coming AI wars. Try asking Bard or ChatGPT what permits have been issued in a given zip code. They can't do it, but we can.
We signed a deal today to get national data starting on June 1. We're working overtime to prep the data pipelines for this influx of data and we'll start reaching out to those of you who have been asking for this for months.