Before I launched my startup, I built an elaborate Excel spreadsheet to forecast how our company would grow. The spreadsheet had tabs for headcount, revenue by product, gross margin, rent, and cashflow.
At the heart of the model, was the product revenue model. Since we were launching multiple products per year, this was the most important part.
However, building this model wasn’t easy. The typical product in my industry would take six months development time and there would be a 50% chance you could release it on the first iteration.
So, the question was what should I do? Should I assume that a product could be introduced without any rework? And, if there was rework, how long would the rework take?
You use your best estimates to guess at revenue.
You get the idea. There was a lot of guesswork in building our initial model, just like there will be for you. But, you’ll end up using your experience to guide you.
I decided to build our model on the basic assumption that every other product would introduce on “first silicon”. The other half of our products, I assumed would need to an extra 3 months of engineering time to be reworked.
I also knew that the typical product in our industry required one design engineer to complete. So, I could now model how quickly we were going to grow based on how many design engineers we could hire.
The rest of the model flowed from there. There were ratios for how many marketing people we needed per design engineer, how many test engineers we needed per design engineer, and so forth.
You adjust your model once you start getting real data.
I used the excel model to build our initial revenue forecast. In fact, I used the excel model to build out our income statement, balance sheet, and cash flow statement.
And, I cut and pasted a simplified income statement from the excel model into the financial slide on our pitch deck. It was good enough for us to raise our initial $12 million in funding.
I started seeing the flaws in the model once we got started. For example, some projects took considerably longer than six months. Revenue ramped slightly slower than expected.
And then you just keep tweaking your model forever.
However, I would regard these as tweaks, so the model was still quite useful and accurate. That’s all you’re looking for.
There’s an old saying that most startups take twice as long and twice as much money to be successful as the founders believed. If your model is more accurate than that, then the model has done the job.
Then, as I said, you just keep getting better and more realistic about your forecasting.