- In reality, you will control certain inputs that influence revenue: salespeople (note you are the first salesperson), marketing spend across various channels, etc. A simple subset of these inputs should be inputs to the model – numbers you can change that automatically change the projections.
- Some measures of effectiveness (e.g. leads generated per $ spent on AdWords, $ sales per salesperson per time period, etc.) will be observable, and will be items you’ll try to optimize, but are not so easy to control as the inputs described in #1. You should also call out these measures as discrete assumptions the model. That way, you can easily change these assumptions to see what impact an increase in salesperson productivity might have, for example. You will also modify these assumptions as you get more data in the real world about your effectiveness.
- For your first sales, you may input the revenue and support costs manually if you can guess at them. The same manual input process might be appropriate for various line items – if you know the cost of something, don’t create a formula to calculate it, just input it. But you should also layer in the ability to modify assumptions for revenue on top of what you know and/or further into the future (e.g. 6-12 months out and longer).
These are fundamental principles of good projections. If followed, they should give you a model that can “live” with you as you grow the business. You can change the inputs and assumptions as you learn more about them to get better projections over time. You can easily change those inputs to see different scenarios for growth, and share the model with investors who can then easily review additional scenarios.
Much has been written about financial modeling for startups. Some of the best and most practical content can be found on foresight and was created by Taylor Davidson.