Financial foresight for a new project

Greetings,
with a friend of mine, we will be applying to a governmental program that financially assists young entrepreneurs to start their own startup company.

Our project is a form of product recommendation system where site owners embedd our system to their sites and we will be sharing the revenue of the advertisement with our users. We plan to go live after 6 months of software and business development concurrently. I am well aware this sounds like Google Adwords but it is actually different.

Both of us are computer scientists thus we have a rather weak financial background. For the product we want to develop, we have been asked to do a rough estimation of 2 years of annual revenue of the company. We are having really hard time to make such predictions.

• How can we estimate the growth of site owners that embedd our system and the amount successful advertisement.
• How can we estimate/guess the base users at the begining of the project once it goes live.
• What percentage of the revenue should be used on the marketing of the system.
• Is there any statistical data regarding a such system as ours, where we can use to do such calculations.

I will be grateful for any suggestion or any way to do such financial foresight for a brand new project.

Firstly, I do not know of any available data that would be a relevant.

Your financial projections at this early stage are going to be based on part guess, part estimate. Once you have actual data you can refine your projections, but that doesn't help you very much now.

The benefit comes from seeing how the numbers interact over a number of different scenarios. You end up with a range of different financial forecasts and ultimately it a case of using your best judgement as to how likely each one is.

Bear in mind the purpose of the projections. If they are for the purpose of applying for a government grant you might think about what they want the numbers to look like. Not much point adopting such a pessimistic outlook that you are rejected immediately. Not suggesting you lie, but within the range of possible outcomes there will be some outcomes that are better than others.

• Start with what you do know, even if it just the relationship between variables. eg if we have x users our bandwidth will cost x * y dollars.
• Plug in some estimates that seem reasonable as a starting point.
• Assume that adoption by customers will start off slow and then increase, rather than evenly over time. eg if you expect 6,000 users over 12 months you might get 50 in month 1 and 1,000 in month 12. You won't get 500 every month in a even distribution.
• once you have a model that seems to hang together ok you can start tweaking the inputs to see what happens in different scenarios.
• Let someone else sanity check your model. They will see things that you have missed and be able to question your assumptions.
• You will probably need to invest in marketing before you get much in the way of revenue, so initially the percentage of marketing to revenue will be much more than 100%. Over time that will decrease as revenue increases.
• A common technique is to do a sensitivity analysis - model what happens if key inputs are 20% higher or lower than a base line. What does do your outcomes?