Startup Financial Modeling is Broken
Let me be perfectly clear, businesses need plans. Models help turn a plan into numbers. The problem is: Startups usually just build the wrong ones at the wrong time. I benefited tremendously from the status quo. My early career was essentially built on 5 year operating models. But I still think it’s time for a change.
The realization started to sink in in 2018. An investor had asked: "Why is this only a 5-year plan?“ We were a 2 year old company. I broke a sweat, the model already took 30 seconds to recalculate every time a number was changed. Excel had regressed to the speed of a dial-up modem. For weeks on end, I had spent nights building waterfalls calculating cohorted subscription revenues, asset values and loss rates for the next half decade. Excel was at its breaking-point, so was I. And they wanted me to extend it further?
2018 Lukas isn’t alone. Founders & finance leaders in our industry often spend months building out operating models that only back into a target number that was preordained for board, investors or both. What a waste. Yet here we are - and everyone in the industry shares a bit of blame.
For sure, Startups need a big, audacious goal to connect to the vision. It rallies the troops and investors. Nothing wrong with aiming for $100 million in revenue - quite the opposite actually. Just don’t spend months building the 17-tab model to match.
Instead, I believe there are three different financial "models" that will help people become better entrepreneurs or startup operators: A simple formula, a detailed unit economics model, and a rolling cash forecast.
A simple business formula. The formula should create a simple narrative that everyone internally and externally can understand and translate into how things are going. They are often based on some combination of TAM, market-share, number of customers, and revenue or gross profit per unit. Be careful - the number going up should pretty binarily mean progress.
E.g. Payment Processor Revenue = # of Customers x Avg. GMV per Company x Take-Rate
A detailed unit econ model. Each company has a few key inputs that either make or break its business model. The team should have these down cold. What they are, what influences them, where they are today and where they will be at scale.
From experience, this is the model that has worked best in early round fundraising both at Pipe and at Fair. It makes it easy for anyone new to the business (or industry) to understand the micro scale first and then extrapolate to the macro scale.
A rolling cash flow forecast. I don’t know who needs to hear this but GAAP is not the real world. And waiting 45-90 days on your external bookkeepers to clean up your quickbooks also is not a viable option to understand your financial health. There is some sophisticated tooling out there to help you with this (most of which is quite expensive) - but at every company that I’ve ever been at, I’ve implemented a weekly process of importing transactions, tagging them and rolling forward what we expect to happen cash-wise. You should too. It’s the easiest way to learn the natural ebbs and flows of your business. And if things ever get tight, you will know well ahead and you can react when the necessary changes might still be manageable, not catastrophic.
You might ask yourself: What about a budget? A budget is crucial - but only for companies that can actually manage to one. And most young companies can’t. You would need department or team leads that can plan and manage effectively. In lieu of that being the case, this mostly would mostly just fall on you. So don’t waste your time (or your team's) to get more bureaucratic than needed. The three alternatives we discussed will give you warning signals the moment things go out of whack.
Obviously, there’s a point when this all is no longer accurate. Maturing companies have more stable financials, experienced leaders and sufficient finance talent. At that point you start a journey that ideally ends with an IPO and a 90 day shot- clock to report & forecast financials. And the 17-tab operating model from earlier might quickly turn out to not be nearly detailed enough.
Cheers,
LW