all models are wrong

some are useful

Business is easy. Money goes in, something happens and money comes out. A model is a dynamic representation of that 'something'. If you understand a businesses good enough to build a sound model, you can likely grow it in size and profitability.

My first business was a deli in west London. I had no idea how to forecast sales nor costs, and that's because I didn't understand the business. Fortunately for me food isn't a complex one and I was able to figure it out as I went. However if I was able to model it right from the start I'd probably would have done it very differently.

All models are wrong. That's right, embrace it. I didn't know this at the start of my career, I couldn’t see the value of any model given they were all likely going to be wrong. It took me a few years to understand that the value of a model isn’t its actual deliverable (whatever format).

It’s not about the destination, it’s about the journey (sorry couldn’t help myself). Learning about all the inputs stocks and outputs that make your business is the true value. By synthesizing your business you learn about what’s important to pay attention to. Oh I love modeling.

Models give confidence to decision making. I’ve been using financial models to run my household decisions for this very reason. What's out projected income, liabilities and cash flow over the next 3,6,9 months? Can we afford this risk? The answer becomes binary.

And what’s funny, is that it turns out that I always underestimate our true appetite for risk. Before modeling a decision I always assume it’s going to be harder to achieve decent returns than it turns out to be.

I've ran financial projections for our household before moving to Paris, Los Angeles and San Francisco, the picture they showed me was the confidence I needed to take practical strategic decision that would have major repercussions well beyond the impact on our savings account.

Modeling a business gives the same type of hindsight. Ask yourself: how do I get this outcome? What levers are reasonable to pull in order to get there? Models are decision making MVPs (minimum viable products).

And the answer will likely be wrong, that's fine, the answer is not the point. As long as you have conviction that your model is the best it can be done, move on.

You can iterate on models (I think I start a new cash flow analysis from scratch for every time we make a big decision, I’ve done at least twenty different by now). Starting from a very basic income statement all the way to complex seasonal trends and deferred sales cycles. Just because you can't get a precise answer today doesn't mean you won't get one eventually, especially if you focus on it.

Once your first draft is out it's easy to figure out how to improve it, what data are you missing? Add it in, if you are collecting now you know you have to. You know where the resources are, how to arrange the data, what pivots you need to see every aspect of the idea you are evaluating.

Start bottom up. Record every touchpoint from your current business funnel. It should be easy to pull, don’t overcomplicate (analysis). Once you have conversion rates build a harness for the data to come to life (synthesis). Plug in your assumptions and done. What's not to love about that?

Pro tip: build a switch into your model. Base better and best assumptions, always. What works at 10x likely doesn’t at 100.

If you wanna chat about modeling reach out. Would love to hear your best practices.