Tech

Models on their own are unreliable. Utterly pointless. Too often, we see misleading decisions being made from that spike in an excel line graph or the absolute value from a financial model. But models are only a simplification of reality. A reflection of history. A benchmark or an anchor for future predictions.

Models are powerful when they’re showcased in a story — contextualising the data from introduction, analysis and recommendations.I recently had drinks with a friend who works in investment banking we came to a simplified (overly simplified) conclusion that investment bankers and consultants do the same thing. We validate an opportunity by building a model around the future-state. This is the ground work. We then hook the deal by injecting industry insights, playing with variables of financial/operating models, and aligning proposed options with strategic objectives of the organisation. So how do you turn a beautifully complex model into a compelling story?

If you haven’t built the model yourself, its well worth investing the time to understand its mechanics — i.e. what are the data sources and how are they applied to produce an output? Ideally, most models should have a buffer in a range where a set of options might sit. A basic modelling principle which requires a thorough understanding of the quality of the data, the context of the business and how it might impact the overall result. For instance, you might apply a 20% contingency buffer to an opex forecast to factor incidentals, or use industry benchmarks for a valuation to include worst and best case scenarios around the actual value.

There’s a clear story here. Adding a buffer to a model says this model isn’t perfect, but there’s a science to how we manage the downside of assumptions or variability in the data.

p>The fun with modelling is in playing with variables to produce different scenarios — be it changing prices, rates of return, stock variances, labor utilisation or even failed projects. You might do this with fancy Monte Carlo simulations (which I’ve never got right!), or manually change a model’s inputs to test its impact on the final outcome.

We create stories through these scenarios. How would the outcome differ if we canned a project half-way through its lifespan? What would happen if the price of wool suddenly skyrocketed up by 80%? What would happen if we acquired company Y instead of company X…or both?

Creating alternate worlds with our models allows us to make decisions about the risks we need to mitigate and the extent to which we might want to control variables depending on their relative impact.

A model becomes meaningful when it inspires action. Action which comes from a decision. A decision which comes from seeing an opportunity that ties directly with a business and social need.

A model is ultimately used to make recommendations on a set of initiatives, drawing upon the vision and strategic intent of the enterprise. But it takes a lot to get to the final cut. It takes a deep understanding of factors that would disrupt the organisation internally, in the broader industry and how these factors translate to a problem statement and call for action.

The recommendations itself must tell a story through a solid business case. It tells us:

• Purpose and need
• Opportunity costs (of choosing one initiative of another)
• Risk mitigation
• Measuring performance

At the end of the day, the model is really only a small part of a pitch. The story is in the detail, the assumptions, the data sources and all the little nuances which contextualise its existence.