The Sydney Opera House was more than 10x over budget.

Welcome back to the Behavioural Unicorn, the newsletter that’s smoother than your latte. Let’s crack in:

Projects almost always cost waaaaay more than we think

• The Sydney Opera House was predicted to cost $7 million but ended up being $102 million

• The James Webb telescope was predicted to cost $4 billion but ended up costing more than double at $9.4 billion.

• The Montreal Olympics took an enormous toll on the taxpayer. It was expected to cost $300 million but took $1.4 billion! A bill that took almost 30 years to pay

Holey Moley, those are some bad days at the office.

Why is it so hard to make these predictions?

Well, there are 2 common mistakes people make

1. Uniqueness Bias: People think their project is unlike any other. Therefore, they can’t make comparisons with similar projects. No one has built something like this before!

2. Optimism Bias: When people are too optimistic about a project, bad things can happen:

How Can We Get Better At Making Forecasts?

Great question my fellow unicorn.

One way is to use reference class forecasting. Here’s how it works:

Imagine we’re building a train station and we want to predict how much it’ll cost.

Here’s how to do it:

Step 1: Ask How Long Do Projects Like This Tend To Cost?

Let’s look at other train stations or similar-sized projects and ask: How much did they cost?

Say the average is $ 1 million (I have no clue, unicorns hate trains). This number is your anchor point

This sounds simple (and it is), but most people don’t make comparisons like this. They think their project is unique, but it rarely is.

Sometimes comparison is the thief of joy, sometimes comparison saves you bucket loads of money

Step 2: Adjust The Anchor Point

Take your anchor point and bump it up or down depending on your project.

Maybe you anticipate some delays with the project so you add $50,000 on to the cost.

But perhaps you can get a deal with some supplies, so you lower the estimate by $10,000.

You’re Probably Not Building A Train Station

No stress my friend, you can use this strategy to predict more normal things.

• How much will your home renovation cost?

• How much will Sally’s 5th birthday set you back?

• How much will I spend at the casino this weekend?

You Can Also Use This Method To Predict Things Other Than Cost

• How many CVs will I have to write before I get a job offer?

• How long will it take me to run my first marathon?

• How long will it take me to read this darn email?

To wrap it up:

To make better predictions use reference class forecasting

1. Make comparisons: Ask how much do these things tend to cost

2. Adjust: Bump that number up or down

This idea is from How Big Things Get Done (and it’s awesome)

Cheers,

Theodore