The Gambler's Guide to Software Development

How breaking big technology investments into smaller, smarter bets can transform project outcomes and reduce risk

11/6/2021
Wanna bet?

The High-Stakes Game of Software Development

Imagine walking into a casino and placing your entire project budget on a single hand of poker – before you’ve even seen your cards. Sounds absurd, right? Yet this is exactly how many organizations approach software development. They commit millions of dollars and years of effort based on initial requirements and assumptions, essentially betting everything on their ability to predict the future.

This approach persists despite a fundamental truth: software development isn’t like other engineering disciplines. When you want to build a bridge in Chicago, you can examine successful bridges in Milwaukee and replicate their design. When you want to buy a sports car, you can purchase the exact same model someone else is driving. But software? If the exact solution you need already existed, you’d simply buy it off the shelf – after all, software is infinitely reproducible at nearly zero cost.

The fact that you’re building custom software means you’re creating something new. And new things, by definition, come with uncertainty.

The Cone of Uncertainty

In project planning, there’s a concept called the cone of uncertainty. Like weather forecasting, our ability to predict outcomes becomes increasingly unreliable as we look further into the future. A tiny change in initial conditions – a new industry standard, an emerging technology, a shift in user preferences – can cascade into major deviations over time.

This uncertainty creates a peculiar challenge in software development. When companies engage with development partners, they often encounter conversations like this:

“How much will it cost to build our system?” “It will take two years and cost $5 million.”

Let’s apply Annie Duke’s “thinking in bets” framework here: Would you bet $5 million that every single assumption you’re making today about your needs, user requirements, technical choices, and market conditions will still be correct two years from now?

The Smart Way to Place Your Bets

This is where agile development – stripped of its buzzword status and returned to its fundamental principles – becomes not just useful but essential. True agile development isn’t about daily standups or sprint planning (though these can help). It’s about transforming one massive, risky bet into a series of smaller, more manageable ones.

Consider this alternative approach:

  1. Start with Core Value: Instead of building an entire system with user management, file handling, and complex interfaces, first prove that your core value proposition works. If you’re building legal analysis software powered by large language models, validate that the AI can actually perform the analysis effectively before investing in surrounding infrastructure.

  2. Incremental Investment: Rather than committing your entire budget upfront, allocate resources in stages. Each stage should provide concrete validation of your assumptions and create actual value. This allows you to adjust course based on real-world feedback rather than theoretical plans.

  3. Risk-First Development: Identify your riskiest assumptions and test them early. A good technical partner will help you spot these risks and create targeted proof-of-concepts to validate or invalidate them before significant investment occurs.

Making Better Bets

When approaching any software project, ask yourself these questions:

  • What am I actually betting on? (Identify your core assumptions)
  • How much am I willing to risk on each assumption?
  • What’s the smallest bet I can make to validate my riskiest assumption?
  • How can I break this project into smaller, safer bets?

Remember: You’re not just building software – you’re navigating uncertainty. Each decision is a bet, and like any good gambler, you want to manage your risk and maximize your information before placing large wagers.

The Path Forward

The next time you’re planning a software project, resist the urge to plan everything upfront. Instead:

  1. Identify your core value proposition
  2. Break the project into small, testable increments
  3. Start with your riskiest assumptions
  4. Validate each step before proceeding to the next
  5. Maintain flexibility to adjust based on what you learn

By thinking in bets, you transform software development from a single, high-risk gamble into a series of thoughtful, calculated investments. Each step provides new information that helps you make better decisions about the next one. This isn’t just a better way to build software – it’s a better way to manage any complex, uncertain endeavor. After all, when you can’t eliminate uncertainty, the next best thing is to manage it intelligently.

Written by Alexander North