📎 Webclip
Drive development with budgets not estimates
The post argues that software development should be planned by budget rather than by estimate. A feature’s worth sets the budget, and that budget can be informed by an estimate, but should not be controlled by it.
It says features can vary widely in scope, so a full implementation is often far more costly than a useful partial one. On a constrained budget, teams should challenge assumptions, avoid gold-plating every feature, and accept that invention happens within limits.
Reading notes#
- Development planning often relies on estimates, but the post says this model has been broken for a long time.
- A better approach is to state what a feature is worth and treat that as the budget.
- The useful budget may be informed by an estimate, but it is not driven by one.
- Feature scope can vary a lot, from a version that takes 2 weeks to one that takes 6 months.
- In many cases, 80% of a feature can be built for 20% of the effort.
- That tradeoff can make it better to deliver several partial features, improvements, or fixes instead of one fully polished implementation.
- This matters especially when money and people are limited.
- Spending too much on polish for every feature can leave nothing left to ship the product.
- A budget based on worth forces attention onto assumptions that can be challenged or ignored.
- Accepting a budget means accepting constraints, and the post presents invention as something that happens inside those constraints.
