The most important product decision is what you don't build
This article argues that the most critical product decision is often what not to build, citing the hidden costs of maintenance and feature creep. It resonates deeply on Hacker News, where developers and product managers constantly battle featuritis, technical debt, and the institutional bias towards creation over subtraction.
The Lowdown
Liam Nugent's piece posits a provocative truth: the most impactful product decision isn't what you ship, but what you choose not to build. He illustrates this with common pitfalls like generic "document hubs" and "notification centers" in financial apps, which start simple but balloon into complex, high-maintenance monsters.
- The Cost of Maintenance: The author highlights that while building new features might seem cheap, the long-term running costs and maintenance burden (e.g., adapting to OS updates) are often ignored, leading to significant overhead.
- Battling the Bias for Addition: Nugent explains that organizations and even human psychology are wired to prefer additive solutions, making it difficult to advocate for removing features or declining new ones. Rewarding creation over maintenance perpetuates this problem.
- Henry Ford's Wisdom & Steve Jobs' Example: He draws parallels to Henry Ford's 'faster horses' analogy, suggesting that users don't always know what they truly need. Steve Jobs' 1997 product matrix at Apple is presented as a prime example of radical subtraction leading to intense focus and monumental success.
- AI as a Pruning Tool: The article briefly touches on DHH's concept of "endless execution" with agents, suggesting that AI might be better utilized for maintenance and judicious removal rather than just building more.
Ultimately, Nugent advises teams to be extremely selective about what gets added and militant about what gets taken out, emphasizing that success lies in meeting customer needs efficiently, not in an ever-expanding feature set.
The Gossip
Pruning Pains & Gains
Many commenters strongly agree with the article's premise, sharing personal anecdotes of saving sprints and reducing maintenance by successfully killing features. They highlight the relief and efficiency gained from avoiding unnecessary builds and advocating for feature cuts against product/design pressure.
Custom Reporting Conundrum
A recurring example of a feature often built (or requested to be built) unnecessarily is 'custom reporting.' Commenters detail how this vague request frequently stems from a lack of clarity on what data users truly need or that existing tools are insufficient, rather than a genuine need for a complex, new reporting engine. Solutions range from asking better questions about underlying needs to providing raw data APIs.
The Allure & Trap of Easy Building
Several discussions revolve around the double-edged sword of modern development, particularly with LLMs, making it 'nearly free' or 'fast and inexpensive' to build. While this lowers the barrier to experimentation, it also increases the risk of falling into the trap of over-building or creating prototypes that become obligations, exacerbating the problem the article describes.
Beyond the Build: User Expectations vs. Needs
Some commenters discuss the challenge of managing user expectations. Users often ask for features they are familiar with (e.g., 'hypnograms' in sleep tech) even if those features don't provide real value, or they are just trying to address a core problem in a suboptimal way. The discussion highlights the need to understand the underlying 'why' before building a requested 'what'.