Bad taste is not simply liking ugly things.

That definition is too narrow and too aesthetic. In operating work, bad taste is failed discernment. It is the inability to tell what matters, what fits, what will age badly, what is merely polished, what is hiding risk, or what deserves rejection.

Bad taste can look sophisticated. That is what makes it dangerous.

The first failure mode is mistaking polish for quality.

A polished deck can hide a weak strategy. A clean interface can hide an unresolved workflow. A fluent AI answer can hide unsupported claims. A charismatic candidate can hide shallow operating habits. A beautifully written customer email can still avoid ownership.

Polish lowers resistance. It makes people want to believe the work is done. Taste asks whether the polish is sitting on top of something true.

The second failure mode is mistaking novelty for quality.

New feels alive. New creates energy. New can also be an escape from the boring truth that the current problem has not been solved. Teams with bad taste chase new formats, new tools, new campaigns, new frameworks, new positioning, new workflows, and new AI use cases before they understand the actual constraint.

Novelty is valuable when it opens a better path. It is expensive when it helps the organization avoid discipline.

The third failure mode is mistaking cleverness for quality.

Clever work announces the intelligence of the maker. Good work serves the situation. The clever abstraction, clever tagline, clever org design, clever pricing mechanic, or clever automation may be enjoyable to build and painful to operate.

Taste has a low tolerance for cleverness that transfers cost downstream.

The fourth failure mode is mistaking category convention for quality.

Every category develops a look, language, and rhythm. B2B SaaS pages start sounding the same. Strategy decks start using the same shapes. Product onboarding starts copying the same patterns. Executive updates start speaking in the same safe phrases. AI-generated content accelerates this because models are trained to reproduce the middle of the category.

Category convention is not always wrong. It can reduce friction. But bad taste confuses recognizable with good. It produces work that looks like it belongs and therefore says nothing.

The fifth failure mode is mistaking personal preference for judgment.

This is common among senior people. They have strong reactions, some of which are right. Over time, they stop explaining the mechanism. The team learns the leader’s taste as a weather system: sometimes sunny, sometimes stormy, always requiring interpretation.

Bad taste hides behind authority. Good taste makes itself accountable.

The sixth failure mode is mistaking intensity for standards.

Some leaders think high standards mean reacting strongly. They tear work apart, reject drafts late, demand excellence, and create fear around review. But intensity is not a standard. A standard is observable, teachable, and enforced consistently. If people leave the review shaken but not sharper, the organization did not receive taste. It received pressure.

The seventh failure mode is mistaking local optimization for quality.

A sales deck that helps close one deal may damage positioning. A custom feature may protect one renewal while complicating the roadmap. A metric may make one team look good while distorting company behavior. A process may improve control while slowing the work that matters. A manager may optimize their team for spotless weekly reporting while starving the cross-functional work that actually moves the company. A technical team may improve its own deploy cadence by pushing integration ambiguity onto support and customer success.

Taste looks across the system. Bad taste celebrates the local win and ignores the bill.

The eighth failure mode is mistaking confidence for calibration.

People with bad taste are often either too certain or too vague. They know, but cannot explain. Or they refuse to judge at all because everything is subjective. Both are weak. Taste requires confidence with reasons and humility about what could change the judgment.

This matters in technical and AI-assisted work. A non-engineer operator with good taste does not pretend to be the architect. They ask sharp questions about consequences, ownership, reversibility, and failure modes. A weak operator either rubber-stamps the technical answer or overrules it based on surface confidence.

The ninth failure mode is stale taste.

Taste can be right for one market moment and wrong for another. Customer expectations move. Tools change. Competitive baselines rise. Company strategy shifts. What once signaled quality can become ceremony. What once felt risky can become standard. What once worked at small scale can fail at larger scale.

Uncalibrated taste becomes nostalgia.

The tenth failure mode is taste without responsibility.

This is the critic’s disease. It is easy to see what is wrong when you do not own constraints. The work could be sharper, cleaner, more original, more elegant, more robust. Fine. What should be done by Friday with the team we have, the customer waiting, the risk profile in front of us, and the budget already committed?

Taste that cannot operate under constraint is not enough. It may be perceptive, but it is not useful.

The practical question is how to detect bad taste early.

Look for repeated vague corrections: “make it better,” “more premium,” “not strategic enough,” “cleaner,” “stronger,” “more polished.” These may point to real issues, but if the reasons never become explicit, the team cannot learn.

Look for work that is impressive in review and disappointing in use. That usually means the organization is judging presentation instead of consequence.

Look for too many accepted “fine” artifacts. Fine is often where quality goes to decay. Fine documents, fine designs, fine hires, fine AI outputs, fine meetings, fine handoffs. Nothing is bad enough to trigger alarm. Everything is weak enough to lower the bar.

Look for standards that depend on one person. If the founder, head of product, editor, or principal engineer has to personally catch every issue, the company does not have team taste. It has centralized inspection.

Look for AI workflows that celebrate generation volume without acceptance discipline. If the team tracks how much output was produced but not how much survived review, improved decisions, reduced rework, or created risk, bad taste is entering the system.

Bad taste is not fixed by telling people to have better taste. It is fixed by making judgment visible: examples, anti-examples, critique with reasons, aftermath reviews, quality rubrics, decision criteria, and standards with owners.

The goal is not to make everyone an aesthete. The goal is to reduce preventable mediocrity.

Bad taste lets mediocre work pass for complete. Good taste catches it while it is still cheap to change.