The tech industry has spent the last decade optimising everything. We A/B test button colours, streamline user flows, and personalise experiences down to the individual pixel. The goal is always the same: reduce friction, increase conversion, scale infinitely.
Meanwhile, luxury brands are doing the exact opposite. Hermès makes their products harder to buy. Bottega Veneta deleted their Instagram account. These brands intentionally create friction, limit access, and resist scale, and they're outperforming their algorithmic competitors.
This isn't just interesting: it's counterpoint to how we think about building products in an AI-dominated world.
The Scarcity Engine
Hermès operates fewer than 300 stores globally. Each Birkin bag takes a single artisan 18+ hours to complete. They could scale this process, open more stores, hire more artisans. Instead, they maintain artificial constraints that keep supply permanently below demand.
This flies in the face of Silicon Valley orthodoxy, where the goal is always to eliminate bottlenecks and scale efficiently. But luxury brands understand something we've forgotten: scarcity creates value in ways that abundance cannot.
Research in behavioural economics shows that scarcity fundamentally alters how we process value. When resources feel abundant, we rely on quick, automatic decision-making. When they feel scarce, we engage deeper cognitive processes that create stronger emotional attachment and willingness to pay premium prices.
The implication for tech products is profound. Instead of optimising for maximum accessibility, what if we designed for meaningful scarcity? Not artificial paywalls or arbitrary limits, but genuine constraints that make the experience more valuable.
The Personalisation Trap
The other counterintuitive insight: luxury brands are deliberately bad at personalisation by tech standards.
While we obsess over 1:1 customisation and dynamic content, Hermès has essentially the same product line they had decades ago. Their creative directors make aesthetic decisions based on their own vision, not user research or behavioural data. Customers either appreciate their taste or they don't.
While most of us want personalised experiences, we are deeply concerned about data privacy, and few of us trust companies to use our information responsibly. We've created a personalisation arms race that's making users increasingly uncomfortable.
Luxury brands skip this entirely. Instead of asking “what does the user want?” they ask “what should exist?” They exercise editorial judgment and taste, creating products that customers didn't know they wanted until they saw them.
This suggests a different approach to product development: instead of optimising for revealed preferences, optimise for latent desires. Create things that shape taste rather than merely reflect it.
Designed Imperfection
Perhaps the most radical luxury principle is the deliberate inclusion of "flaws." Hand-stitched seams that show slight variations. Asymmetries that prove human involvement. Irregularities that algorithms would eliminate.
As AI becomes capable of generating technically perfect designs, these imperfections become increasingly valuable differentiators. Current design trends in 2025 are moving toward authenticity through brutalist elements and intentional imperfections that strengthen user connections, demonstrating that websites don't need conventional perfection to engage users effectively.
This hunger for imperfection reflects a broader backlash against algorithmic uniformity. Early indicators suggest human curation is gaining premium positioning—from Substack's human-curated recommendations to Spotify's editorial playlists commanding higher engagement than algorithmic ones. “Non AI-generated” will become the new “non GMO” sticker as companies differentiate through exclusive human curation rather than automated optimisation.
The Return to Human Judgment
The value in such an approach isn’t just sentimental—it's measurable. Research from Carnegie Mellon published in 2023 found that combining human curation with algorithmic recommendations can increase engagement by up to 13%, with human experts providing "broad knowledge accumulated over a professional career" that algorithms lack. The study showed that humans performed better than algorithms on days with more attention-grabbing news, suggesting human judgment excels when context and taste matter most.
Even platforms built on algorithmic feeds are facing user resistance. When Mastodon started experimenting with recommendations in late 2024, users pushed back, preferring human curated spaces over machine recommendations because algorithms create even deeper echo chambers and narrow recommendations down to the op performers. One I personally use, lynkmi.com, feels like your best friend group chat for recommendations.
Implications for Product Strategy
The implication isn't to abandon algorithms entirely, but to understand their limits. As AI capabilities become commoditised, sustainable differentiation increasingly comes from distinctly human qualities: editorial judgment, aesthetic choices that can't be A/B tested, and the confidence to create deliberate imperfections that signal authentic human involvement rather than algorithmic optimisation.
The anti-algorithm isn't about rejecting progress—it's about preserving space for the irreplaceable human elements that create lasting value.
These luxury principles suggest several counterintuitive approaches for tech products:
Intentional Constraints: Instead of optimising for optionality, ruthlessly limit choices to create focus and clarity. The paradox of choice is real—too many options create decision paralysis and decrease satisfaction.
Editorial Decision-Making: Use data to inform decisions, but don't let it make decisions. Human judgment and taste become more valuable as algorithmic optimisation becomes commoditised.
Premium Friction: Not all friction is bad. Some experiences should require effort, investment, and intentionality. The best communities often have application processes. The most valuable products have learning curves.
Identity Over Utility: Design for what people want to become, not just what they need to accomplish. Products that become part of users' identity create stronger loyalty than those that merely solve problems.
The Algorithmic Resistance
None of this is about rejecting technology. Luxury brands use sophisticated logistics, materials science, and manufacturing techniques. But they subordinate these capabilities to human vision and judgment.
The opportunity for tech companies isn't to abandon algorithms, but to understand their limits. AI excels at working within known parameters, but luxury creates new parameters of desire. Algorithms can improve efficiency, but human taste determines what's worth being efficient about.
As AI capabilities become commoditised, sustainable differentiation will increasingly come from distinctly human qualities: judgment, taste, vision, and the confidence to make opinionated choices about what matters.
It is true that many luxury brands operate on different timescales than tech companies. They measure success in decades rather than quarters. Products from the 1950s are still desirable today. This temporal perspective allows them to resist much of the short-term over-optimisation tech trends suffer from in favour of leaning into long-term value creation.
This suggests a fundamental question for product builders: are we prioritising for next quarter's metrics or next decade's relevance? The brands that endure will be those that balance algorithmic efficiency with human meaning.
The anti-algorithm is an acknowledgement that all taste isn’t binary encoded. It's about preserving space for the irreplaceable human elements that create lasting value: judgment, craftsmanship, and the wisdom to know what shouldn't be optimised away.