Why I'm Building Recommendation Engines for Art (Not Just Netflix)

Last month, I found myself down a rabbit hole trying to solve what seemed like a simple problem: finding art that doesn't suck for my home office. As someone who's spent years optimizing recommendation algorithms for e-commerce, I was shocked by how primitive art discovery still feels online.

Most art platforms still rely on basic category filters—"abstract," "landscape," "portrait"—as if we're browsing a 1990s directory. Meanwhile, Spotify can surface obscure indie tracks that perfectly match my coding mood, and GitHub's explore page consistently serves up repositories I didn't know I needed.

This got me thinking: why hasn't the art world embraced the same data-driven discovery mechanisms that work everywhere else?

The Data Problem

The challenge isn't technical complexity—it's data richness. When Netflix recommends a show, they're working with viewing history, completion rates, genre preferences, and even the time of day you watch. Art platforms typically have purchase history and maybe some favoriting behavior. That's it.

But what if we could capture more nuanced signals? Color palette preferences, compositional elements, even the emotional response to certain artistic movements. The computer vision tools to extract these features exist. We're just not using them creatively enough.

Beyond the Algorithm

I've been experimenting with some prototype recommendation engines for visual art, and the most interesting insights come from combining multiple data sources. Image analysis reveals color and texture preferences. Browsing patterns show style evolution over time. Cross-referencing with music streaming data (with permission) can even surface correlations between sonic and visual aesthetics.

One Australian marketplace I've been following has taken an interesting approach to this problem. Their weekly art curation combines algorithmic suggestions with human editorial insight—essentially creating a hybrid system that feels both personal and serendipitous. It's the kind of thoughtful intersection between data and intuition that makes discovery feel less mechanical.

The Creative Tools Renaissance

What excites me most isn't just better discovery, but how these same technologies are empowering artists themselves. AI-assisted color palette generation, automated social media optimization, blockchain-based provenance tracking—we're seeing a creative tools renaissance that rivals what happened in music production over the past decade.

The artists who understand these tools aren't replacing creativity with automation; they're amplifying their creative capacity. They're spending less time on administrative overhead and more time on what matters: making work that resonates.

Building Better Bridges

As developers, we have an opportunity to build better bridges between creators and audiences. Not through disruptive blockchain art marketplaces or NFT speculation, but through thoughtful tools that solve real problems in art discovery and creator sustainability.

The most successful artsale platforms of the next decade won't just be marketplaces—they'll be recommendation engines, creative tools, and community platforms rolled into one. They'll understand that buying art online isn't just about transactions; it's about relationships between creators, curators, and collectors.

What would you build to improve how we discover and connect with visual art?