FragIndex

FragIndex needed to display a lot of information without making the page feel overwhelming. Each fragrance page includes brand information, notes, accords, ratings, clone recommendations, community votes, and buy links, so the UI had to organize that information into clear sections that users could scan quickly.

I designed the interface around a clean, card-based layout with strong spacing, readable typography, and separate content blocks for each major decision point. For the fragrance pages, I focused on making the most important information visible early: the fragrance name, brand, overall rating, short description, and best clone. Supporting details like notes, accords, performance ratings, season, occasion, and similarity were grouped into their own sections so users could dig deeper without losing the main context.

I also wanted the voting UI to feel approachable. Instead of hiding ratings behind a complicated form, each category was designed as its own interaction. The overall UI direction was inspired by modern discovery platforms like Letterboxd and Backloggd, where the product feels clean, community-driven, and easy to browse while still supporting a large amount of structured data.

I chose to separate voting into individual categories instead of using one generic rating system. A fragrance can smell great but have weak longevity. A clone can be affordable but not very similar to the original. Because of that, FragIndex lets users vote on categories like scent, longevity, projection, occasion, season, time of day, and similarity separately.

This makes the data more useful than a single average rating.

FragIndex was built around searchable, indexable pages rather than only app-style navigation. The platform uses dynamic routes for fragrance and brand pages, structured metadata, sitemap support, and crawlable content so that users can discover pages through Google.

This matters because many fragrance searches are long-tail queries, such as notes, clones, comparisons, and "smells like" searches. The platform is designed to grow through programmatic SEO and internal linking rather than relying only on social media or paid traffic.

I wanted FragIndex to feel more like a discovery platform than a static database. Internal linking is important because users rarely stop after researching one fragrance. They may want to compare similar scents, browse a brand, find a clone, or discover fragrances with similar notes.

The long-term product direction is to make FragIndex easier to browse through related fragrances, best clone rankings, note-based discovery, trending pages, and comparison pages.

FragIndex includes outbound retailer links so users can move from research to purchase. This creates a natural monetization path through affiliate links or retailer partnerships. The key product decision is that monetization should support the research experience instead of interrupting it.

The goal is not to overload pages with ads immediately, but to place helpful purchase options where users already have buying intent.

One of the biggest challenges was balancing SEO with product UX. FragIndex needed pages that were crawlable, structured, and useful for Google, but the site also needed to feel like a real discovery product rather than a thin SEO page generator.

Another challenge was designing the voting system. A simple star rating would have been easier, but fragrance decisions depend on multiple traits. I had to think through how users should rate categories independently while keeping the interface simple enough that contributing felt lightweight.

Performance and scale were also important challenges. Since the site has dynamic pages, voting data, images, and aggregate ratings, I had to think through caching, ISR, API usage, and how to avoid unnecessary database calls as traffic grew.

Building FragIndex taught me that SEO products need more than pages. They need useful page templates, strong internal linking, fast performance, and a reason for users to stay after they land from search.

I also learned that community data is only valuable when it is structured around real user decisions. For fragrance research, one generic rating is less useful than separate signals for scent, performance, season, occasion, and similarity.

Another major lesson was that traffic growth and monetization require different layers. Ranking pages can bring users in, but the product still needs discovery paths, trust signals, buy links, and comparison features to turn traffic into long-term value.