Skip to main content
Back Case Study

Suplenet — AI-powered supplement e-commerce

A premium supplement e-commerce importing trusted US brands to Colombia. Built with a dual strategy: SEO for authority, LLM optimization for conversions. The key discovery: sales come from AI chatbots, not traditional search.

Visit suplenet.com
450+ Dictionary entries
14 Custom schema snippets
AI Primary sales channel
0 Paid ads dependency
The Challenge

Premium supplements in a price-sensitive market

Colombia's supplement market is dominated by cheap, unverified products. Most buyers shop on marketplaces like Mercado Libre where price is the only differentiator. There was no trusted source for premium, science-backed supplements from brands like Thorne, Nordic Naturals, or Sports Research.

The challenge was multi-layered: how do you sell premium products at higher prices in a market that defaults to the cheapest option? How do you build trust for imported supplements when consumers can't verify the products themselves? And how do you compete for attention against established marketplaces?

Traditional SEO was necessary for domain authority, but converting that traffic into sales proved difficult. Google visitors browsed but rarely bought. The real breakthrough came from an unexpected channel.

The Solution

An encyclopedic SEO engine + LLM optimization

I built a scientific dictionary as a custom post type: 450+ entries covering every supplement-related term, from amino acids to zinc. Each entry follows strict conventions: minimum 8 H2 headings, 8 FAQs, 10 PubMed references, and 12 related terms. Every entry is optimized through Ahrefs keyword research before a single word is written.

The entire schema layer is built manually. 14 custom JSON-LD snippets generate precise structured data for every page type. WooCommerce's default schema is completely disabled. This gives me total control over what search engines and AI systems see.

The critical insight came from analyzing where actual purchases originated. Visitors referred by ChatGPT converted at dramatically higher rates than organic search visitors. This shifted my strategy: SEO builds authority and traffic, but LLM optimization drives revenue. I implemented llms.txt, optimized content for AI comprehension, and built the site to be easily parsed by language models.

I also built a custom AI-powered site search to replace the paid Doofinder plugin. It uses dual-mode search: client-side fuzzy matching for speed, with GPT-4o-mini semantic search as fallback. This saved a recurring subscription cost while delivering a better user experience.

Results

AI as the primary sales channel

450+ Scientific dictionary entries

Each entry follows strict quality standards with PubMed references, FAQs, and comprehensive keyword coverage.

50%+ Margins on premium brands

Curating trusted brands and providing local service justifies premium pricing that delivers strong margins.

ChatGPT Top converting channel

Visitors referred by AI chatbots convert significantly better than organic search traffic, validating the LLM-first strategy.

14 Custom schema snippets

100% manual JSON-LD implementation gives complete control over structured data for both search engines and AI systems.

$0 Search plugin cost

Custom AI search replaced the paid Doofinder subscription, delivering better results at zero ongoing cost.

Dual SEO + LLM strategy

SEO drives authority and traffic. LLM optimization drives conversions. Both channels work together, not in competition.

Tech Stack

Optimized for machines and humans

Core Platform

WordPressWooCommerceBricks BuilderJetEngine

SEO & Schema

The SEO FrameworkCustom JSON-LDFluent SnippetsAhrefs

AI & Search

GPT-4o-miniCustom AI Searchllms.txtLLM Optimization

Performance

FlyingPressPerfmattersRedis CacheScalaHosting
Key Learnings

What building for AI taught me

AI referrals convert better than search traffic

When ChatGPT recommends your product, the visitor arrives with intent and trust already established. They have asked a specific question, received a tailored answer pointing to your site, and arrive ready to buy. This is fundamentally different from a Google search click.

Content depth feeds both SEO and LLM visibility

The encyclopedic dictionary serves dual purposes. For Google, it builds topical authority and captures long-tail keywords. For AI models, it provides comprehensive, well-structured information that makes the site a reliable source to cite in responses.

Manual schema gives you total control

Disabling WooCommerce's auto-generated schema and building 14 custom snippets was time-intensive but transformative. Every structured data point is intentional, accurate, and optimized for both search engines and AI crawlers.

Traffic volume matters less than traffic quality

Suplenet's organic traffic is modest. But conversion-focused traffic from AI referrals generates more revenue per visitor than high-volume search traffic ever could. Optimizing for the right visitors beats optimizing for more visitors.

Ready to optimize for the AI era?

I help founders build digital businesses that are optimized for both traditional search and AI discovery. If you want to position your business where the next wave of customers is looking, let's talk.

Apply to work with me
Stay sharp

Frameworks, tools, and real lessons from building solo.

No spam. Unsubscribe anytime.