TL;DR
Building Nexeo taught us: server rendering is 7x more impactful than any other factor, brand authority compounds (first citation is hardest), entity consistency affects 40-60% of citations, and product pages get cited 3-5x more than blog posts for purchase-intent queries.
The question we couldn't answer
It started on a client call in late 2025. We were running a brand portfolio for Paris Watch Company and their founder asked: "When someone asks ChatGPT for a watch recommendation, do we come up?" We had no idea. Nothing in our stack could answer that.
We checked manually. Sometimes yes, sometimes no. The citations were inconsistent across queries, sessions, and models. That was the moment — every commerce brand had the same blind spot.
Why we built it ourselves
Existing tools approached AI monitoring from the SEO side — bolt-on features that treated citations like another SERP feature. They couldn't answer operator questions: Which products? To whom? What's the revenue impact? We had the engineering team, the commerce context, and clients who needed answers yesterday.
What we got wrong
Mistake 1: We tried to measure "rankings" — positions in AI responses. AI doesn't work that way. We rebuilt around citation presence, context, and share. Mistake 2: We celebrated citation volume without tracking sentiment. Being cited as a cautionary example isn't a win. Mistake 3: We assumed blog content would drive citations. Product pages are cited 3-5x more in purchase-intent queries.
What the data revealed
Server rendering = 7x citation advantage. Brand authority compounds — each citation makes the next more likely. Entity naming inconsistency reduces citations 40-60%. FAQ schema provides 2-3x uplift. These aren't theories. This is what we measured across our portfolio of commerce brands.
What operators should take from this
Server-render product pages. Standardize your entity naming. Add FAQ schema to top pages. Invest in product pages, not just blog content. And measure — you can't optimize what you can't see.