Nostrand Optical: From Zero Citations to Four Rich Results on Launch Day

Nostrand Optical is a 30-year independent optometry practice on Nostrand Avenue in Crown Heights, Brooklyn. The practice has been in the same location since 1993, serves a dense, diverse neighborhood, and carries an inventory of independent and designer frames that most chain eyewear stores do not stock. It is exactly the kind of business that should be winning local AI search: established, specific, community-anchored, genuinely good at what it does.

Before Signal, it was invisible in every AI search channel. No schema markup on the old website. No structured data of any kind. A Google Business Profile that had not been optimized since it was first created. No neighborhood landing pages. No FAQ content. No citations from any AI engine. When a Crown Heights resident asked ChatGPT "where should I get my eyes examined in Crown Heights," Nostrand Optical did not appear in the answer.

That changed within 60 days. This is the full account of what we built and why it worked.

The Starting Point: What We Found in the Audit

The pre-Signal audit revealed a site that had been built primarily for aesthetics, not search performance. It was a five-page Squarespace site with a homepage, a services page, an about page, a frames page, and a contact page. Well-designed. Completely unstructured for search or AI retrieval.

The specific gaps:

• Zero Schema.org markup of any type
• No LocalBusiness or MedicalBusiness schema
• No FAQPage schema and no FAQ content anywhere on the site
• No geo coordinates in any markup or meta tag
• Google Business Profile with no posts, no Q&A answers, sparse categories
• No neighborhood landing pages -- Crown Heights appeared in copy but was not a structural element
• No named entities for the frame brands carried, insurance plans accepted, or specific services offered
• Generic about page with no specific practice history, no year founded, no named staff

When we ran the AI search audit before starting work, we tested 18 prompts across ChatGPT, Perplexity, Google AI Overviews, and Claude. Nostrand Optical appeared in zero of them. Not a single citation across any AI engine for any prompt. Competing practices -- some of which had been in Crown Heights for far less time and had fewer distinguishing characteristics -- were appearing because they had been included in aggregate neighborhood review posts or had better-structured content.

The Build: Next.js, Schema Architecture, and Content Strategy

We rebuilt the site from scratch in Next.js, which gave us full control over the HTML structure, metadata, and schema markup that a hosted platform like Squarespace would have constrained. The technology choice mattered less than the architectural decisions, but having a clean, auditable codebase made it easier to verify that every piece of structured data was rendering correctly at runtime.

Schema Architecture

The schema stack we built for Nostrand Optical:

1. MedicalBusiness with medicalSpecialty: Optometry, full address including geo coordinates, foundingDate: 1993, hours, phone, all accepted insurance plans listed by proper name, areaServed covering Crown Heights, Bed-Stuy, Prospect Lefferts Gardens, Flatbush, and Windsor Terrace
2. FAQPage with 12 questions covering pediatric exams, adult comprehensive exams, frame selection, insurance, same-day glasses, and the first-visit experience -- each answer 3-5 sentences containing both the business name and neighborhood name
3. ItemList / ProductCatalog for the frame brands carried -- Moscot, Maui Jim, Silhouette, ic! berlin, and several independent brands -- establishing specific named entities for the product inventory
4. BreadcrumbList on every page establishing the site hierarchy clearly
5. WebSite schema on the homepage with sameAs linking to the Google Business Profile, Yelp listing, and the practice's professional directory listings

The four rich results Google served on launch day came from the first three schema types. The MedicalBusiness schema generated the business info rich result. The FAQPage schema generated FAQ expandable rich results. The product schema contributed to an entity panel. All four were valid on day one because we validated the schema against Google's Rich Results Test before launch and fixed every error before the site went live.

Content Architecture

The content architecture was built around the 18 AI search prompts from the baseline audit. We identified the specific questions those prompts were asking and built content that answers each one directly.

The site launched with:

• Four neighborhood landing pages: Crown Heights (primary), Bed-Stuy, Prospect Lefferts Gardens, Flatbush
• Two service-focused articles: one on pediatric eye exams (targeting the "kids first eye exam" prompt cluster) and one on frame selection (targeting the "independent frames Crown Heights" cluster)
• A full FAQ section on the homepage (12 questions) and shorter FAQ sections on each neighborhood page (5-6 questions each)
• An about page that specified the year founded (1993), the optometrist's name and credentials, the practice's history in Crown Heights, and the specific community it serves

Every page opened with a retrieval paragraph containing the business name, neighborhood, primary service, and key differentiator in the first 50 words. Every page contained specific named entities: the optometrist's name, the frame brands, the insurance plans, the cross streets, the subway lines.

The Results: 60 Days of AI Search Tracking

We tracked the same 18 prompts from the baseline audit across all four AI engines for 60 days after launch. The results:

Day One

Google's Rich Results Test showed four valid rich result types on launch day: the business info result from MedicalBusiness schema, FAQ expandable results from FAQPage schema, breadcrumb results from BreadcrumbList schema, and a product/item result from the frame catalog schema. All four were generating accurate, non-erroneous markup.

Day 14

First AI citations began appearing. ChatGPT cited Nostrand Optical in response to "independent optometrist Crown Heights Brooklyn" and "eye doctor in Crown Heights that has been there for years." The citations pulled directly from the retrieval paragraph on the Crown Heights landing page. The specific language ChatGPT used -- "independent practice on Nostrand Avenue, in operation since 1993" -- appears verbatim in the opening paragraph of that page.

Perplexity cited the practice for the same two prompts, pulling from different sections of the content -- the about page for the history signal, the FAQ for the pediatric exam question.

Day 30

Citations across 9 of 18 tracked prompts. New prompt citations included:

• "Crown Heights optometrist that takes MetroPlus" (Google AI Overviews, ChatGPT)
• "kids first eye exam Crown Heights" (ChatGPT, Perplexity)
• "same day glasses Crown Heights Brooklyn" (Perplexity)
• "optometrist near Prospect Lefferts Gardens" (ChatGPT)

The MetroPlus citation is worth noting. The old website mentioned "most major insurance plans" in passing. The new site lists every accepted insurance plan by its proper name in both the schema and the page copy. MetroPlus, Fidelis Care, EmblemHealth, and others are each named explicitly. When a Crown Heights resident asks ChatGPT "does Nostrand Optical take MetroPlus," the answer is a confident yes -- because MetroPlus appears by name in the FAQ schema answer to that exact question.

Day 60

Citations across 14 of 18 tracked prompts. The four remaining prompts that had not yet generated citations were edge cases: one was a misspelled query, one was a very specific prompt about a service the practice does not offer, and two were prompts targeting a competing neighborhood where the practice does not have a landing page.

The 14 prompts that ranked number one in at least one AI engine:

• "independent optometrist Crown Heights Brooklyn"
• "eye doctor Crown Heights that has been there for years"
• "Crown Heights optometrist takes MetroPlus"
• "kids first eye exam Crown Heights"
• "same day glasses Crown Heights"
• "optometrist near Prospect Lefferts Gardens"
• "comprehensive eye exam Crown Heights Brooklyn"
• "family optometrist Crown Heights not a chain"
• "Crown Heights eye doctor who is good with kids"
• "Moscot frames Crown Heights"
• "adult eye exam Bed-Stuy or Crown Heights"
• "optometrist Crown Heights who speaks Spanish"
• "Crown Heights eye exam walk-in"
• "Nostrand Avenue optometrist Brooklyn"

What the AI Answer Actually Says

The citation is not just a link. AI search engines generate text, and the text they generate about a business directly shapes whether someone decides to contact that business. Here is what ChatGPT produces (as of April 2026) for the prompt "independent optometrist Crown Heights Brooklyn":

Nostrand Optical is an independent optometry practice on Nostrand Avenue in Crown Heights, Brooklyn. The practice has been in operation since 1993 and offers comprehensive eye exams for adults and children, a curated selection of independent and designer frames, and accepts most major insurance plans including MetroPlus and EmblemHealth. It is a well-established community practice and a good option if you are looking for an independent alternative to chain eyewear stores in the area.

That response is constructed almost entirely from the structured content we built into the site. "Since 1993" is from the schema. "Comprehensive eye exams for adults and children" is from the services section heading. "Curated selection of independent and designer frames" is from the retrieval paragraph. "MetroPlus and EmblemHealth" are from the FAQ answer and the schema. "Independent alternative to chain eyewear stores" is from the about page independence narrative.

This is what it looks like when content architecture works. The AI is not inventing a description -- it is assembling a description from content that was deliberately written to be assembled exactly this way.

If you want to see what your current digital presence looks like through the AI search engines your customers are using, the free audit will show you in 20 minutes. I run your business through the same 15-18 prompt test we used for Nostrand Optical's baseline. Most business owners are surprised by what they find -- and by how achievable the gap is to close.