GEO INSIGHT
SUMMITFEED Website GEO Before-and-After Report: Comparing JSON-LD Implementation
We compare the code before and after applying JSON-LD on the SUMMITFEED website with Git history and current operating code, and explain the structure of Organization·WebSite·WebPage·Service·FAQPage.
Quick answer
Before the change, the website rendered the screen UI, but there was no JSON-LD insert code to describe the website entities.
The current code concatenates Organization, WebSite, WebPage, BreadcrumbList, Service, and FAQPage with the absolute path ID and actual screen data.
The before-and-after comparison was not an estimate, but was based on the app/page.tsx and current main code in the Git commit, and only publicly available code was included in the capture.
The key change is not to increase the number of schemas, but to change the relationship between brand, site, page, service, and FAQ to convey it consistently in the initial HTML.
Before change
JsonLd, schema.org, application/ld+json connection code 0
After change
Prints 6 types of Schema.org as JSON-LD on website
Verification criteria
Comparison of actual files in Git history and current operating main code
Interpretation Principles
This is to strengthen the basis for understanding the site and does not guarantee exposure or citation performance.
Report scope and comparison criteria
This report records the differences between the operating code before and now the GEO related structured data is linked to the SUMMITFEED website. The comparison target is app/page.tsx in the repository, the baseline before the change is commit e519dc0 on July 10, 2026, just before the introduction of JSON-LD, and the baseline after the change is main commit c2cf075 on July 24, 2026.
We distinguished between confirmed facts and interpretation. What we have confirmed is the import, Schema.org type, @id, URL, and rendering code that exist in each point-in-time file. Since the weighting of this structure by search engines or generative AI is not disclosed, it cannot be assumed to be the direct cause of a specific ranking or citation result.
| category | reference point | check file | How to check |
|---|---|---|---|
| Before change | 2026-07-10 · e519dc0 | app/page.tsx | JsonLd·schema.org·application/ld+json Check absence of string and insertion code |
| After change | 2026-07-24 · c2cf075 | app/page.tsx | Check the current JSON-LD array and JsonLd component insertion code |
The commit hash is a reproducible basis for comparison; subsequent changes to the code may differ from the current state.
Before change: There was screen UI, but no website entity descriptions.
Before the change, app/page.tsx had a structure that started with use client and directly managed the screen status and interactions. There was a function to display the screen, but the connection that outputs the relationship between the official company, site, page, service, and FAQ as JSON-LD was not confirmed.
This condition should not be interpreted directly as not crawlable. Plain HTML text and links may be collected separately. However, this means that there was no explicit entity graph that allowed the machine to interpret which organization's official information was the information appearing on the page and which service it was connected to.
After change 1: Organization and WebSite were connected by absolute path ID.
The current code creates unique @ids for organization, website, and webpage based on https://www.summitfeed.co.kr. Organization specifies SUMMITFEED and SUMMITFEED notation, official URL, inquiry email, logo absolute path, and main professional topic.
WebSite connects publisher to organizationId. Instead of writing the same company name or URL repeatedly in multiple schemas, you can use @id to reference a single organizational entity, allowing you to express your brand's relationship with your site more consistently.
After change 2: Insert WebPage·Service·FAQPage into the initial HTML
WebPage references WebSite by isPartOf and Organization by about. The Service connects the provider of GEO analysis and execution to the same Organization, and FAQPage generates questions and answers from faqItems displayed on the website screen.
The JsonLd component outputs this array to the application/ld+json script. Structured data can be read independently of whether JavaScript is running or not, as it is configured to be checked in the initial HTML returned by the server rather than in the values it creates after waiting for client interaction.
6 schemas and roles currently applied to the website
The current website uses 6 types: Schema.org. Having a lot of types is not the goal in itself. The key is to express only information that actually exists on the screen and connect the relationships between organizations and websites, pages and services using the same @id system.
| Type Schema.org | Information currently expressed | complementary relationship |
|---|---|---|
| Organization | Company name, English name, official URL, email, logo, professional topic | Another schema references #organization |
| WebSite | Official website name, URL, language | publisher refers to Organization |
| WebPage | website title, description, language | isPartOf refers to WebSite, about refers to Organization |
| BreadcrumbList | Home location of current page | Represent home URL as a list item |
| Service | GEO Analysis and Execution Services and Areas Served | provider refers to Organization |
| FAQPage | Questions and answers displayed on the website | Use the same faqItems data as your screen |
Structured data must be kept consistent with what users can actually see on screen.
What has changed from GEO perspective
Before the change, the brand name and service description existed only in screen context. After the change, the relationship between the official organization, site, website, and service was clearly expressed, and the company name and logo URL were also unified under one standard.
This structure is the foundation that search systems and AI-based collection systems can use to interpret the subject and topic of a page. However, adding JSON-LD alone does not automatically improve the accuracy of your content, external sources, site reputation, crawlability status, or question relevance.
Therefore, in GEO operation, robots.txt, sitemap.xml, canonical, internal links, screen body, author, publication and modification dates, and external sources must be checked along with structured data. After application, actual search results and changes in mentions and citations for each question must be measured separately.
| Check items | Apply code | Additional information to check in operations |
|---|---|---|
| brand entity | Link Organization and official URL | Company name, logo, and contact information match throughout the site |
| content meaning | Expression WebPage·Service·FAQPage | Match screen body and schema content |
| collection path | Print JSON-LD in initial HTML | Check robots.txt·canonical·sitemap·status code |
| performance judgment | Schema validation | Separately measure exposure, mention, citation, and recommendation by question |
Pre-publication verification checklist
Structured data does more than just verify that code exists. You must also check the initial HTML of the final operational URL, matching schema values and screen content, absolute path asset accessibility, and duplicate entities.
| Check items | How to check | Based on current report |
|---|---|---|
| Includes initial HTML | Check application/ld+json in page source | Check JsonLd component insertion code |
| Official URL | Check whether url·@id·logo is an absolute https path | summitfeed.co.kr Use absolute path |
| Match screen and schema | Compare company name, service, and FAQ with screen | FAQs use the same faqItems data |
| Entity Association | Check @id of publisher·about·provider | See #organization and #website |
| Performance interpretation | Separate structure application and exposure performance | Do not use ranking/AI citation guarantee expressions |
JSON-LD Operation Checklist
- ✓ Make sure the company name, English name, official URL, email and logo are consistent throughout the site.
- ✓ Check whether the URL including the logo and canonical is an absolute https path.
- ✓ Ensure that services and FAQs in structured data are present in the actual screen body
- ✓ Use @id to check whether the relationships Organization, WebSite, and WebPage are not broken.
- ✓ After operational deployment, check the page source to see if application/ld+json is included in the initial HTML
- ✓ Do not express JSON-LD application as search ranking or AI citation guarantee
Frequently asked questions
If I enter JSON-LD, will it be cited directly in AI search?
no. JSON-LD is a foundation that helps interpret the entities and relationships on the page. Actual mentions or citations depend on a number of conditions, including suitability of the question, quality of the text, collectability, reliability of the source, timing and platform, etc.
Were search engines unable to read the website at all before the change?
You can't be so sure. Plain HTML body and links could be collected. The difference identified in this report is the presence or absence of an explicit JSON-LD link code that describes the website entity.
Why do I write the company name and logo URL as absolute paths?
This is to clearly express the location of official organizations and assets in one address system, and to refer to the same entity regardless of page location.
FAQPage Can I add questions that are not on the screen?
It's safer to base your questions and answers on the actual questions and answers users can see on the screen. The SUMMITFEED website uses the same faqItems data as the screen and JSON-LD.
Is the more schemas the better?
no. You should only use information that actually exists and can be described on the page. Entity relationships, accuracy of values, and consistency with screen content are more important than number of types.
Conclusion: Structuring is the basis of GEO's operations, not a performance guarantee device
The biggest change in the SUMMITFEED website is not the number of schemas (from 0 to 6). Official organizations, websites, homepages, services and FAQs are linked by a single URL and @id system and can be read from early HTML.
This structure communicates your brand and page meaning more clearly, but by itself does not guarantee search rankings or generative AI citations. It must be managed together with accurate text, consistent brand information, collectible technical structures, external sources and repeatable measurements.
SUMMITFEED operates the GEO structure by not only checking whether or not the code is applied, but also checking the consistency of screen information and schema, collection path, and changes in exposure for each question.
Tags
Continue reading
Sources and references
Search results and generated AI answers may vary depending on platform, question, timing, and user conditions, and do not guarantee specific exposure, ranking, or citations.