Three Glossaries, One SERP: How a Thin Page Inside a Strong Entity Architecture Ranked
An IBM FlashSystem search turned up three glossary pages across three domains, one of them barely more than two blurbs and a CTA. The content depth didn't match the result, and that's what killed the glossary-schema theory.
Content depth across the three results wasn't remotely comparable, which undercuts a pure glossary-schema explanation.
The claim, evidence, limits, and open loop.
Active experiment ... baseline contaminatedWhat was true before the work
One IBM FlashSystem search surfaced three glossary pages tied to Krisada's work across three domains. One page contained only two short blurbs and a CTA, so page depth did not match the observed result.
What changed
The three implementations were compared directly. The thin FlashSystem page was then rebuilt with deeper content, structured data, and internal links under the Entity Boot Camp intervention.
What the evidence showed
The observation weakened the original glossary-format theory and produced a narrower working hypothesis: strong entity context may help explain why a thin page surfaced. The intervention has not yet proved that hypothesis.
What Google reported
The starting signal was a live Google result containing all three glossary pages. Historical search and warehouse data are being used as the closest available baseline because the page was improved before a clean screenshot and frozen control were captured.
What the server logs reported
No server-log evidence is used to support the ranking explanation in this case. Bot access and ranking causation remain separate questions.
What this does not prove
The control was contaminated immediately after discovery. One shared result page cannot isolate entity architecture from links, history, domain relevance, or other ranking systems.
Where the case stands now
The case remains open and is tracked through the Portfolio Entity Architecture SERP Co-Occurrence experiment and the FlashSystem Entity Boot Camp proof entry.
What gets measured next
Compare the rebuilt page's recurring visibility against the historical baseline and the older Midland page without treating one screenshot as a durable ranking change.
What was measured.
Three readings from a single day observation. The third one limits what the first two can claim, which is why it is here rather than in a footnote.
The weakest of the three held its position carrying two short blurbs and a call to action. That page is the reason this case exists.
The stronger explanation is the architecture the thin page sat inside, not the glossary schema or anything on the page itself.
One observation with a compromised control suggests something. It does not establish it. A single natural control remains untouched, which is why the experiment is still marked active.
Stated as a limit rather than buried, because an experiment that loses its control and does not say so is worth less than one that never ran.
The short version.
This is an active, open-ended case study. It documents an IBM FlashSystem SERP where three glossary pages tied to Krisada's work appeared together across as400ibmsystems.com, the flashsystem.as400system.com subdomain, and midlandinfosys.com. Content depth across the three pages was wildly inconsistent, which weakened the initial glossary-architecture theory and pointed toward entity architecture as the more likely explanation. The thin page has since been rebuilt as a live intervention, tracked separately as Entity Boot Camp.
Explain it to me like I'm ten.
We searched Google for an IBM storage term and noticed three of our glossary pages showing up together, even though one of them barely had any content. That made us think it's less about how good one page is and more about the whole website, and the sites around it, clearly being about that topic.
What was the system?
The observation started with a routine IBM FlashSystem search. Three results connected to Krisada's own work appeared together: "IBM FlashSystem Glossary ... AS400 IBM Systems" at as400ibmsystems.com/glossary/ibm-flashsystem/, "FlashSystem Glossary ... AS400System" at flashsystem.as400system.com/glossary/flashsystem/, and "IBM FlashSystems Glossary ... Midland Information Systems" at midlandinfosys.com/about/glossary-of-ibm-iseries-hardware-terms/ibm-flashsystems. The first instinct was a glossary-architecture theory: maybe schema, maybe definition-oriented content, maybe something intrinsically glossary-related that Google rewards.
What changed?
The three implementations were compared directly. AS400IBMSystems.com's glossary entry had recently been substantially built out. The flashsystem.as400system.com entry was extremely thin at the time it was observed, two short blurbs and a CTA. The Midland Information Systems entry came from an older glossary and directory architecture that hasn't been actively optimized in years, and none of its content is unusually deep either. Comparing content depth across the three pages against a single-variable theory (glossary schema or format) was the test: if the theory held, depth and implementation should track with each other. They didn't.
What did the evidence show?
Content depth across the three results wasn't remotely comparable, which undercuts a pure glossary-schema explanation. The most interesting result turned out to be the weakest page: flashsystem.as400system.com/glossary/flashsystem/, a FlashSystem-specific subdomain, on an AS400-focused domain, with a FlashSystem glossary entry, sitting inside a growing network of connected IBM and AS400 properties. The page-level content was weak. The entity-level context around it (subdomain naming, domain topical focus, and the surrounding IBM/AS400 property network) was strong. The Midland result reinforces the same read from the opposite direction: an old, unmaintained implementation with thin directory content still surfaces for important IBM and AS400 categories, with nobody actively optimizing it.
What can be reused?
The revised working hypothesis is that entity architecture, not page-level content or glossary schema specifically, is the stronger explanation for why a thin page held a position in a competitive SERP. That doesn't establish causation from a single observation. The formerly thin flashsystem.as400system.com page has been substantially rebuilt (more content, semantic depth, structured data, internal links) under the name Entity Boot Camp, specifically to test whether combining that entity architecture with a strong document changes the outcome. Note: the control was contaminated before a baseline screenshot was taken, since the page was improved immediately after being noticed. Historical search and warehouse data are being used as the closest available baseline instead.
Keep Browsing the Case Library
Move from this evidence file back into the full proof system.