Exp-004 Planned D Activates 2026-09-15
Entity stacking using legitimate Technology Partner pages
Research question Do transparent Technology Partner / attribution pages on third-party sites improve discoverability and correct entity association for the credited vendor, compared to sites without such pages?
Motivation
“Entity stacking” is sometimes used loosely to justify networks of thin partner pages. This lab tests the narrower claim: transparent, accurate attribution pages on real client/partner properties.
Facts
- Many vendors publish Technology Partner or Built by pages.
- Search systems historically treat manipulative doorway / partner spam as policy violations (Google spam policies); this experiment does not implement those patterns.
- Vendor under study: InfoWebPlus; Person platform: barbu.es.
Assumptions
- Honest attribution can still be a discoverability signal when content is unique and relationship-true.
- AI systems may cite partner pages when asked who built or supports a site.
Unknowns
- Threshold at which partner-page volume becomes noise or risk.
- Whether AI citations prefer partner pages over the vendor’s own domain.
Hypothesis
H₀: Adding legitimate Technology Partner pages does not increase correct vendor association or citation of the vendor’s canonical site in AI probes.
H₁: Transparent partner pages increase (a) correct vendor naming for the host site and (b) citation of the vendor canonical URL, without increasing false partner claims elsewhere.
Methodology
Design type
Matched comparison across consenting properties: attribution page present vs absent (or before/after on the same property).
Eligibility criteria (all required):
- Real commercial or editorial relationship exists.
- Page discloses the relationship accurately.
- Page is reachable from the site (nav, footer, or about).
- No fabricated testimonials or fake partner logos.
Environment
| Dimension | Value |
|---|---|
| Primary domains | Consenting client/partner sites + InfoWebPlus canonical |
| Surfaces under test | /partners/, /technology/, footer credits, case-study asides |
| AI systems probed | ChatGPT, Claude, Gemini, Perplexity |
| Measurement window | ≥ 30 days post-publication per property |
| Geographic / language scope | Per property language; log language of probes |
| Tools used | Page inventory, crawl checks, probe logs |
Variables
Independent variables
- Presence of Technology Partner page
- Depth (stub vs substantive: scope of work, stack, dates)
Dependent variables
- Correct vendor association rate for “Who built/supports [SITE]?”
- Canonical vendor URL citation rate
- False association rate on unrelated sites (safety metric)
Controlled / held constant
- No PBNs, expired-domain networks, or mass identical page templates across unrelated hosts without disclosure of templating
Confounds to monitor
- Concurrent PR / launches
- Client site authority differences
- Exp-003 chain reinforcement overlapping the same URLs
Procedure
- Select N≥3 consenting properties; document relationship evidence.
- Baseline probes without partner page (or with credit footer only - record which).
- Publish partner page variant A (minimal) or B (substantive); randomize if multiple.
- Verify indexability (no accidental
noindex). - Probe on schedule; include negative controls (“Who built [unrelated site]?”).
Prompt protocol
Who provides technology or SEO services for [HOST_SITE]?
Is [VENDOR] associated with [HOST_SITE]? Cite sources if possible.
Name technology partners of [HOST_SITE], if any.Results
Status: Planned. No results claimed.
| Metric | Baseline | Post-intervention | Notes |
|---|---|---|---|
| Correct vendor association | n/a | n/a | Not collected |
| Vendor URL citation rate | n/a | n/a | Not collected |
| False association rate | n/a | n/a | Not collected |
Observations
- Pending partner selection.
Limitations
- Small N; site-level confounds dominate.
- “Entity stacking” effects may be SEO-community folklore; this experiment may fail to detect lift even if traditional crawlers react differently.
- Policy risk if templates are abused - mitigated by eligibility criteria.
Future Work
- Compare footer-only credit vs dedicated partner page.
- Measure human trust and AI citation as separate outcomes.
References
- Google Search spam policies (doorway / affiliate abuse boundaries)
- Organization / attribution types: schema.org/Organization, schema.org/Brand where applicable
- Vendor canonical: InfoWebPlus; Person platform: barbu.es
- Lab Exp-003 (supporting entity chain)
- Lab Exp-010 (case studies vs generic articles)
Replication Notes
Reject any implementation that would mislead users or invent partnerships. Publish page URLs and relationship evidence with results.
Status: Planned
Author: George Barbu, Founder of InfoWebPlus