Operational Methodology · Post-Adversarial Draft · v0.4 — March 2026
The Coetzee Liquidity Protocol
An Operational Model for Niche Commerce Authority in High-Velocity Search Markets
A structured methodology for dominating machine-readable commerce surfaces through attribute specificity and feed architecture — designed specifically for niche commodity markets where AI agents verify inventory before humans browse, and where the cost of missing a Boolean True is a lost transaction to a generalist marketplace.
Published by Erwee Coetzee — SEO Gurus · Cape Town, South Africa · CC BY 4.0
5
Operational Pillars
Rf
Binary Resilience Gate
+300%
Long-Tail Organic Spike — Feb 2026 Blackout
6
Rounds of Adversarial Stress-Testing
The Core Insight
In commodity markets, buyers do not audit — they route. The Coetzee Liquidity Protocol is the engineering specification for ensuring your inventory is chosen by the machine before the human reaches the comparison page. Authority is a gradient in high-friction markets. In high-velocity markets, it is a binary gate.
The Mechanism
Search Confidence in commodity markets is a function of five compounding variables — Liquidity, Use-Case Interpretability, Trust Floor, Buyer-Goal Fit, and Economic Viability — governed by a Resilience Floor that must pass as a binary precondition before the formula operates at all.
The Differentiator
SMEs do not beat Amazon by being more liquid. They beat Amazon by being liquid enough to qualify — and semantically superior in the narrow niche where the marketplace stays generic. The CLP defines exactly how to find that niche, own it, and know when to move on.
1. The Liquidity Hypothesis: Why Commodity Markets Need a Different Architecture
The CLP is predicated on a specific and testable behavioural distinction. In high-friction search environments, users audit. In low-friction, high-velocity environments, users — and in 2026, increasingly their AI agents — route. The routing decision happens before the human is involved.
Auditing Behaviour (High-Friction)
Search for a “bespoke diamond engagement ring Cape Town.” The buyer’s cost of error is high — financial, emotional, and irreversible. They verify credibility before enquiring. Trust is built over a long decision cycle. The Coetzee Convergence Framework (CCF) is the correct architecture for this market.
Routing Behaviour (High-Velocity)
Search — or instruct an agent — to “find a 3-person ultralight tent with 20D ripstop nylon, in stock, delivered Cape Town by Friday.” The buyer does not research; they verify attributes and transact. In 2026, an agentic layer often executes this query before the human sees a results page. The CLP is the architecture for this market.
This distinction has structural implications for the type of digital presence that wins a high-velocity transaction. A standard SEO approach focused on content relevance does not pass an agentic routing check. The CLP is the engineering protocol for a presence that does — and that survives when the routing infrastructure fails.
“In high-velocity markets, search engine confidence and agentic routing confidence are not separate problems. They are the same problem expressed at different speeds. The methodology that wins the feed wins the transaction.”
— Erwee Coetzee, CLP v0.4
2. What Is a High-Velocity Commodity Market?
A high-velocity market qualifying for the CLP is defined by three concurrent conditions — all of which must be present for the protocol to apply:
Condition 1 — Attribute-Driven Selection
The buyer’s decision is primarily determined by product attributes — specifications, compatibility, availability, and price — rather than by brand trust or relationship history.
Condition 2 — Machine-Routable Intent
The search intent is specific enough to be resolved by structured data. An AI agent or shopping surface can match the query to an SKU if the merchant’s product data is sufficiently granular. Generic intent cannot be routed; specific intent can.
Condition 3 — Margin-Viable Niche
The category gross margin is sufficient to fund the data maintenance cost required to sustain Attribute Certainty. Generic commodity categories — where price is the only variable and margins are sub-10% — do not qualify. The CLP requires a niche, not a category.
Qualifying Market Examples
- Specialist outdoor and hiking equipment (technical SKUs)
- Compatibility-dependent automotive and mechanical components
- Industrial replacement parts and tooling
- Specialised electronics accessories with fitment constraints
- Pet nutrition and veterinary dietary products
- Hunting optics and precision sport equipment
- Horticultural and agricultural speciality inputs
Where the CLP Does Not Apply
- True commodity markets — AA batteries, generic office supplies, undifferentiated consumables. Amazon wins these on volume economics. CLP fails below the Specificity Threshold.
- High-friction trust markets — luxury goods, professional services, bespoke manufacturing. Apply the Coetzee Convergence Framework instead.
- Markets where category margin cannot absorb a data maintenance cost of R20,000+ per year — the economic viability gate (Ec) will fail.
- Service businesses — hours and expertise cannot be exposed as machine-readable inventory unless the service is first productised into defined SKU-equivalent deliverables.
3. The Search Confidence Formula
The CLP formalises commodity Search Confidence (SC) as a governed equation. The Resilience Floor (Rf) is a binary precondition — if it fails, the formula produces zero regardless of the values of any other variable. This is not a metaphor; it is the architectural lesson of the February 2026 Merchant Centre outage.
If Rf = 1 → SC = (Lq × Ui × Tf) × Bg × Ec
— CLP v0.4 Master Formula
| Variable | Name | Definition | Role in Formula |
|---|---|---|---|
| Rf | Resilience Floor | Binary gate: can the merchant sustain a minimum revenue floor during a 14-day feed outage? Requires ≥30% of revenue independent of Merchant Centre feeds. | Precondition. If Rf = 0, SC = 0. |
| Lq | Liquidity Baseline | Accuracy and freshness of stock, pricing, and availability data across all feed surfaces. The table-stakes requirement for agentic discoverability. | Multiplier. Enables discovery. Does not differentiate. |
| Ui | Use-Case Interpretability | The degree to which taxonomy, attributes, and supporting content make a product unambiguously matchable to a narrow buyer intent. Niche taxonomy depth, compatibility data, and attribute specificity. | Primary moat variable. Where SMEs structurally outperform generalist marketplaces. |
| Tf | Trust Floor | Merchant reliability signals available in 2026 infrastructure: Google Merchant Centre account history, verified GBP presence, structured return and shipping policies, review signal integrity, domain consistency. | Binary pass/fail in practice. Must reach a minimum threshold to avoid agentic exclusion. |
| Bg | Buyer-Goal Fit | The merchant’s ability to satisfy a specific constrained intent better than a generalist marketplace. Measured by the alignment between the merchant’s niche taxonomy and the query cluster they target. | Amplifier. The answer to: why does an agent choose you when Amazon exists? |
| Ec | Economic Viability | Whether the margin captured by the niche exceeds the data maintenance cost. Profitability condition: (Sv × Cr × Cm + Cv + Ac) > Dc. | Business model gate. High Ui is worthless if the niche cannot fund its own data layer. |
“A niche is only viable if it wins both the search algorithm and the profit equation. Visibility without margin is not a strategy — it is a subsidy to the marketplace.”
— Signal Viability Principle, CLP v0.4
4. The Five Operational Pillars
The CLP is applied as a five-pillar architecture. Pillar I (Resilience Floor) is a structural precondition — no feed strategy is viable without it. Pillars II–IV are the discovery and dominance layer. Pillar V is the competitive defence discipline that maintains the moat over time.
Pillar I
The Resilience Floor (Rf)
Objective: Ensure the business can sustain a minimum revenue floor during a 14-day platform outage. The February 2026 Merchant Centre disruption demonstrated that feed-dependent merchants with no organic fallback lost up to 40% of monthly revenue. The Resilience Floor is not an optional resilience measure — it is the architectural precondition for the entire protocol.
Required components: crawlable product pages with organic ranking capability independent of feed inclusion; Product schema on every PDP mirroring all feed attributes; multi-platform feed distribution (Google, Bing, Meta); email subscriber and returning customer base. Minimum threshold: ≥30% of revenue not dependent on Merchant Centre.
Pillar II
The Liquid Tip (Lq)
Objective: Achieve and maintain feed accuracy sufficient to qualify for agentic routing consideration. Liquidity is table stakes — not the moat. Real-time or near-real-time stock and pricing signals pushed via the Content API for Shopping. Out-of-stock items are an authority killer in agentic surfaces: an agent that routes to an unavailable product does not route to that merchant again.
Stop treating the website as the primary destination. The Merchant Centre feed is the discovery surface. Every product attribute in the feed must also appear in on-page Product schema — this is the JSON-LD parity requirement that creates the Crawlable Shadow for Pillar I.
Pillar III
Attribute Certainty (Ui)
Objective: Return a Boolean True on compatibility and specification queries where the generalist marketplace returns a “Maybe.” In 2026, agentic routing is programmed to avoid ambiguity because the cost of returns is embedded in the agent’s confidence model. A merchant that answers “does this part fit a 2014 Jeep Grand Cherokee 3.0 CRD” with a verified Boolean True wins the routing decision over 50 sponsored results that might fit.
Attribute Certainty does not require building a proprietary database. For most SME niches, leasing specialist fitment or compatibility data via APIs (R500–R4,000 per month depending on vertical) is sufficient. The economic gate: four sales at 35% margin on an R8,000 AOV cover the annual data cost. Identify the niche before committing to the layer.
Pillar IV
Verified Trust (Tf)
Objective: Meet the Trust Floor required for agents and users to consider the merchant a credible transaction partner — without relying on speculative verification infrastructure. In 2026, the Trust Floor is built from available signals: Google Merchant Centre account history and policy compliance; a verified Google Business Profile; structured return and shipping policy markup; review integrity from a single consolidating source; and domain age and consistency.
An “Expert-Verified Returns” layer — where a qualified human checks specification fit before dispatch in technical categories — creates a trust signal that an agentic system prioritises because it minimises the Total Cost of Ownership risk embedded in the agent’s routing logic. This is the Trust Moat for compatibility-dependent products.
Pillar V
The Moving Taxonomy
Objective: Maintain a moving specificity frontier that extends taxonomy depth faster than generalist marketplaces can standardise the discovered attributes. Amazon’s feed team is reactive: the moment a niche generates visible transaction volume, they enrich their own attribute data for that category. The Moving Taxonomy is the discipline that ensures the SME’s Ui advantage survives marketplace discovery.
In practice: continuously extend category depth beyond the currently visible query cluster; add new compatibility attributes before they appear in marketplace schema standards; use customer query logs and support interactions to discover edge-case attributes that no external database yet formalises. The SME’s proprietary edge-case dataset — accumulated from real buyer behaviour — is the one data layer that cannot be purchased or leased by the marketplace.
5. The Specificity Threshold
The CLP does not apply universally to commodity markets. The Specificity Threshold defines the minimum conditions a product category must meet before the protocol generates a positive return. Below the threshold, Amazon wins on volume economics. Above the threshold, the SME can dominate on semantic depth.
The Threshold Assessment Criteria
| Criterion | Below Threshold | Above Threshold | Assessment Method |
|---|---|---|---|
| Attribute Complexity | Product is fully described in 3–5 generic attributes (size, colour, quantity) | Product requires 10+ attributes including compatibility, fitment, material spec, or certification data | Count required disambiguating attributes for a confident purchase decision |
| Marketplace Taxonomy Depth | Amazon or Takealot provides equivalent attribute granularity in their existing category schema | Marketplace listing for equivalent product returns ambiguous or incomplete attribute coverage | Manual audit of top 3 marketplace listings for the category |
| Category Gross Margin | Sub-15% — data maintenance cost cannot be amortised | ≥25% — data layer can be funded from category revenue | Category margin calculation against estimated Dc (data cost) |
| Long-Tail Query Specificity | Primary query clusters are 1–2 word head terms dominated by marketplace PPC | Primary query clusters are 5–10 word specification queries with technical vocabulary | Keyword research filtered for attribute-specific long-tail terms |
| Decision Risk | Wrong product is easily returned; no material consequence | Wrong product creates significant cost, delay, or operational failure — driving pre-purchase verification | Customer journey audit: does the buyer verify before transacting? |
The Economic Viability Calculation
Before committing to a niche, calculate the Net Moat against the Data Tax:
Net Moat = (Margin × ConversionAgent) − (Data Tax + Friction Cost)
— CLP Economic Viability Gate
Data Tax Reference Ranges (South Africa, 2026)
- Lightweight (internal matrix): R20,000–R80,000 per year. Suitable for small niche categories with limited SKU range.
- Licensed database (API subscription): R100,000–R400,000 per year. Suitable for automotive, technical equipment — categories with established fitment database ecosystems.
- Full data engineering (proprietary build): R500,000+ per year. Justified only in high-margin verticals where the proprietary dataset itself becomes a defensible asset.
Profitability Condition
The niche qualifies as a viable Attribute Certainty lane if:
(Sv × Cr × Cm) + Cv + Ac > Dc
Where Sv = monthly search demand, Cr = conversion rate improvement from Attribute Certainty, Cm = category gross margin per unit, Cv = customer lifetime value for repeat buyers, Ac = acquisition cost avoided through organic discovery, and Dc = annual data maintenance cost.
6. The Entropy Signal and Harvest / Pivot Rule
Every niche has a discovery half-life. The moment a niche generates sufficient transaction volume to become visible to marketplace analytics teams, the incumbent’s response begins. The Entropy Signal is the set of observable indicators that the niche’s Ui advantage is being eroded. The Harvest / Pivot Rule defines the required response before the moat is fully breached.
Entropy Signal Table
| Entropy Signal | Indicator | Required Action | Decision Window |
|---|---|---|---|
| Marketplace Attribute Parity | Amazon or Takealot adds your specific niche attributes to their core category schema. Their listings begin returning comparable attribute completeness on your primary query cluster. | Pivot: Deepen taxonomy immediately. Extend to the next specificity layer — sub-niche attributes, edge-case compatibility data, or application-specific bundles not yet in any catalogue standard. | 180 days from first observable parity signal |
| Margin Compression | CPC for niche keywords exceeds 30% of category gross margin. Organic click share for your primary query cluster begins declining versus previous 90-day baseline. | Harvest: Cease aggressive new customer acquisition spend. Redirect budget to LTV maximisation — email, repeat purchase optimisation, bundle architecture. The niche is becoming paid-media dependent. | Immediate on breach of the 30% CPC/margin threshold |
| Data Provider Concentration Risk | Your leased compatibility or fitment database provider is acquired by or enters a preferred partnership with a major marketplace competitor. | Exit / Build: Migrate immediately to an alternative provider. Simultaneously accelerate proprietary edge-case dataset development using customer query logs. The leased moat is now a shared moat. | 60 days from acquisition announcement |
| Query Cluster Commoditisation | Your long-tail specification query cluster begins attracting high-volume generic PPC from marketplace players. Head-term CPCs for adjacent category terms increase by ≥40%. | Brand / Community Shift: The query cluster is becoming a paid-media commodity. Shift visibility strategy to brand-search demand and direct channel development. The CLP moat in this cluster has expired. | 90 days from first significant CPC escalation |
“If a marketplace achieves 90% attribute parity with your niche taxonomy, the moat is breached. You have 180 days to either extend the specificity frontier by 20%, or pivot the revenue floor to brand and community loyalty.”
— CLP Harvest Rule, v0.4
7. Falsifiability Condition
The CLP predicts that implementing the full five-pillar architecture for a verified above-threshold commodity niche will produce measurable improvement in at least two of the following outcomes within 90 days of full implementation, measured against a documented pre-intervention baseline:
- Improvement in Google Search Console impressions and click share for the targeted long-tail specification query cluster
- Increase in Merchant Centre product approval rate and feed attribute completeness score
- Measurable organic traffic to product and category pages independent of feed-driven Shopping surfaces
- Improvement in qualified conversion rate from specification-query organic traffic versus head-term traffic baseline
The protocol is weakened if a niche merchant with materially better taxonomy depth, compatibility metadata, and feed accuracy still fails to outperform a generalist marketplace on the targeted long-tail intent cluster over a 90-day observation period. In such cases, the engagement should be reviewed for threshold misclassification or implementation failure before the protocol itself is invalidated.
Note: The CLP was developed and stress-tested through six rounds of structured adversarial audit in Session WBT-2026-03-08-ZA-PILLAR-2-STRESS-TEST and revised in response to valid findings at each stage. Empirical validation in a live commerce category is the identified next phase.
Relationship to the CCF
The CLP and the Coetzee Convergence Framework are not competing models. They address structurally different search economies and are selected based on market-type diagnosis, not preference.
| Dimension | CCF | CLP |
|---|---|---|
| Market Type | High-friction, high-trust | High-velocity, high-specificity |
| Buyer Behaviour | Auditing | Routing |
| Primary Signal | Authority and credibility | Inventory interpretability |
| Moat Variable | Entity signal convergence | Attribute Certainty (Ui) |
| Platform Risk | Low — not feed-dependent | High — requires Rf gate |
| Competitor | Trust-deficit operators | Generalist marketplaces |
The correct diagnostic question before selecting a framework is: does the buyer in this market audit or route? If they audit, apply the CCF. If they route, apply the CLP. If the market contains both buyer types at different intent stages, a hybrid architecture is indicated.
Frequently Asked Questions
Can a small South African e-commerce store genuinely beat Amazon on a product category?
Yes — in a constrained niche where attribute complexity exceeds marketplace taxonomy depth and category margins justify the data layer. The CLP does not claim SMEs can beat Amazon on generic commodity economics. They cannot. The protocol is specific to above-threshold niches where Amazon’s catalogue is a consensus of third-party noise and the SME’s single-source attribute certainty produces a higher agent confidence score on the specific match. Below the threshold, Amazon wins. Above the threshold, the SME can dominate — for as long as the Moving Taxonomy discipline is maintained.
What is the Crawlable Shadow?
The Crawlable Shadow is the on-site semantic layer that mirrors every product attribute present in the Merchant Centre feed as structured Product schema on the corresponding Product Detail Page. When the feed is operational, this layer reinforces feed signals and contributes to organic shopping surfaces. When the feed is disrupted — as occurred in the February 2026 Merchant Centre outage — the Crawlable Shadow allows Google’s secondary organic crawl to still surface the product in blue-link results. During the February outage, long-tail organic traffic for technical SKUs spiked by 300% as users reverted to organic search. Merchants with a functioning Crawlable Shadow captured that traffic. Feed-only merchants did not.
Do I need to build a proprietary compatibility database?
For most SME niches, no. The CLP recommends leasing specialist fitment or compatibility data via established API providers as a cost-effective entry point — the Data Tax is then a monthly operating expense rather than a capital project. The critical risk to model is data provider concentration: if the provider you lease from is acquired by a marketplace competitor, your leased moat becomes a shared moat. For this reason, the CLP recommends supplementing any leased dataset with a proprietary edge-case layer built from customer query logs and support interactions. This layer — accumulated from real buyer behaviour in your specific niche — is the one data asset that cannot be purchased by a marketplace.
How does the Resilience Floor work in practice?
The Resilience Floor is a binary gate, not a gradient. The minimum operational requirement is that ≥30% of revenue can be sustained without Merchant Centre feed distribution — through direct organic traffic, returning customers, email channels, or alternative shopping feeds (Bing, Meta). Achieving this floor requires maintaining indexable product pages with Product schema, a category taxonomy that ranks organically for specification queries, and a minimum direct audience. An SME that implements the full CLP feed architecture without first establishing the Resilience Floor is building on a single point of failure. The feed is the spear; the semantic site is the shaft. Both are required.
Is the CLP peer-reviewed?
Version 0.4 is published as a post-adversarial draft. It underwent six rounds of structured red-team stress testing in a documented session (WBT-2026-03-08-ZA-PILLAR-2-STRESS-TEST), with the framework revised in response to valid kills at each stage — including the cold-start problem, the Data Tax challenge, the platform concentration risk, and the absence of a moat defence mechanism. The current version represents the framework that survived those kills. Empirical validation in a live commerce category against a documented baseline remains the next phase. Practitioners who apply the protocol and observe results are invited to submit case observations for inclusion in the evolving validation dataset.
Apply the CLP to Your Commerce Operation
If your e-commerce operation sells in a niche where attribute complexity exceeds marketplace taxonomy depth — and your current feed architecture is not structured to win that comparison — request a CLP Specificity Threshold Assessment. We will identify whether your category qualifies, model the economic viability gate, and define the five-pillar implementation roadmap.
The CLP is published under Creative Commons Attribution 4.0. You may cite, reference, and build upon it — with attribution to Erwee Coetzee, SEO Gurus, March 2026.