This is the maintenance layer of the BlitzMetrics Content Factory — the SOP for auditing, improving, and recursively updating the SOPs themselves. It sits alongside the Definitive Article Guide (which explains how to build the SEO tree), the Meta-Article Prompt (which documents how each article was made), and the Blog Posting Guidelines (which govern execution). This article covers what happens after content is published — how we keep the tree healthy, how agents learn from themselves, and how knowledge stays portable across any AI platform.
Version 1.0 — March 2026 — BlitzMetrics Content Factory
The Problem This Article Solves
BlitzMetrics has three SOPs that govern content creation: how to structure a definitive article, how to document the process behind each article, and how to execute the publishing workflow. Together, these form a recursive learning system — agents create content, document how they created it, and future agents learn from that documentation.
But there is a missing layer. None of those three documents address what happens after content goes live. Articles drift. Someone adds a video about Pilot Plumbing to an article about ARDMOR Windows. Someone embeds five tangentially related videos into what is supposed to be a definitive guide. The SEO tree grows branches that point in random directions, and nobody notices because there is no audit cycle.
This article fills that gap. It is the SOP for maintaining the system that creates SOPs — the truly recursive layer that makes the Content Factory self-correcting rather than just self-replicating.
The Five Maintenance Processes
Content maintenance at BlitzMetrics operates through five interconnected processes. Each one addresses a specific failure mode we have observed in production.
Process 1: The Quarterly Definitive Article Audit
The Definitive Article Guide contains a status table that tracks every major BlitzMetrics concept with green, yellow, or red indicators. The problem is that a green checkmark only means an article met the standard when it was created. It does not mean the article currently meets the standard.
Every quarter, each definitive article must be audited against four criteria. First, topic coherence: does every element on the page — every video, image, link, and paragraph — directly support the article’s declared topic? This is the SEO Tree test. If a video about Company A appears in an article about Company B, the article fails. If an embedded video covers a general concept that is only tangentially related to the specific topic, the article fails. The standard from the SEO Tree is absolute — every leaf must connect to its branch, and every branch must connect to the trunk.
Second, structural completeness: does the article still follow the eight-step framework from the Definitive Article Guide? Has the table of contents drifted from the actual content? Are the internal links still pointing to live pages? Are the schema markup and Rank Math settings still accurate?
Third, information currency: has new data from Google Business Profile, Ahrefs, client Zoom calls, or campaign results invalidated or supplemented what the article says? A definitive article on Dollar a Day that does not reference a successful campaign from the last quarter is falling behind.
Fourth, cross-reference integrity: do other articles that link to this definitive article still accurately describe what it contains? If a meta-article says “see the Dollar a Day guide for campaign setup instructions” but those instructions have been moved or removed, the cross-reference is broken.
After each audit, the status table in the Definitive Article Guide must be updated with the audit date and current status. A new column — “Last Audited” — tracks when each article was last reviewed. Green means it passed all four criteria in the most recent audit. Yellow means it passed with minor issues that are scheduled for correction. Red means it failed one or more criteria and needs immediate attention.
Process 2: The SOP Update Protocol
When an agent or human encounters a recurring problem that is not addressed by the existing SOPs, they need a formalized way to propose an update. Currently, SOP refinement happens when Dennis or someone senior notices a pattern and manually updates the documents. This creates a bottleneck and means valuable operational knowledge gets lost between the moment someone discovers a problem and the moment a senior person has time to address it.
The SOP Update Protocol works as follows. When an agent identifies a gap — for example, “I keep finding articles with videos from unrelated companies and there is no documented standard for which videos belong on which article” — the agent creates an SOP Amendment Proposal. This is a short document, no more than 500 words, that describes the problem encountered, identifies which existing SOP should be updated (or whether a new SOP is needed), proposes specific language to add or modify, and provides at least two examples from real articles where the gap caused a quality issue.
Amendment proposals are tagged with the SOP they affect and placed in a review queue. A senior team member reviews proposals weekly. Approved amendments are integrated into the relevant SOP with a version number increment and a changelog note. Rejected proposals receive a brief explanation of why the existing SOP already covers the case or why the proposed change would create conflicts.
This is how the system learns from itself. The agents doing the work are the ones who discover the edge cases, and the protocol gives them a structured way to feed those discoveries back into the system without requiring senior intervention for every small improvement.
Process 3: Knowledge Capture From Live Interactions
BlitzMetrics generates knowledge through Zoom calls with clients, analysis of Google Business Profile data, Ahrefs audits, campaign performance reviews, and dozens of other live interactions every week. This knowledge is currently trapped in recordings, screenshots, and team members’ heads. The pipeline from “we learned something new” to “that knowledge is documented in the right definitive article” does not have a formal process.
The Knowledge Capture Pipeline operates on a simple principle: every interaction that produces a reusable insight must generate a Knowledge Capture Note within 24 hours. A Knowledge Capture Note contains the source (which call, which data source, which campaign), the insight (what we learned, stated in one to three sentences), the destination (which definitive article or SOP should be updated), and the priority (does this change something we currently tell clients, or is it supplementary).
Agents processing Zoom recordings, campaign data, or audit results are specifically tasked with generating these notes as part of their workflow. The notes feed into the SOP Update Protocol when they affect processes, or directly into the relevant definitive article when they add examples, data points, or case studies.
This is the mechanism that makes the system smarter over time. More client calls means more insights. More campaigns means more data. More audits means more examples. Each one flows through the capture pipeline into the content architecture, and the next agent who reads the definitive article benefits from all previous agents’ experience.
Process 4: Platform Portability Discipline
BlitzMetrics documents knowledge outside of any particular AI system so that the entire Content Factory can be moved from Claude to ChatGPT to Gemini to whatever comes next without losing institutional knowledge. This is not just a philosophical principle — it requires specific structural discipline in how we write SOPs and articles.
The portability discipline separates every process document into two layers. The methodology layer describes what we do and why — this is the transferable knowledge that works regardless of which tool executes it. The implementation layer describes which buttons to click in which tool — this is the platform-specific detail that changes when we change tools.
For example, the Blog Posting Guidelines currently reference specific tools: Descript for video transcription, Rank Math for SEO settings, WordPress for publishing, Grammarly for proofreading. The methodology — transcribe video, optimize for search, publish to web, check grammar — is platform-independent. The implementation — “open Descript, click Import, select the video file” — is platform-specific.
When writing or updating any SOP, the author must maintain this separation. Methodology sections use tool-agnostic language. Implementation sections are clearly marked and can be swapped out when tools change. This means that if BlitzMetrics moves from WordPress to a different CMS, or from Claude to a different AI platform, only the implementation layers need to be rewritten. The methodology — which represents the actual institutional knowledge — remains intact.
This discipline also applies to how agents document their work in meta-articles. The Meta-Article Prompt captures both what the agent did (methodology) and how it did it (implementation). Future agents on different platforms can learn from the methodology even if they use completely different tools.
Process 5: The Recursive Improvement Cycle
The four processes above — auditing, SOP updates, knowledge capture, and portability discipline — are themselves subject to improvement. This article is not exempt from its own standards. The recursive improvement cycle ensures that the maintenance system maintains itself.
Every six months, this article undergoes the same audit that definitive articles receive. The auditor asks: are the five processes still the right processes? Has operational experience revealed a sixth process that needs to be added? Have any of the five processes proven unnecessary or redundant? Are the specific procedures within each process still producing the intended results?
The auditor also reviews the SOP Amendment Proposals from the past six months to identify patterns. If multiple proposals address the same gap, it suggests a structural issue that requires a new process rather than a patch to an existing one. If no proposals have been submitted, it suggests either that the system is working perfectly (unlikely) or that the proposal mechanism itself has a friction problem that needs to be addressed.
This is what makes the system genuinely self-improving rather than merely self-documenting. The documentation layer (meta-articles) captures what happened. The maintenance layer (this article) ensures what happened was correct. And the recursive layer ensures the maintenance layer itself stays correct. Each layer watches the one below it, and the whole system gets better with every cycle.
How This Connects to the SEO Tree
The Definitive Article Guide uses the SEO Tree metaphor — trunk, branches, and leaves — to explain how content should be organized. This maintenance article extends the metaphor to cover what arborists call “tree care.”
Building the tree is the Definitive Article Guide’s job. The trunk is the BlitzMetrics entity. The branches are definitive articles — one per major concept. The leaves are case studies, meta-articles, and supporting content that connect back to their branch.
Maintaining the tree is this article’s job. Pruning removes content that does not belong on a particular branch — like a Pilot Plumbing video on an ARDMOR article. Fertilizing adds new data, examples, and insights from live interactions — the Knowledge Capture Pipeline. Inspecting checks that every branch is still healthy and connected to the trunk — the Quarterly Audit. Grafting carefully integrates new knowledge into existing branches without damaging the structure — the SOP Update Protocol.
Without maintenance, even a well-built tree becomes overgrown. Branches cross each other. Dead wood accumulates. The tree looks impressive from a distance but does not produce good fruit. The Content Factory works the same way — without active maintenance, the content library grows but the quality degrades, and the SEO value erodes as Google encounters pages with incoherent signals.
Why Humans Must Understand This, Not Just Agents
AI agents can execute every process in this article. They can run audits, generate amendment proposals, create knowledge capture notes, and maintain portability discipline. But if the humans directing those agents do not understand the underlying principles, the agents will make mistakes that look correct on the surface.
An agent told to “add a video to this article” will add the video. It will format it correctly, add schema markup, and update the table of contents. But if the human directing the agent does not understand the SEO Tree — does not understand that every element on a page must directly support the page’s topic — the agent will dutifully add a video about the wrong company to the wrong article. The agent’s work will be technically perfect and strategically wrong.
This is the fundamental challenge of AI-assisted content production. The tool amplifies whatever understanding the operator brings to it. A human who understands the SEO Tree, the Definitive Article framework, and the maintenance principles in this article will direct agents to produce content that strengthens the entire system. A human who does not understand these things will direct agents to produce content that looks good but slowly degrades the system’s coherence.
Every team member at BlitzMetrics — not just the senior strategists, but the VAs, the editors, and anyone who touches content — must be able to explain why a Pilot Plumbing video does not belong on an ARDMOR article. Not because they memorized a rule, but because they understand the tree. The articles we write and the SOPs we maintain are teaching tools first and process documents second.
Implementation Checklist
For teams adopting this maintenance system, these are the concrete actions to take.
Add a “Last Audited” column to the status table in the Definitive Article Guide. Set the initial value to “Not yet audited” for every article. Schedule the first round of quarterly audits and assign an auditor to each definitive article.
Create a shared location for SOP Amendment Proposals. This can be a Google Doc, a WordPress draft category, or a project management board — the tool does not matter as long as proposals are visible to reviewers and tagged with the SOP they affect. Assign a weekly reviewer.
Add Knowledge Capture Note generation to every workflow that processes Zoom recordings, campaign data, or audit results. The note template — source, insight, destination, priority — should be available to every agent and team member.
Review existing SOPs for portability discipline. Identify sections where methodology and implementation are mixed. Begin separating them so that tool-specific instructions are clearly marked and swappable.
Schedule the first six-month recursive review of this article. The reviewer should assess whether the five processes are still the right processes and whether operational experience has revealed gaps.
Related Articles
How to Create a Definitive Article for Any BlitzMetrics Concept — The architecture SOP. Explains the SEO Tree, the eight-step framework for building definitive articles, and the status table that this maintenance article audits.
Meta-Article Prompt Template — The documentation SOP. Explains how to capture the process behind each article so that future agents can learn from it.
Blog Posting Guidelines — The execution SOP. Covers the step-by-step workflow for publishing content, from video transcription through final QA.
Dennis Yu
Dennis Yu is the CEO of Local Service Spotlight, a platform that amplifies the reputations of contractors and local service businesses using the Content Factory process. He is a former search engine engineer who has spent a billion dollars on Google and Facebook ads for Nike, Quiznos, Ashley Furniture, Red Bull, State Farm, and other brands.
Dennis has achieved 25% of his goal of creating a million digital marketing jobs by partnering with universities, professional organizations, and agencies. Through Local Service Spotlight, he teaches the Dollar a Day strategy and Content Factory training to help local service businesses enhance their existing local reputation and make the phone ring.
Dennis coaches young adult agency owners serving plumbers, AC technicians, landscapers, roofers, electricians, and believes there should be a standard in measuring local marketing efforts, much like doctors and plumbers must be certified.