Secrecy is what you sell when you can't show receipts. This page is the complete methodology — the five-step loop every asset runs through, the exact scoring weights, and the compliance gate that nothing skips. Show it to your agency. Ask them for theirs.
Nothing we publish is a one-off. Every asset — article, rewrite, social post, review response — moves through the same five stations, and what the last station learns feeds the first one on the next pass. That loop, running monthly on everything you own, is the product.
Every page you own is graded on the rubric below, exactly as an AI answer engine reads it: does it answer the question in the first sentence? Is the entity structure clear? Is there anything here a competitor couldn't copy-paste? The score decides what gets written, rewritten, or left alone — not a hunch, not a "content calendar."
The questions your buyers actually ask, in your county, in their words — pulled from real query data, not a national keyword list rented from a tool subscription. A DUI question in Travis County is not a DUI question in Cook County, and the research treats them that way.
Before anything publishes, it's checked against your jurisdiction's advertising rules: required disclaimers inserted, guarantee language blocked, competitor references suppressed, bar-specific restrictions applied. Failures stop the line, with the reason attached. This step has no override switch — that's the point.
Answer-first structure, entity-clear copy, complete schema — built so an answer engine can lift your paragraph with your name attached. Then distributed the way each platform actually works: LinkedIn thought leadership is not a Google Business Profile update is not an Instagram caption.
Every asset is tagged when it ships and reports its own numbers to the ledger: traffic, impressions, rankings, clicks, engagement — per page, per post. When your firm confirms a consultation or a signed matter, the ledger ties that outcome to the asset that started it. We don't claim to read your phone lines or your CRM; we claim receipts for what's provable and confirmation for what isn't.
What the ledger learns, the next cycle uses. Underperforming assets get rescored and rewritten; winning patterns get repeated across your site. The shared brain is why month twelve beats month one — and why the longer it runs, the harder it is to replace.
Every asset is scored out of 100 before it ships, and every free SiteScan runs the same rubric on your existing pages. These are the weights — not a summary of them, the actual weights.
Would a rival attorney learn something from this page? Half the entire grade is whether the content says something a competitor couldn't copy-paste — a real case-handling insight, a jurisdiction-specific wrinkle, a number with a source. This is the half agencies skip, because it's the half that can't be produced by a content farm. It's also what generative engines cite: an LLM has no reason to quote the 400th generic explainer.
Experience, expertise, authoritativeness, trust — demonstrated, not asserted. Named authorship, credentials that check out, citations to primary sources, and the structural signals search engines use to decide whether a page deserves to rank for a consequential query.
Answer-first ordering: the question answered plainly in the opening, then the depth. Clear headings that match how questions are asked, scannable structure, and entity clarity — so both a human in a hurry and an answer engine assembling a response can lift the right paragraph.
The 2019 stuff still has to be right: titles, internal linking, canonical hygiene, speed, crawlability. Ten points because it's necessary, not fifty — because it stopped being sufficient years ago.
Complete structured data — FAQ, LegalService, Organization, Article — so machines know exactly what the page is, who it's from, and what question it answers. Five points, near-zero excuses for missing them.
// Version note: the rubric is versioned (this is v3) and re-weighted as answer engines change. When the weights move, this page moves with them — that's what "published" means.
Competitors sell "pending patents" and proprietary engines you have to take on faith. We'd rather you check our work: run the free SiteScan, watch your own pages get graded on the rubric above, and judge the rewrite it produces before you spend a dollar. If the methodology is real, showing it costs nothing. If it isn't, hiding it is the business model.