Five rounds of critique. A corpus-comparative quality bar. A dedicated brand identity for every build. This is what it takes to produce something worth owning.
Most studios ship an MVP and move on. The Factory runs on a different principle: nothing counts as done until it has been torn apart and rebuilt at least twice. The process is systematic, not aspirational. Every tick of the loop advances a specific artifact toward a specific quality threshold, and the threshold is not negotiable. A product that clears only four of the five gates does not ship. A landing page that reads well but has broken links does not ship. The bar exists because the bar is the product.
Before a single line of code is written, every product concept goes through an intake pipeline that scores it across four dimensions: market signal, ICP clarity, competitive white space, and monetization fit. Concepts that do not score in the top quartile do not advance. There is no holding queue for "good ideas."
Ideas are sourced from market signals, sales conversations, and observed friction in specific industries. No open-ended brainstorming sessions. Every concept that enters the pipeline starts with documented evidence of a real problem in a named workflow.
A product without a named customer is not a product. Before a concept advances, the target role, the specific workflow disrupted, and the measurable pain must be stated explicitly. Vague ICPs are rejected at this gate.
Concepts are evaluated against a weighted rubric. High scores earn a build slot. Low scores go back to the inbox for refinement. There is no "we will revisit this later" holding pattern. If it does not score well today, it is not ready.
Before committing to a build slot, the full portfolio is reviewed for thematic overlap. If three or more existing products share a root category, the next candidate must come from a different vertical. Concept diversity is enforced, not hoped for.
Every product gets its own visual language before anything is built. Not a studio template with a color swap. A distinct palette, a typography pairing chosen for the actual customer audience, and a one-line voice description that governs every headline and CTA on the page.
A property management tool does not get the same look as a grant writing platform. A restaurant menu assistant does not share typography with a law firm intake system. Each audience carries specific visual expectations, and the build meets those expectations precisely.
The brand brief is saved to file before build starts, and it is read back when any polish action fires on the same product. Identity drift across critic rounds is how products lose their voice.
The Factory Director runs on an adaptive tick loop. Every tick, it surveys the entire portfolio, applies a thirteen-option decision matrix, and picks the single highest-value action available. No action is taken without a stated rationale. The loop runs overnight and returns a complete, battle-tested artifact by morning.
Missing or thin pages in the studio micro-site. Every page must pass a minimum word count, have real photography, and present varied section types before it is considered live.
An existing artifact is audited. Five or more specific weaknesses are identified. A full v2 is written with every issue resolved, not patched. The file is replaced entirely, not appended.
A net-new product from the scored picks file. Brand brief is written first. Single-page landing with real photography, complete feature story, three-tier pricing, and a working FAQ. Minimum 1500 words.
Pricing pages, FAQ sections, case studies, security pages, onboarding checklists, integrations. A product that only has a landing page is not done, and the loop does not treat it as done.
Email drip sequences, LinkedIn outreach packs, and 90-second demo scripts. Every completed product ships with three GTM assets before it is marked done.
Corpus-comparative critique against twelve reference sites. Produces 12 to 25 specific demands per pass. This is round three of the critic pipeline, never round one.
A product that passes one round of critique is not ready. The standard is five progressively demanding reviews, each targeting a different failure mode. Rounds are not redundant. They are additive. Later rounds assume earlier gates have already been cleared.
Broken links, placeholder content, missing sections, href="#" dead-ends, absent pricing tiers. The page must be functionally complete before any conceptual critique begins. A beautiful page with five broken links fails this round.
Does the headline address the named ICP? Is the value proposition specific or generic? Are features described in outcomes, not specifications? Does the CTA vocabulary match the buyer's stage? Does the page have a clear villain (the problem) and a clear hero (the product)?
A corpus-comparative review against twelve reference sites selected for design and conversion benchmarks. The Foreman produces specific, numbered demands rather than general impressions. Every demand has a target element, a specific failure mode, and a recommended fix. Round three is always the Foreman. Never another internal pass.
A separate model reads the page cold and identifies what a skeptical buyer would object to. This step catches assumptions baked in so deeply that the original author cannot see them. Objections that feel obvious in hindsight are exactly the ones that were invisible before this round.
All demands from rounds one through four are closed in a single complete rewrite. Not a patch job applied to the existing file. The page is rebuilt with every identified failure mode addressed. After this pass, the product is marked done and moved to the next phase.
These are not aspirational targets. They are gates. An artifact that does not meet them does not advance to the next phase, regardless of how polished it appears.
The word "done" is used loosely in most studios. Here it has a specific definition. A product is done when it could be handed to a salesperson tomorrow and they would have everything they need to start conversations without any additional production work.
That is a higher bar than "the landing page looks good." It means the go-to-market collateral exists, the pricing is specific and defensible, the common objections are addressed on the page, and a 90-second demo can be recorded without prep or editing.
Products that have cleared rounds one and two but not round five are not done. They are in progress. The log tracks the distinction publicly, so anyone looking at the portfolio can see exactly which products are finished and which are still moving through the pipeline.
A polished landing page with no outreach strategy is a brochure. Every completed product ships with three go-to-market assets that make the first thirty sales conversations possible without additional production work from the client.
A cold outreach sequence targeting the named ICP. Each email is under 150 words, opens with a specific pain signal, and closes with a low-friction call to action. Subject lines and send-order guidance included. Voice matches the product brand brief.
Short-form messages under 300 characters, written for specific named roles in the ICP organization. Designed for direct message, not connection request spam. Each message targets a different angle on the same core problem. Copy-paste ready.
A structured walkthrough covering problem statement, product demonstration, and close. Written for screen recording. The structure is tight enough that Wes can read it cold and capture usable content in one take without editing or prep time.
Browse the current build portfolio or read more about the studio behind the process.