Method
Editorial log
Every published page on this site, who reviewed it, when, and what the review changed.
Scaled content is penalised when it is published without review. Rather than assert that review happens, the record is kept here and the build will not publish a page whose review block is missing or dated before the page was written.
One collection is reviewed differently and the table says so rather than hiding it. Market pages carry no third-party figure — every publishable fact on one is a named authority, a section number, an institution or a link — so each is checked by retrieving the source and searching it, and the URL, the status, the hash and the matched passage are recorded. A field that cannot be established that way is dropped; a page whose required fields cannot be is not published. Everything that carries a figure somebody else measured is still read by a person before it ships.
| Page | Type | Reviewed by | Reviewed | What the review changed |
|---|---|---|---|---|
| Should you hire a dispatcher or an office coordinator? | Answer | Answer Production Engine | 2026-08-23 | The draft used a single loaded-cost multiplier, which the provenance trail on that multiplier shows to be an origin-less figure. Replaced it with the statutory components themselves, each of which is sourced and behaves differently by earnings, and added the concurrency-versus-capacity distinction because it is the test that decides whether the hire is the right instrument at all. |
| Should you hire or outsource? | Answer | Answer Production Engine | 2026-08-23 | The draft compared the two on cost, which is the comparison that cannot be made cleanly because the visibility of the two costs differs. Rebuilt around what the work leaves behind. Added the reversibility asymmetry, which makes renting a legitimate way to test whether a function is needed rather than merely a compromise. |
| Should you let AI reply to customers without a human? | Answer | Answer Production Engine | 2026-08-23 | The draft treated autonomy as one permission. Splitting it into acknowledge, answer and act is what makes the boundary usable, since the three have different error costs and businesses currently apply one rule to all of them. Added the escalation-on-uncertainty property, which does more work than the category boundaries and is testable before anything is bought. |
| Should you offer a guarantee? | Answer | Answer Production Engine | 2026-08-23 | The draft treated a guarantee as a conversion device and did not address what happens when one is claimed. That behaviour determines whether it helps or harms, so it became the central test. Added the substitutes, since a business that cannot afford or control a guarantee should not be left without an answer to the same fear. |
| Should you open on weekends? | Answer | Answer Production Engine | 2026-08-23 | The draft argued for weekend opening on competitive grounds. Being reachable captures most of that benefit at a fraction of the cost, so the two were separated and the reachable option placed first. Added the extended-week cost, which the first version omitted and which is what makes weekend arrangements fail after a few months. |
| Should you publish your availability? | Answer | Answer Production Engine | 2026-08-23 | The draft proposed a live availability calendar. That is a larger commitment than the question requires and it exposes quiet periods, so the dated range replaced it. Added the per-service ranges, since a single figure across short and long lead work misleads every reader. |
| Should you raise your prices? | Answer | Answer Production Engine | 2026-08-23 | The draft used cost increases as the trigger. Costs explain why a price must move and say nothing about what the work is worth, so the signals were changed to win rate and capacity. Added the disagreement case, since a high win rate with spare capacity is a demand problem and raising prices into it does damage. |
| Should you record customer calls? | Answer | Answer Production Engine | 2026-08-23 | The draft treated consent as a single national rule. It differs by state and some require all parties, so the page now states that and gives the practice that satisfies both. Added the note-taking alternative, since a business that decides against recording should not be left without the training benefit. |
| Should you require a phone number on your enquiry form? | Answer | Answer Production Engine | 2026-08-23 | The draft recommended requiring the number on speed-to-contact grounds. The fake-number effect undermines that: a required field converts an abandonment into an unreachable enquiry that looks real in every report, which is worse than the blank. Reversed the recommendation and added the preferred-channel field, which carries more information than the number itself. |
| Should you run ads or improve your site first? | Answer | Answer Production Engine | 2026-08-23 | The draft answered site first without qualification. That is wrong where nothing reaches the site at all, so the two-number test was added to distinguish the cases. The unanswered-enquiry case was added after noticing it presents identically to a conversion failure and is the cheapest of the three to fix. |
| Should you send a newsletter? | Answer | Answer Production Engine | 2026-08-23 | The draft recommended a fixed monthly cadence. Committing to a frequency rather than to having something is what produces filler and spends the list, so the ordering was reversed. Added targeted contact as the alternative, since most businesses asking this question would be better served by it than by any newsletter. |
| Should you specialise or stay general? | Answer | Answer Production Engine | 2026-08-23 | The draft argued for specialising without distinguishing marketing from capability, which made the recommendation sound riskier than it is and would have been wrong advice for a shop with expensive general-purpose equipment. Separating the two decisions is the correction, and the market-size caveat was added because it is the one case where a narrow position genuinely fails. |
| Should you take a deposit? | Answer | Answer Production Engine | 2026-08-23 | The draft justified deposits on cash flow and suggested a standard percentage. The commitment-test framing is the stronger one and it produces a better rule for the amount, so both were changed. Added the deferred-work case, which a deposit does not address and which businesses assume it does. |
| Should you take on a big client? | Answer | Answer Production Engine | 2026-08-23 | The draft gave a concentration percentage as the threshold. No source supports a general figure and the survivability question is the one that actually matters, so it was replaced. Added the volume-shortfall term, which the first version omitted and which is the term most often regretted. |
| Should you track competitor prices? | Answer | Answer Production Engine | 2026-08-23 | The draft recommended regular competitor price monitoring. Continuous observation invites reaction to a figure that means something different in another business, so it was reduced to periodic with scope attached. Added the competition-law boundary, which is easiest to cross in exactly the informal settings small businesses operate in. |
| Should you use a CRM? | Answer | Answer Production Engine | 2026-08-23 | The draft used a revenue threshold as the trigger. Size predicts this poorly and the handoff predicts it well, so the trigger was changed. Added the future-state trap, which is the most common way a correctly chosen system fails after purchase. |
| Should you work with friends and family? | Answer | Answer Production Engine | 2026-08-23 | The draft advised against these engagements. That is neither realistic nor necessary — the failures trace to informality rather than to the relationship — so the page now gives the structure that makes them work. Added the decline category, since some of these jobs require a conversation the relationship makes harder and no amount of paperwork addresses that. |
| Should you write for people or for machines? | Answer | Answer Production Engine | 2026-08-23 | The draft accepted the trade-off and advised balancing the two audiences. The requirements have converged enough that the balancing framing is misleading, so the page now states the convergence and isolates the narrow genuine difference. Added the overconfident-prose failure, which is new and specifically produced by writing for extraction. |
| What AI automation is actually useful for a small business? | Answer | Answer Production Engine | 2026-08-23 | The draft was organised as a list of use cases, which reproduces the vendor framing the page is arguing against and would have aged badly. Reorganised around the handoff-versus-judgement division, which is the property that predicts survival, and added the middle band because draft-for-approval is where most of the current practical value sits and a two-way split would have excluded it. |
| What can go wrong when you automate customer messages? | Answer | Answer Production Engine | 2026-08-23 | The draft led with model error, which is the fear rather than the risk. Reordered by observed frequency and cost, which puts the unmonitored return path first. Added tone drift, which the first version missed entirely because it is the only failure on the list that no single message causes. |
| What data should you never put into an AI tool? | Answer | Answer Production Engine | 2026-08-23 | The draft was organised around data sensitivity levels, which requires a judgement most staff will not make consistently under time pressure. Reorganised around custody — whose data is it — which produces the same boundary and is answerable in a second. Added the account-arrangement section, since it is the control that actually changes what is permitted. |
| What does a good week look like in a service business? | Answer | Answer Production Engine | 2026-08-23 | The draft included revenue and jobs completed. Neither can be acted on within a week and both reflect decisions taken earlier, so they were excluded with the reasoning stated. Added the compounding work item, without which the definition would describe a well-run business that never changes. |
| What does AI automation cost a small business? | Answer | Answer Production Engine | 2026-08-23 | The draft quoted a price range for small-business automation. There is no admissible source for one and the range would have been read as a benchmark, so it was cut. The page now gives the cost structure instead, which is what a buyer actually needs in order to evaluate whatever quote they are holding. |
| What is a customer actually worth? | Answer | Answer Production Engine | 2026-08-23 | The draft included a discounted cash flow treatment. The error in the repeat assumption dwarfs the discounting effect for a small business, so it was removed in favour of a conservative repeat estimate, which addresses the same concern and can be explained to whoever uses the number. |
| What is a lead magnet and does it work? | Answer | Answer Production Engine | 2026-08-23 | The draft treated gating as the default. Gated material is invisible to search and retrieval, which is a real cost for a business trying to be found, so the open-publication alternative was added with the interactive-assessment approach that captures both. Also changed the success measure from downloads to customers. |
What the build refuses to publish.
Four checks run before any page in this system is generated, and each one stops the build rather than producing a warning nobody reads.
A page must name, in writing, what is on it that could not appear unchanged on a sibling page — and no two pages may give the same answer. Independently of that, every page is reduced to a fingerprint with its own subject and place names masked out; if two fingerprints match, or come close, both pages are named and the build stops. That is the check that catches a page whose only distinguishing feature was the city in the heading.
A page must also carry at least two datapoints that are specific to its own subject rather than to its category, and if the same figure is claimed as specific on two different pages, it was specific to neither. Pages ship in cohorts with a size cap so indexation can be observed between them, and every page carries the review block that produces the table above.
All four are properties of the schema, not of anyone remembering. Reviewed by Siddharth Sharma, except the 0 market page(s) above, which are checked against their sources and carry the record of it.
GATES
A gate enforced by a person reading pages stops being enforced somewhere around page 60.