Editorial Policy

Last updated: August 2026

This policy sets out the standards that calculator content on the UK Calculator Platform is held to, and how those standards are enforced in practice. It is written to be checkable: most of what follows is something you could hold a specific page against and find it wanting.

1. Scope

This policy covers everything editorial on the platform — calculator guides, methodology explanations, worked examples, assumptions, limitations, category descriptions, FAQs and the governance pages themselves. It does not cover the arithmetic itself, which is governed by code and automated verification and is described in how we check our figures.

2. Content principles

• Say what the calculator actually does. Every guide states the model being applied, not a flattering description of it.
• Show the working. The governing formula and the ordered steps behind it are published, so a reader can follow the reasoning rather than take the number on trust.
• State the limits as prominently as the capability. Assumptions and limitations are required sections, not optional extras appended when there is room.
• Plain English. Technical terms are used where they are the accurate word and explained on first use. Internal engineering identifiers never appear in reader-facing content.
• No content written to rank. Pages are written to answer the question the reader arrived with. FAQs cover questions people genuinely ask, not keyword variants of the page title.

3. UK-first approach

The platform models UK rules and states figures in pounds sterling using UK conventions. Where a rule differs between England, Northern Ireland, Scotland and Wales — Income Tax bands, Stamp Duty and its devolved equivalents, and student loan plans among them — the difference is modelled and named rather than averaged into a single national figure. Where a calculator applies to one nation only, its guide says so.

4. Source hierarchy

Sources are ranked, and the rank determines what a source is allowed to prove. The hierarchy is enforced automatically: a source is placed in a tier by the publishing body it comes from, not by an author's assessment of it.

Four rules follow from that ranking, and they are the ones that do the real work:

• A primary official source is mandatory for any statutory value — a rate, threshold, band, allowance, benefit amount or statutory formula. A Tier 2 or Tier 3 source may explain such a figure but may never be the only evidence for it.
• A search-result snippet is not a source. Neither is an AI-generated summary, nor a secondary page quoting a figure without citing where it came from. The underlying official page must be located and read; if it cannot be, the figure is not published as verified.
• Inaccessible sources are recorded, not worked around. If an official page has moved, been withdrawn or cannot be reached, the claim it supported is marked as requiring verification and is shown that way on the page. It is not quietly re-attributed to a weaker source that happens to agree.
• Conflicts escalate towards legislation. Where two sources disagree, the more primary one governs, with legislation above departmental guidance and guidance above explanatory material. Where two Tier 1 sources genuinely conflict, the figure is treated as unresolved and marked as requiring verification until it is settled, rather than resolved by picking the more convenient answer.

Every cited source records the rule it supports and whether it has been verified, and — where the source itself stated one — the period the figure applies to. An absent period means the source did not say, which is deliberately different from a guess.

5. Methodology writing

Each guide explains how the calculation works in prose, then states the governing formula plainly and breaks it into ordered steps. The methodology describes what the engine actually does. Where the engine simplifies — and it usually does — the simplification is named in the assumptions rather than glossed over in the method.

6. Worked examples

Every guide carries a worked example built from a realistic scenario, with the inputs shown as the reader would enter them and the steps followed through to the result.

Published example figures are not typed in by hand. Each example stores the exact inputs it was computed from, and the automated test suite re-runs the calculation engine against those inputs and compares the result to every figure published in the guide. If an engine change moves a number, the suite fails. A worked example on this platform therefore cannot drift away from the arithmetic it claims to demonstrate — the two are checked against each other on every build. Examples that depend on the current date pin the date they were computed on, so they neither drift nor quietly fall outside the calculator's own validation window as time passes.

7. Assumptions and limitations

Both are required sections. Assumptions record the modelling choices the reader is entitled to know about — which reliefs are and are not applied, what is assumed about growth or inflation, what is treated as constant. Limitations state where the model stops being a reliable guide, including the circumstances it does not attempt to handle.

A calculator that presents itself as more capable than it is causes more harm than one that is plainly bounded, because the reader has no way to know when to stop trusting it.

8. Corrections policy

Errors are corrected rather than quietly overwritten.

• Report. Anyone can report a suspected error via the contact page. A report with inputs, the observed result and a supporting official source can be investigated immediately.
• Reproduce. The reported inputs are run against the engine to establish what the platform actually produces, before any judgement is made about whether it is wrong.
• Verify independently. The disputed rule is checked against a primary official source. The expected value is derived independently rather than by reading it off the code that is under suspicion.
• Fix and prove. A confirmed defect is corrected, and regression cases covering it are added so the same error cannot return unnoticed. Benchmark expectations are never edited to match what the code happens to produce.
• Record. Corrections that change a published result are listed on the updates page, naming the calculators affected.

9. Rules-sensitive calculators

A calculator is treated as rules-sensitive when its output depends on rules that can change when legislation or official rates change. Those calculators carry extra obligations: the tax year is named on the page, the figures are stated against a specific versioned ruleset, every statutory value is evidenced by a primary official source, and the review date is shown so a reader can see how current the content is.

Review freshness is derived from the editorial content itself rather than maintained as a separate list, so a review date cannot exist for work nobody did. Rules-sensitive content falls due for review at the start of each UK tax year; general content is revisited on a fixed cycle to catch link rot and drifting explanation.

10. AI-assisted development and editorial use

AI-assisted tooling is used in building and maintaining this platform, and we would rather say so plainly than let you discover it. It supports software development, organising research, drafting explanatory prose, generating test scaffolding and quality assurance work.

What matters is what it is not trusted to do. No calculation result, statutory rate, threshold or rule reaches this platform on the authority of a language model. Figures are anchored to four things instead:

• Deterministic code. Results come from a calculation engine that produces the same output for the same input every time, not from generated text.
• Approved rule configuration. Rates and thresholds live in versioned rule data tied to a named tax year, reviewed as a deliberate change.
• Independently derived benchmark cases. Expected values are worked out from the rules themselves, not by running the code and recording whatever it produced. A benchmark that simply agreed with the implementation would prove nothing.
• Official sources. Rule-sensitive figures are checked against primary official sources under the hierarchy above, and published with a link so a reader can check them too.

An AI-drafted explanation is treated as a draft requiring verification, exactly like any other unchecked draft. Where a figure has not been confirmed against an official source, the page says so rather than presenting it as verified.

11. Human accountability

The platform is operated by Jomovate, and responsibility for what is published rests there. Changes to calculation logic, statutory rules and editorial content are reviewed before release and are recorded in version control, so any published figure can be traced to the change that introduced it.

We do not claim an editorial board, named external reviewers, professional memberships or third-party certification, because we have none. Where content would benefit from professional review we say so rather than implying it has already happened.

12. Advertising and independence

The platform carries no advertising, no sponsored content and no affiliate links, and no product provider pays for placement or mention. Nothing in a calculator result or a guide is influenced by a commercial relationship, because there are none to influence it. If that ever changes, this section will be updated to say so, and any paid or affiliate relationship will be disclosed on the pages it affects.

The platform does not use marketing, advertising or tracking cookies — see the privacy policy.

13. Updates and review

Content is reviewed on a schedule rather than only when something breaks. The review cycle covers the annual UK tax year rollover, Budgets and other fiscal events, benefit and pension uprating, mid-year statutory changes, and periodic checks that cited sources still resolve and still say what they said. Material user-facing changes are published on the updates page.

14. Medical, financial and tax boundaries

The platform provides general information and estimates. It is not regulated, and it does not provide financial, investment, tax, legal or medical advice, nor personal recommendations of any kind.

• Tax. Calculators apply simplified standard rules and cannot account for the full range of individual circumstances, reliefs and elections. HMRC or a qualified accountant is the authority on your position.
• Financial products. Projections are illustrations built on assumptions that will not hold exactly. They are not lending offers, quotations or recommendations. Advice on a financial product should come from an FCA-regulated adviser.
• Health. Health calculators are general educational tools using population-level formulas. They do not diagnose, do not account for individual clinical circumstances, and do not replace advice from a medical professional.

Where health content cites clinical guidance, it is sourced from the NHS or NICE. We do not claim clinical review of that content by named practitioners, because none has taken place.

Questions about this policy

Get in touch via the contact page.