Editorial policy

How we decide what to publish — and what we refuse to publish.

Last updated 16 October 2026

Our principles

  1. We don’t fabricate. No invented prices, features, statistics, testimonials, reviews or test results.
  2. We label uncertainty. If we haven’t verified something, the page says “Not yet verified” and links to the official source.
  3. We write our own words. We don’t copy vendor descriptions or scrape other websites.
  4. Money doesn’t buy opinions. Affiliate relationships and sponsorships never change what we write.

How tools are added

We add tools that are useful to our readers and that we can describe accurately. A new profile starts with a plain-English description of the product’s core purpose, who it is aimed at and its main use cases, based on publicly available information. At this stage:

  • Pricing is shown as “Not yet verified” until someone has checked the vendor’s official pricing page and recorded the date.
  • Features describe only the product’s well-known core purpose and are marked as not verified hands-on.
  • Pros, cons and ratings are left empty.

How the tool database works

Each profile mixes three kinds of information, and we label which is which:

  • Verified facts: pricing, free plans and country availability. We show these only after someone has checked them on the vendor’s own pages, and we record the date. Until then the profile says “Not yet verified”.
  • Publicly listed facts: platforms (web, Windows, iOS…), integration types and notable integrations. These come from what each vendor publishes on its website, app-store listings and integration directories. They’re marked “not yet verified by ToolTarvern” until we confirm them and record the date. When we’re unsure, we leave the field blank, and filters treat it as unknown.
  • ToolTarvern classifications: our editorial judgement, designed to help you filter. They are not quality scores.
    • AI capability — “AI is the core product” (the tool exists to generate or analyse with AI), “Has AI features” (AI assists an otherwise conventional product), or “No notable AI features”.
    • Built for — the business sizes the product is designed and priced for: solo/freelancer, small business (2–50 people), mid-size (51–250) or large organisation. A tool can suit several.
    • Use case — plain-language jobs (such as “Accounting & invoicing”), derived from the categories a tool is listed in.

Search results are ranked by how closely your words match: tool name first, then tags and categories, then use cases, then the description and features. No tool can pay to rank higher. A sponsored placement, if we ever sell one, is labelled “Sponsored”.

Reviews and ratings

We publish pros, cons and a rating only after a person on our team has used the tool for a real task. The profile then shows a “Last reviewed” date. If a review is based on a trial or a limited plan, we say so.

No tool currently has a ToolTarvern rating. We never show star ratings from other sites, and we don’t calculate scores from vendor claims. When we do rate a tool, this is exactly how the score is worked out:

  1. The reviewer completes at least three realistic tasks for the tool’s main use case and writes down what they did.
  2. They score five criteria from 1 to 5: ease of use, how well it does the main job, value for money (at the verified price on the review date), help and documentation, and fit for South African users (currency, payments, local support and availability).
  3. The overall rating is the plain average of the five scores, rounded to the nearest 0.5. All five criteria count equally.
  4. The profile shows all five scores, the tasks tested, the plan used and the review date, so you can see where the number comes from and weigh the criteria differently yourself.

Ratings are re-checked when we re-review a tool. A rating more than 12 months old is marked as out of date.

Guides and comparisons

Every guide and comparison includes a “How we made this” note describing its basis — for example, desk research from public product information, or hands-on testing. Comparisons explain trade-offs; we do not declare winners we haven’t tested.

Commercial independence

  • Search results, listings and comparisons are ordered by relevance or alphabetically. Affiliate status, sponsorship and advertising play no part, and our automated tests check this.
  • Advertisers, sponsors and affiliate partners can’t see, approve or change our content before it’s published.
  • Paid placements sit in separate, labelled slots. They never replace or reorder editorial content.
  • If a commercial relationship ever conflicts with giving readers an honest answer, the honest answer wins.

Affiliate links and sponsorships

  • Affiliate links are added only where a real agreement exists, and are labelled next to the link.
  • Sponsored placements are always labelled “Sponsored”.
  • Display ads, if shown, are labelled “Advertisement” and never placed inside a free tool’s interface.

See our affiliate disclosure for details.

Use of AI

We may use AI tools to help with research, drafting or editing. A person is responsible for every page we publish, checks facts against primary sources and ensures it follows this policy.

Corrections

Prices and features change. If you spot something wrong or out of date, email support@digitaltarvern.co.za. We correct factual errors promptly and update the “last updated” date on the page.