This page is the standard we hold ourselves to: who decides what gets published, what counts as evidence, what people and software each do, how commercial relationships are handled, and what happens when we get something wrong. How the testing itself works is on the methodology page; this page is about the rules around it.
Editorial independence
Vendors may provide corrections, documentation, briefings or test access, but they do not control coverage, review drafts before publication, choose conclusions, or buy placement or positive verdicts. Nobody reviews our drafts before publication except us. We do not sell placement, we do not sell positive verdicts, and we do not publish sponsored or guest posts.
Service Tech Select does not currently use affiliate links. If that changes, a link that can earn us a commission will be labelled where it appears, and commission will never determine our rankings.
Commercial relationships cannot affect eligibility, evaluation criteria, ranking or verdict. Our affiliate disclosure lists current relationships and explains how a commercial link is labelled.
We publish product reviews, comparisons and alternatives, trade buying and requirements guides, pricing and feature guides, workflow explainers, and methodology and policy pages. A page must add useful organization, analysis, verification, comparison or original hands-on evidence. A pricing guide can add value by making scattered vendor information easier to evaluate even when the underlying facts come from first-party sources; a page that adds nothing is not published.
Sources and evidence
Not all evidence is worth the same, so we label it and use the source that best fits the claim. For how software behaves in use, a controlled test of our own is usually strongest. For current pricing, plan limits and formal policies, current first-party documentation may be controlling. Vendor support can clarify ambiguous documentation, credible third-party evidence covers attributable real-world experience, and inference is always marked as inference.
When what we observe contradicts a vendor’s marketing copy, or two of a vendor’s own pages disagree, we record the conflict and investigate before we publish anything. We do not average the two, and where the conflict stays unresolved the page says so.
Material mutable facts, such as a price, a plan limit, a feature gate or an integration, are maintained in structured records where practical, so pages reuse the same source, value, market qualifier and verification date instead of restating a number that can drift.
Human responsibility and AI assistance
A person runs the product interactions attributed as hands-on, in an ordinary browser, on an authorized account. A person keeps every part that needs judgment or authority: vendor terms, test accounts, anything that spends money or has a real-world consequence, what is worth testing, any subjective judgment about a product, and the final review of unusually strong claims.
Software, including AI, may assist with collecting and organizing source material, drafting, editing, calculations and quality checks. A human runs the product tests, verifies material claims against the cited evidence, approves subjective assessments, and takes responsibility for publication. AI output is never presented as first-hand experience, and an observation produced by our analysis rather than by the person at the keyboard is never presented as somebody’s lived experience with the product.
We separate what was done from how it felt. Because a person runs the workflow, a page reporting a run can say plainly what was done: “we created the customer, sent the estimate, scheduled the job”. Writing such as “we found it intuitive” or “our team loved it” requires the person who ran that session to have recorded that judgment, and “users find” requires actual user evidence to attribute it to. Without that, the page says what happened instead.
Commercial relationships and material benefits
- Affiliate programs: listed on the affiliate disclosure page and disclosed beside every commercial link, when any exists.
- Test accounts, extended access and credits from a vendor: accepted, used only within the terms they come with, disclosed on the page that relies on them and on the disclosure page, and never a condition of coverage or a factor in a conclusion. An ordinary public free trial needs no special disclosure.
- Vendor briefings and product updates: welcome; nothing said in one changes a finding unless it can be verified.
- Gifts, travel and sponsored content: not accepted.
- Ownership: Service Tech Select is not owned by a software vendor. Any ownership or investment conflict is disclosed on the affected pages before publication.
Freshness and corrections
A claim without a date is close to worthless. Each material fact carries the date it was verified, and that date moves only when that fact is rechecked. A page’s updated date changes when the page changes. A page-wide “facts verified” date appears only when every material mutable fact on the page was rechecked. We do not refresh dates to look current.
We get things wrong. Report a correction by email through our contact page, with the page address and, where you can, a link to the current vendor information. Corrections are the highest priority in that inbox.
What happens next: we verify the claim against the source ourselves rather than taking either side’s word for it. If we were wrong, we fix the page and move the affected fact’s verification date to the day we rechecked it. If a correction changes a conclusion or a decision-relevant fact, the page carries a visible note with the date, what was wrong, and how the conclusion changed. A spelling or style fix gets no note. If we were not wrong, we respond, when practical, with the evidence supporting the current wording. Corrections from vendors, readers and competitors are all handled the same way.
Bylines and accountability
The byline on an editorial page identifies the person accountable for the page’s research review and publication approval: Evan Mercer, Founder & Lead Researcher at Service Tech Select. A “Tested” credit appears only when that person performed the stated test, and only when a recorded test run backs it. AI-assisted drafting does not change the requirement for human evidence review and accountability. A byline is a claim, and we hold it to the same standard as every other claim here.
We do not borrow credentials or invent an editorial team. One accountable person, a documented process, and evidence you can check.