Editorial policy

How is our content researched?

Every page starts from the buyer's question: what a site manager or owner needs to know to hire well. We research from primary sources where they exist, such as federal agencies and standards bodies, and from the practical realities of data center roles. Claims that need a source get one, linked on the page.

We prefer official sources over summaries. For credential questions that means the agency that issues or defines the credential, such as OSHA for outreach training cards or the EPA for refrigerant certification. For industry definitions it means the association that publishes them. Each source link records the date we retrieved it, because official pages change.

We avoid claims we cannot support. We do not publish statistics without a source, invented case studies, fabricated testimonials or speed promises. Where something depends on your site, your jurisdiction or your contract, we say so rather than offering a false general answer.

Content is written in plain language for buyers. We avoid jargon where a plain word works, explain acronyms the first time they appear and keep each answer short enough to read on a phone between site walks.

How do we use BLS wage data?

We use the Bureau of Labor Statistics Occupational Employment and Wage Statistics program for pay references. Each role is matched to the closest federal occupation, labeled with its title and SOC code, and shown with the release it came from. Federal medians describe everyone in that occupation, not data center offers.

Many data center roles do not have their own federal occupation. A Critical Facilities Engineer may be matched to Stationary Engineers and Boiler Operators, a broad occupation that includes people in hospitals, plants and campuses. We say when a match is broader than the role so readers do not mistake a federal median for a market rate.

We show the current release and keep the prior release available for comparison where it helps. When the BLS publishes a new release, figures are refreshed and the release label on each page changes with them.

What do our data flags mean?

Flags tell you when a number is missing or capped. Top-coded means BLS caps the published wage at a threshold, so we show the threshold and no exact median. Not available means BLS did not publish a figure. Broader occupation means the match includes work outside data centers. We never estimate around a flag.

Filling gaps with estimates would make pages look more complete and be less honest. A hiring manager who sees a number tends to anchor on it. If the source does not support a number, we would rather show a gap and explain it.

Flags also appear in comparisons. When two roles are compared and one has a top-coded or missing figure, the comparison says so instead of dropping the role or inventing a value.

Where do we use templates?

We use templates for shared, factual elements that should read the same everywhere: forms, fee statements, the replacement guarantee wording, trust notes, process step names and data flags. Templates keep legal and commercial language consistent. Editorial prose that answers a page's own question is written for that page and is not counted as template.

Template blocks are marked in the page source so our checks can separate them from editorial content. When we measure how much original writing a page carries, shared forms, trust strips, navigation and lists of linked titles are excluded. A page cannot reach its word floor by repeating a block that appears elsewhere.

Role pages share a structure, but structure is not content. Two pages may both have a section on screening, yet the screening points for a commissioning agent and a security supervisor should have nothing in common beyond the heading.

How are drafts checked before release?

Drafts pass automated validation first, then human review before release. Automated checks look for missing sources, unsupported numbers, banned phrasing, em dashes, duplicate paragraphs, word floors on unique prose and required fee and guarantee wording. A person then reviews the page for accuracy, tone and usefulness before it is published or updated.

Automated checks catch consistency problems that people miss when reading quickly: a fee sentence that drifted from the approved wording, a page whose answer paragraph grew too long, or a figure that no longer matches the current release. They cannot judge whether advice is sound, which is why the human review step exists.

This policy describes the workflow every page is meant to pass through. A page that fails an automated check is held until it is fixed. A page that has not yet been reviewed by a person is not considered final.

How do we correct mistakes?

If you find an error, contact us and tell us the page and the problem. We check the claim against the source, correct the page if it is wrong and update the retrieval date on any source we recheck. Material corrections to data or definitions are made openly rather than quietly rewritten.

Our pages are written to help you hire, and some of them invite you to contact us. That commercial interest is real, so we keep advice separate from sales. Where another model, such as staffing or RPO, would serve you better, our guides say so.