How to choose a white-label content provider: 12 questions

When you resell a provider's articles, your name sits on their work. If a number is wrong or the prose reads like a template, your client blames you, not the vendor. To choose a white-label content provider well, you need 12 questions and, for each one, the document you can ask to see, so your choice rests on proof and not on a sales page. We are a provider ourselves, so we answer where we can for our own process too, and we say where we only tell you what to ask.

What "white-label" should mean in practice

White-label means the provider writes and you brand and resell. The client sees your agency, not the vendor. That is the whole definition, and it is why the quality question matters more than usual: you cannot point at the vendor when something goes wrong.

The difference from hiring freelancers is mostly who holds the process. A freelancer's quality depends on one person's habits. A provider should have a process you can inspect: a brief, a check by someone other than the writer, and records. Ask for those records, not for assurances.

Quality and process

1. How is each article briefed?

A brief decides what the article must answer. A brief written from what currently ranks for the query is better than a keyword list and a word count. A good answer names where the brief comes from and what it contains: the reader's situation, the questions to answer, the outline, the sources to use.

Ask to see: a sample brief for a real query, not a template.

2. Who checks the draft, and are they independent of the writer?

A writer who checks their own work misses the same errors twice. Look for a separate checker who works from the draft, the brief and the sources only. In our case the editor is a separate AI session that never sees the writer's reasoning, only the draft, the brief, the sources and the voice doc.

Ask to see: the checker's notes on a past draft, including one that was sent back.

3. Is there a published rubric and a pass mark? What happens to drafts that fail?

A score means little unless you can read what it measures. A good answer is a written rubric, a fixed pass mark, and a rule that failing drafts are rewritten, not shipped. We fail drafts below 80 out of 100. Our public rubric is the one to compare with whatever any other provider shows you.

Ask to see: the rubric and the score on each article you receive.

4. Are numbers and claims traced to sources you can open?

Unsourced statistics are the fastest way to embarrass an agency. A good provider cites every factual claim to a source, and the checker compares each number with that source. Where a claim cannot be traced, it should be cut, not softened.

Ask to see: the source list for a sample article. Open five links yourself.

5. How is originality checked, and can you see the result?

You need to know the article is not lifted from elsewhere. A good answer names the check, says when it runs, and hands you the output. Be wary of "we check for plagiarism" with nothing to show.

Ask to see: an originality report for a delivered article, and the rule for what happens when it flags a match.

6. Does it match your client's voice?

Generic output is the common complaint about AI-written articles. A provider should hold a voice document per client: audience, tone, point of view, banned words, preferred terms, sample paragraphs. The checker should score against it.

Ask to see: a voice document and an article written against it. Check whether the banned words really are absent.

Operations

7. How do articles reach your client's CMS, and can they arrive as drafts for approval?

You want to approve before anything goes live under your client's brand. Ask whether delivery is by file, by direct CMS connection, or both, and whether a draft-only mode exists. Get the answer in writing, because this is easy to promise and hard to retrofit.

Ask to see: a demonstration of a draft arriving in a test site.

8. What are the turnaround and revision terms?

Your client deadlines are yours, not the vendor's. A good answer gives a stated turnaround per article length, says how many revisions are included, and says what counts as a revision versus a new brief.

Ask to see: the terms in writing.

9. How is volume handled without quality dropping?

Ask what changes at 5 articles a month versus 50. Google's guidance says that using generative AI tools to produce many pages that add no value for users may break its spam policy on scaled content abuse [1]. The risk at volume is not AI itself but pages that add nothing. A good provider can show that the same checks apply to every article, not a sample of them.

Ask to see: scores for a batch, not a hand-picked example.

10. How are AI tools used and disclosed, and who is accountable?

Google's guidance does not prohibit AI-written content; it describes generative AI as useful and warns about scale without value [1]. It also calls it critical to review AI-generated content by hand for accuracy and trustworthiness before publishing [1]. So ask who does that review and where it is recorded. Google adds that telling readers how a piece was made can give them useful context [1]. Decide with your client whether to disclose, and ask the provider to support either choice.

Ask to see: a written description of where AI is used in the process and who signs off.

Commercial

11. Who owns the content, and is your client told about the provider?

Get ownership and confidentiality in the contract: that you own delivered articles, that the provider will not be named to your client, and that no provider branding appears in the output. We have not set out our own terms here; ask any provider, us included, for the contract text before you commit.

Ask to see: the ownership and confidentiality clauses, and any NDA.

12. What does it cost, and is pricing public?

Public pricing lets you check the margin before a sales call. Gated pricing is not automatically bad, but it should come with a written quote and a sample before you spend time. Work out your cost per article against what you charge the client, including revision rounds. See our plans and prices.

Ask to see: a written quote that lists what is included.

Red flags that should end the conversation

• Pricing hidden and no sample offered.
• No named process, only adjectives.
• Promised rankings. Nobody honest can promise them.
• Refusal to show sources or the checker's notes.
• Vague quality claims with no score, rubric or report behind them.
• No answer to who is accountable when an article is wrong.

How to test a provider: a small paid pilot

Order two or three articles on real client queries. Pay for them, so the provider treats them as real work.

Score each article with your own rubric before you read the provider's score. Open every source and check the number against the page. Run your own originality check. Then compare your scores with the provider's: if they differ widely, you have learned how much to trust the provider's numbers.

If you want a starting point for your own scoring sheet, read the rubric every article must pass.

Where we fit (and where we don't)

We write articles from a brief built on what is ranking, with sources you can open, and an editor that fails drafts below 80 out of 100. We do not promise rankings, and we make no claims here about customer results. If you need a provider who will say yes to anything, we are the wrong fit.

See how white-label works for agencies.

Sources

1. Google Search's guidance on using generative AI content (context section) — Google Search Central, 2026-10-01

This article was researched, written and edited by AI under human supervision, and checked against Rankwrite's public editorial rubric.