How to Hire Your First Freelance Writer When You Don't Know How to Judge Writing

If you are about to hire first freelance writer help and you don't think of yourself as a writer, the hard part is not finding candidates. It's knowing whether what they send back is any good. This guide gives you a repeatable process you can run yourself: a short brief, the same paid test for every finalist, and a simple rubric that scores the work on things you can actually check. You don't need a good ear for prose to do it. You need to know your product, your reader, and which facts can't be wrong, and you already know those better than any outside writer will.

Here is the short version of how to hire a content writer without being one. Define one real assignment, shortlist from portfolios, ask every finalist the same questions, give each of them the same paid test, score the drafts against a rubric you wrote before reading them, and check the facts. Then test one revision before you commit.

Step 1. Define the writing work you actually need

Before you post anywhere or message anyone, write down one realistic first assignment. Most founders skip this and end up hiring for "good writing" or "content," which is too vague to test against. That's understandable if you have been writing everything yourself, because founder-led content tends to stall when other work piles up, and a hire is the natural next move.

Put these on one page:

• The deliverable, such as a how-to article, a product explainer, a case study, or a technical guide.
• The reader: who they are, what they already know, and what question or task brings them to the piece.
• The outcome: what the reader should understand or be able to do afterward.
• The business purpose: educate a prospect, support a product evaluation, or help an existing customer. Say whether the piece is mainly meant to be found in search, to convert, or both.
• Who supplies which facts: what you will hand over, what the writer must research, and what someone on your side needs to verify.
• The boundaries: rough length, deadline, format, whether it's bylined or ghostwritten, and any SEO or CMS requirements that really apply.
• The collaboration: who answers the writer's questions, who fact-checks product claims, and who gives feedback.

It helps to separate two things that get mixed up when you hire a content writer for a SaaS company. One is domain familiarity, meaning they already know your category. The other is writing and research ability, meaning they can learn a subject and explain it clearly. If you can give a writer accurate product context and answer their questions, the second one is usually what you are really buying, and your test should measure that.

How you know it's done: you can describe one representative assignment, its reader, its purpose, and its must-have facts, and you can say what a good draft would let the reader do.

Common mistake: hiring for "good writing" without naming the content type or the reader. A polished portfolio in an unrelated genre doesn't show that someone can write the kind of article you need.

Step 2. Write one fair brief and decide your criteria first

Next, turn that page into a short brief you can send to every finalist. It should give a writer enough to do the work without guessing, while still leaving room to see how they think. If you have never written one, a repeatable content briefing system is a good model for what to include.

A practical brief has seven parts:

1. Audience and task: who the reader is, what they know, and what question the article answers.
2. Purpose and main point: what the reader should take away.
3. Deliverable and constraints: format, length range, deadline, required sections, and anything out of scope.
4. Approved company facts: product capabilities, terminology, positioning, and claims that must not be misstated, plus a source document or a person to ask. Keeping your product facts written down in one place makes this part quick.
5. Research expectations: whether outside sources are expected, what quality of source you want, and how claims should be documented.
6. Voice and examples: a short description of the tone you want and, if you have one, an example, with a note on what it shows rather than a request to copy it. If your voice isn't written down yet, documenting your brand voice and tone takes an afternoon and saves a lot of back and forth.
7. Evaluation criteria: tell candidates what the sample will be judged on, such as task fit, research and accuracy, structure, clarity, and following the brief.

Give the writer the facts they need to be accurate, but don't write the piece for them. If something is unknown, mark it as unknown or tell the candidate to ask. Nobody should be expected to invent it.

Pro tip: include a real question your target reader would ask, and ask each candidate to explain the angle, sources, and structure they would use before they write. Their questions tell you a lot. A candidate who is thinking about your reader and the evidence behaves differently from one who just starts producing paragraphs.

How you know it's done: a candidate can tell what to write, for whom, why, by when, and how the result will be evaluated.

Common mistake: testing a writer on product details you never supplied, then marking them down for asking a reasonable question or for clearly labeling an assumption. How a writer handles missing information is part of what you are testing.

Step 3. Source candidates and use portfolios to shortlist only

Now find people. Referrals, professional networks, freelance marketplaces, writer communities, and published work in the kind of content you need are all fair places to look. When you look for a freelance writer for startup work, relevant B2B or SaaS samples can help you build the list, but a familiar company name or a degree shouldn't be an automatic pass or fail.

Ask each candidate for one or two relevant samples and a short explanation of each:

• What they personally did on the piece.
• Whether it was written from scratch, edited from someone's draft, or produced with an expert.
• What research or interviews supported it.
• Whether someone else substantially edited the published version.
• Which part best shows the kind of work you need.

A portfolio can show you relevance, clarity, and range. It can't prove who did what, and it usually shows a writer's best work, picked by the writer. The U.S. Office of Personnel Management makes the same point in its guidance on writing samples as a hiring tool, noting that portfolios can be hard to verify and show candidates' selected best work. So treat a portfolio as a reason to shortlist someone, and nothing more.

Some people can't share client work because of confidentiality. In that case ask for a redacted or alternate sample, or a description of their role, or just move to the paid test. Not having a shareable sample is not proof that someone can't do the work.

How you know it's done: you have a small group of candidates whose samples and explanations suggest they can do the kind of work you need, and you know what each one actually contributed.

Common mistake: picking the person with the most impressive portfolio page without asking about authorship, editorial support, or topic fit.

Step 4. Ask every finalist the same questions

A short interview helps you compare people on the same things, which matters a lot when you don't have a trained eye. Ask the same core questions in the same order, take notes, and rate each answer on one scale. That keeps you from leaning toward whoever you had the best conversation with. This is the same idea behind structured interviews, where every candidate gets the same job-related questions.

Questions that work well when you want to vet a writer:

1. "Walk me through how you'd approach this brief before drafting. What would you need to know, and what would you research?"
2. "How do you decide whether a source supports a claim you plan to make?"
3. "Tell me about a piece that required learning an unfamiliar or technical subject. What was your process?"
4. "How do you adapt a piece for a particular audience, format, and voice?"
5. "What feedback changed a draft for the better, and how did you handle it?"
6. "What kind of content is hard for you to write well, and what do you do about it?"
7. "How do you use AI, if at all, and how do you check the output?"
8. "What would you want to know about our product, reader, publishing process, or success measures before starting?"

When you listen, you are looking for a clear process, curiosity about the audience, a willingness to say what they're unsure of, and concrete examples. A good answer doesn't have to use your vocabulary. The person should be able to explain their decisions plainly and ask relevant questions of their own. If search matters for the assignment, add a question about how they would use the target query and search intent without making the piece awkward or inaccurate.

How you know it's done: every candidate answered the same core questions, and your notes record evidence tied to the job instead of an overall feeling.

Common mistake: asking only broad questions like "Are you a good writer?" or "Do you know SEO?" Nobody says no to those. Ask for a process or an example you can compare.

Step 5. Run one small, paid writing test

This step does most of the work. Give each finalist the same brief and a paid assignment that looks like the work you would actually commission. One sample won't prove everything about how someone will perform later, but it lets you watch the real task.

One workable example, and only an example, is a paid piece of about 600 words from a real brief with a 48-hour turnaround. Adjust the length and the window to fit your actual work and keep the test proportionate. Agree the fee and the deadline before anyone starts, and pay for the requested work whether or not you pick that writer.

Ask every candidate for the same items:

• The draft, in the format you asked for.
• Links or notes for the sources used, with the claims they support marked where practical.
• A short list of assumptions, open questions, or product facts that need confirming.
• Disclosure of any AI assistance, if that matters to your policy and the assignment.

Give everyone the same instructions, company facts, source access, deadline, and chance to ask questions. Tell them whether outside research is expected. If the job is technical, decide whether you are testing technical knowledge or the ability to explain information to your reader, and then supply the same technical context to everyone. The test should measure skills the writer is supposed to bring, not things you plan to teach after you hire them.

Please don't use an unpaid speculative article as a substitute for a paid trial, and don't publish or reuse the trial work unless you and the writer have separately agreed on rights.

How you know it's done: every candidate completed the same paid, job-relevant task under comparable conditions, and you have the draft plus enough source context to review it.

Common mistake: turning the test into a full article or a large deliverable, giving candidates different prompts, or changing the brief halfway through. An oversized or inconsistent test costs candidates time and makes the comparison less useful.

Step 6. Score the drafts with a rubric, then check the facts

This is the step that lets you judge writing without being a writing expert. Choose the rubric before you read any sample, and use it the same way for everyone. The one below is a practical suggestion, not a validated industry standard. Rate each dimension from 1 to 5 and keep each score visible instead of leaning on a single total.

Use steady anchors for the scale. A 1 means the requirement isn't met and substantial correction would be needed. A 3 means it meets the basic requirement and is usable with normal editing. A 5 means it strongly meets the requirement and needs little correction in that dimension. A 2 or a 4 sits between the anchors. These numbers are a comparison aid, not a scientific test, and there is no universal pass mark, so weight the dimensions that matter most for your own brief. If you want to see how a team scores finished drafts, a rubric for scoring draft quality follows a similar approach.

Check the facts without being a writing expert

1. List the most important factual claims, numbers, product statements, and recommendations in the piece.
2. Open the source behind each one and check that it actually supports the claim, not just that it mentions the same topic.
3. Compare company-specific claims with the product facts you supplied, and ask a product expert to verify anything outside your knowledge. A walkthrough on fact-checking a draft covers the same habits.
4. Confirm that numbers, dates, qualifications, and caveats haven't been changed or stripped out.
5. Notice whether the writer separates sourced facts from their own recommendations.
6. If the topic needs first-hand expertise, interviews, or original research, check whether the writer has access to those or has a believable plan to get them.

Then run a quick clarity check. After the opening, can you state the one main point? Can you tell who it's for? Is the most important information near the top? Does each section help finish the task? For a how-to, does each step say what to do and how to know it's done?

A factual error in a central claim, an invented company capability, or a missed critical instruction shouldn't be offset by nice prose. A small style preference shouldn't outweigh accurate, useful, well-organized work. Google's guidance on people-first content leans the same way, asking whether the content gives the reader a complete, trustworthy answer.

How you know it's done: every candidate has a score for each dimension, the important claims are checked, and you can explain your evaluation with examples from the draft instead of "I liked it."

Common mistake: scoring a sample on voice or grammar alone. A polished article can still be wrong, unhelpful, or unrelated to the reader's question.

Step 7. Test one revision, decide, and put the terms in writing

If one sample looks promising, ask that finalist for one small paid revision based on specific feedback, or build a single revision round into the paid test agreement. This shows you how the writer reads notes, handles factual corrections, and improves a draft. Keep the feedback tied to the brief, the reader, and the evidence. Notes like "make it pop" give you nothing to evaluate.

Good revision requests are things like a factual clarification, a clearer explanation for the target reader, or a more direct opening. Then look at whether the writer applied the note accurately, asked a question when the note was unclear, and explained a reasoned concern instead of quietly ignoring it. Don't use a revision exercise to ask for unlimited unpaid work.

Then choose the candidate whose work best meets your must-haves and whose way of working fits how your company can collaborate. A single sample is one job-relevant observation, and it doesn't guarantee future quality. If two candidates are close, consider a second small paid assignment, or have someone you trust who wasn't involved review both samples on their own.

Before the next assignment, put the essential terms in writing:

• The deliverable, format, scope, deadline, and what is excluded.
• The fee and payment timing, agreed directly with the writer.
• How many revision rounds are included and what counts as a new request.
• Who supplies product facts and who approves factual claims.
• Whether the work is bylined or ghostwritten, and how confidentiality is handled.
• What rights, license, or permitted reuse apply, and when they start.
• Any agreed AI-use disclosure or confidentiality rules.

A sample freelance agreement can help you think through scope, but it isn't legal advice. In the United States, don't assume commissioned freelance writing is automatically "work made for hire." The Copyright Office's circular on works made for hire describes conditions that include a written agreement signed by both parties and a work that fits the law's eligible categories. Rules vary by jurisdiction, so use an agreement that fits your location and ask a qualified lawyer if ownership or licensing matters to you.

How you know it's done: you have chosen one writer using the same evidence and criteria you applied to the others, and both sides have a clear written understanding of the next assignment and the rights.

Common mistake: treating a paid sample as an open-ended audition and then asking for revision after revision with no defined scope or agreement.

What to do next

This week, pick the one assignment you would really commission, write the brief, and send it to a short list of two or three finalists with the same paid test and the same deadline. Write the rubric down before the drafts arrive. That is most of the work, and it is the part that lets you judge a writer without having to be one.

A freelancer is one way to get articles written, and it's a good one when you want a person to learn your product and grow with you. Where this hire sits in the bigger picture depends on how far along your content program is, and the stages of content program maturity are a useful way to place yourself. If you are weighing other routes to the same problem, it can help to read about hiring a content person as a first move, and why it isn't the only way to scale organic. Another good read covers the roles a lean content team actually needs. DeepSmith's Content Studio is a separate option for the production side: the Writer turns a planned idea into a researched, brand-grounded article, with a cover image and publish-ready metadata, and you review, edit, and publish it. It doesn't recruit or evaluate freelancers, so it sits alongside the process above rather than in place of it. If you are curious how a production workflow runs from brief to publish, there is a walkthrough for that too. If you'd like to see how that works, you can start a free trial of DeepSmith.