How to Compare Content Agencies by Process

Every content agency pitch leads with a portfolio. Here's a piece that ranked, here's a client whose traffic went up and to the right, here's a sample article written in a convincing voice. The problem is that a portfolio is a highlight reel. It shows you the best output from the best month, often written by a senior person who no longer touches your account after the contract is signed. It tells you almost nothing about whether the agency can produce that quality on the fortieth article, nine months in, after the founder has moved on to the next sale.

Process is what you're actually buying. A monthly content program is a repeatable system: research, briefing, drafting, review, publishing, interlinking, measurement. If any stage is improvised, quality drifts. The way to compare agencies is to interrogate each stage and see who has a real answer versus who gestures at one. This guide gives you the specific questions and what good answers sound like.

Why a portfolio misleads you

A strong sample article proves three things: someone can write, someone picked a good topic once, and the agency can make a sales deck. None of those predict consistent output. The writer on the sample may be a $0.30-per-word subcontractor who won't be assigned to you. The topic may have been chosen by the founder who personally researched it and won't do that at scale.

There's also survivorship bias. Agencies show you the client whose traffic climbed. They don't show you the six clients who churned at month four because nothing moved. You can't evaluate what you can't see, so a portfolio structurally hides the information you most need. Process questions get around this because a weak system can't fake a detailed, specific answer under follow-up questions.

The seven stages to interrogate

Walk every candidate through these. Ask them to describe how each stage works on a normal account, then ask one follow-up. Vague answers and sudden hedging are the signal.

1. Topic research and prioritisation

Ask: "How do you decide what to write next, and who decides?" A good answer references a documented method — mapping the questions buyers actually ask, checking search demand, estimating competitive difficulty, and sequencing so early articles support later ones. A weak answer is "we brainstorm with you each month" or "we track trending topics." Brainstorming doesn't scale and burns your time.

2. Briefing

Ask: "What does a writer receive before they start?" You want to hear about a structured brief: target question, intended reader, required subtopics, internal links to include, word count range, and the specific angle. If briefs are thin or verbal, every article becomes a coin flip on quality.

3. Voice and subject-matter accuracy

Ask: "How do you capture my voice and avoid getting technical details wrong?" Good agencies have an onboarding step — interviews, a style sheet, sample edits you approve. For technical trades, they should have a mechanism to fact-check claims, whether that's a client review loop or a subject expert. If the answer is "our writers are great," that's not a process.

4. Drafting and who writes

Ask directly: "Who writes my articles, and will it be the same people in month six?" Then ask how writers are assigned and trained. You're trying to find out whether the sample writer is the actual writer. Turnover is the quiet killer of content programs — a new writer every few months means your voice resets each time.

5. Editing and quality control

Ask: "What happens between first draft and publish?" You want at least one editing layer separate from the writer, plus a checklist. Ask what's on the checklist. Specific answers — fact-check, link insertion, heading structure, meta fields, voice match — mean it exists. "We proofread everything" means it doesn't.

6. Publishing and interlinking

Ask: "Do you publish into my site, or hand me a document?" Handing you a Google Doc pushes the hardest, most-skipped work onto you. Ask specifically how they handle internal links between articles, because interlinking is where a lot of the compounding value lives and where most programs quietly fail.

7. Measurement and iteration

Ask: "What do you report, and what do you do when an article underperforms?" Good answers tie reporting to a goal you agreed on and describe revisiting or refreshing older pieces. Weak answers report vanity metrics — words published, pages live — with no feedback loop.

What should I ask a content agency before signing?

Ask them to describe their exact workflow from topic selection to published article, naming who does each step and what documentation exists at each handoff. The best single question is "Walk me through what happens to one article from idea to live page." A mature agency answers in minutes with specifics — a brief template, a named editing stage, a publishing method, an interlinking rule. An immature one gives you adjectives. Follow up by asking who personally will write and edit your articles and whether that stays constant. Then ask what their process is when a piece doesn't perform after three months, because the answer reveals whether they treat content as a one-time deliverable or an asset they maintain.

Is a cheaper agency with a good portfolio worth the risk?

Usually no, because a good portfolio paired with a low price almost always means the samples were produced by senior people you won't get, while your actual work goes to the cheapest available writers with minimal oversight. The math forces it: if an agency charges a low monthly rate and still profits, something in the production chain is being cut, and it's rarely the sales effort. That doesn't make cheap always wrong — some lean operations run tight systems — but it means the process questions matter more, not less, at the low end. Ask a budget agency the same seven-stage interrogation. If they have real documented answers, the low price might reflect efficiency. If they don't, you're buying the portfolio's ghost.

A quick scoring worksheet

Score each candidate 0-2 on the seven stages (0 = no real answer, 1 = partial, 2 = specific and documented). Then apply this reading:

Suppose an HVAC company in Tucson evaluates three agencies. Agency A has a gorgeous portfolio but scores 5 — the founder clearly wrote the samples and everything else is vague. Agency B scores 12 with duller samples but crisp answers on briefing, editing, and interlinking. On process, B is the lower-risk choice even though A's sample reads better, because B can repeat its result and A cannot.

Where this leaves you

Treat the sales conversation as an audit, not a demo. Bring the seven stages written down, ask the two question-based prompts above, and watch what happens when you ask the follow-up. The agency that has built a real system — managed research, structured briefs, consistent writers, a separate editing layer, publishing with interlinking, and a measurement loop — is the one worth paying for. At ClearPath Content we built our subscription around exactly those stages, which is why we're comfortable being asked about each one.

The practical takeaway: before you compare a single writing sample, write down the seven stages and make every agency walk you through theirs out loud. The one who can answer in specifics without reaching for adjectives is the one who can still deliver in month twelve.