How to choose a content writing agency: the procurement test that actually works

Last reviewed: May 2026

Content writing agency procurement is unusually noisy compared with other marketing procurement. The capability decks across providers look similar. The customer logo lists overlap. The case studies are written by the same kind of marketing teams using the same kind of language. The procurement step most buyers default to (reviewing decks and shortlisting from inbound) is the procurement step most likely to produce the wrong outcome. The procurement steps that produce the right outcome are different and require more buyer-side work.

The sample-output test as the single most useful procurement step

The sample-output test is the most reliable procurement signal available and is used by less than a third of buyers in the procurement processes that produce the best outcomes. The test works like this: the buyer provides a real cluster brief for an article the engagement would actually produce. The prospective provider produces one full article against that brief at their own cost as part of the selection process. The buyer evaluates the output against the standards the engagement requires.

The test costs the provider one article's worth of time, which is meaningful but proportionate to the deal value. It tells the buyer more about the provider than any number of case studies because it is the actual work product against the actual brief. Providers who decline to participate are eliminated.

The named-writer disclosure question

The procurement question that produces the most signal in a single answer is "who specifically will write our content and what is their relevant operator or sector background." Specialist providers answer this directly with named writers, their backgrounds, and relevant published work. Generalist providers deflect with "our team of experienced writers" or "we assign based on availability."

The deflection is the answer. A provider unwilling to commit to specific named writers during procurement is a provider unwilling to assign sector-specialist writers to the engagement. Specialist content writing services generally publish their writers' bylines on the work itself, which makes the disclosure question simpler to answer.

The reference question that matters

The most valuable reference call is the one with a client who has worked with the provider for at least 12 months and is willing to talk candidly. The most diagnostic reference call is with a client who has recently churned, because it surfaces the failure modes the provider is most likely to repeat.

The procurement red flags

The patterns that predict bad outcomes in content writing procurement include:

• Refusal to disclose the writer bench by name and background. A specialist provider has nothing to hide on this; a generalist provider has reasons to deflect.
• Refusal to produce a paid or unpaid sample article during procurement. The reasons given are usually around protecting team capacity, but the underlying signal is that the sample would not look as good as the case studies suggest.
• Capability decks heavy on logos and light on actual published output. Logos can be acquired through small project work; published output is the proof of capability.
• Pricing materially below the specialist tier band for the claimed work. Either the pricing is unsustainable and quality will degrade, or the work being quoted is not what the buyer thinks they are buying.
• Aggressive sales motion with quick close pressure. Specialist providers with full client books generally do not run aggressive outbound; the procurement pattern that finds them is more relationship-based.

The procurement steps that work

The full working procurement sequence: identify candidate providers through reverse search (find the work, identify the producer) rather than from inbound; ask the named-writer disclosure question and the sample-output test; verify with references in the buyer's exact vertical; evaluate against a defined cluster plan with measurable outcomes; structure the engagement on monthly retainer with a 90-day diagnostic period before committing to longer term.

When agency procurement is the wrong frame entirely

The honest cases include: very small content needs (under 4 articles per month) where a single specialist freelance writer outperforms an agency engagement; ultra-specialist needs in narrow categories where one named expert is the right answer rather than a team; and organisations with mature in-house editorial functions where external production supplements rather than replaces internal capacity.

For most mid-market organisations seeking specialist content production at scale, the agency procurement frame is the right one and the steps above produce materially better outcomes than default procurement.

A worked example: the B2B SaaS firm that used the sample-output test

A $12M ARR HR tech platform in London was selecting a content agency for a 12-month engagement to build a cluster around workforce analytics, people data compliance, and HR technology evaluation. Three agencies responded to their shortlist request. Agency A provided a detailed capability deck with 23 customer logos, a content process diagram, and per-article pricing at £320. Agency B sent a team credentials document, two case studies from the healthcare industry, and pricing at £480 per article. Agency C sent a two-page capability statement and asked if they could submit a sample article against the buyer's brief as part of the selection process.

The buyer agreed. Agency C produced a 2,100-word article on GDPR Article 9 special category data requirements for HR analytics platforms within 5 business days, with named writer credentials (the writer had previously worked as a data protection officer at a mid-size SaaS firm), correct citations to the GDPR text and the ICO's HR-specific guidance, accurate framing of the Article 9 explicit consent requirement versus the substantial public interest condition, and a practical section on how an HR platform's data processing agreement should reflect the Article 9 processing basis. The buyer's head of data science read the article and said it was "actually useful." Agency A and B were not given the sample opportunity because the buyer had already identified their preferred provider.

Agency C was engaged at £680 per article, materially above the initial budget target of £480. The buyer justified the rate increase to the CFO by projecting that Agency A's content would require 3 hours of internal review per article (£95 per hour, 12 articles per month: £41,040 per year in internal review cost) versus Agency C's estimated 30 minutes per article (£7,695 per year). The net cost of Agency C including internal review time was lower. The sample-output test produced this ROI-based procurement justification. A an industry-specialist content writing service offers the sample-output test as a standard procurement step because the content the test produces is evidence the buyer needs to make the investment case internally.

The reference call questions that produce signal

Most reference calls in content agency procurement are wasted because the questions asked are too generic to produce useful signal. "How was working with them?" produces a positive answer regardless of the underlying reality because the referee was chosen by the provider. The questions that produce signal are more specific and more uncomfortable. "What was the first-draft acceptance rate on their content against your brief standards?" A provider whose references report 50% to 70% first-draft acceptance is a Tier 2 provider regardless of their Tier 3 positioning. "What was the compliance failure rate on their content in a regulated context?" A rate above 10% signals a provider whose writers do not have genuine regulatory fluency. "If you started the engagement again, what would you specify differently in the contract?" This question surfaces the failure modes the provider is most likely to repeat because it asks for the buyer's own assessment of where the engagement underdelivered without putting the provider directly on trial.

The most diagnostic reference call is the one with a client who has recently churned. Most providers will not voluntarily offer churned clients as references. A buyer can identify churned clients by asking "who else have you worked with in our vertical" and then independently verifying whether those firms are still clients. If a firm that worked with the provider 18 months ago is no longer a client, asking them for a candid debrief will surface exactly the failure modes that live clients will not discuss. The risk to the provider of this approach is that buyers with the diligence to pursue it are the buyers most likely to hold providers to a high standard throughout the ena sector-trained content writing service service with full client books provides churned client references proactively because the programme failures are recoverable and the transparency builds the buyer's trust in the engagement that follows.

The 90-day diagnostic structure that works

The engagement structure that produces the best outcome for both buyer and provider is a 90-day diagnostic period with defined output commitments and a structured review at the 90-day mark before the longer engagement begins. The diagnostic produces: 8 to 12 articles against a defined cluster brief, with agreed quality metrics including first-draft acceptance rate, compliance pass rate, and internal link completion rate; a cluster planning output covering the full topic taxonomy for the 12-month engagement; and a measurement infrastructure setup including GA4 attribution configuration, CRM field setup, and the initial baseline reporting. At the 90-day review, the buyer evaluates actual output against the quality metrics, decides whether to extend to the 12-month engagement at the same or adjusted rate, and has a cluster plan and measurement infrastructure in place that makes the 12-month engagement productive from month 4 onwards rather than from month 8 or 9 when a slower onboarding would have produced comparable output. A specialist content writing service structures its commercial arrangements to accommodate this diagnostic period because the results it produces are the evidence the provider needs to justify the longer engagement.

Red flags in agency capability decks

Content agency capability decks share a predictable structure across the spectrum from Tier 1 to Tier 4. The elements that correlate with Tier 3 quality are: named writers with verifiable sector-specialist credentials featured prominently; examples of published work in the buyer's specific vertical with SERP position evidence; a clearly described editorial workflow with specific process steps; transparent pricing that aligns with the Tier 3 range; and client references from current engagements in the buyer's vertical. The elements that correlate with Tier 2 agencies presenting as Tier 3 are: customer logo arrays as the primary credibility signal; case studies from adjacent verticals rather than the buyer's specific sector; pricing described as "starting from" a Tier 2 floor with qualification language about specialist premiums that will emerge after signing; vague descriptions of the writing team as "experienced specialists" without named individuals or verifiable credentials; and aggressive close pressure suggesting the agency has capacity to fill rather than a full client book.

The deck element that is most diagnostic is how the agency presents its writers. A Tier 3 specialist provider features specific named writers with verifiable backgrounds in the relevant sector. A Tier 2 provider with Tier 3 positioning features generic descriptions of the writer pool without names or verifiable credentials. The named writer disclosure question in procurement is designed to surface this difference; the capability deck often contains a preview of how the agency will answer that question. A specialist content writing service publishes its writers' credentials alongside their published work, which makes the named writer disclosure question answerable before it is even asked.

Frequently asked questions

Should buyers pay for sample articles during procurement?

Sometimes. Larger engagements often justify a paid sample at agreed rate, which respects the provider's time and provides a more genuine production sample. Unpaid samples are appropriate for smaller engagements or as part of competitive selection between final shortlisted providers.

How many providers should be on a procurement shortlist?

Three to five for most procurement processes. Fewer than three reduces comparison value; more than five exceeds the buyer's evaluation capacity and dilutes engagement on each provider.

Should procurement involve compliance or legal review early?

Yes for regulated verticals. Bringing compliance into the procurement conversation at provider shortlist stage surfaces compatibility issues early and helps select providers whose workflow integrates with the firm's compliance process.

What contract terms are standard for content writing engagements?

Monthly retainer, IP transfer on payment, termination with 30 to 60 days notice, named-author byline rights, and confidentiality covering the buyer's commercial information. Standard terms are negotiable and should be reviewed against the engagement specifics.

How does procurement differ for AI-content services vs human-led services?

AI-content procurement should focus on disclosed AI policy, named human accountability, and the sample-output test. The selection signals are similar to human-led procurement but with explicit attention to the human accountability question.

Sources

• Content Marketing Institute research
• B2B procurement insights - Gartner
• Search Quality Rater Guidelines - Google