Why This Decision Is Harder Than Picking an SEO Vendor
Hiring an SEO company is mostly a mechanics check: do they understand technical SEO, can they build links, do they report rankings honestly. Hiring a content marketing company is a judgment call, because the deliverable is words, and words are cheap to fake and expensive to fix. A contractor can get 40 blog posts in a year that read fine, sound professional, and never move a ranking or earn a call, because nobody checked whether the content matched what people actually search or whether it was structured so a search engine (or an AI answer engine) could use it.
That's the trap this guide exists to help you avoid. Content marketing for contractors is not journalism and it's not a company newsletter. It's infrastructure: pages built in clusters around your core services, written accurately enough that a foreman on your crew would sign off on the details, and organized so each piece reinforces the others instead of sitting alone as an orphan post nobody links to.
The market is full of generalist agencies and freelance marketplaces that treat a roofing company's content the same as a law firm's or a dentist's. They swap in your trade name and city and call it customized. That works fine for search volume math on a spreadsheet. It falls apart the moment a homeowner or a competitor reads the post and it says something no one in the trade would actually say, like calling a ridge vent a soffit vent, or claiming a service you don't offer, or quoting a price range that doesn't match how your trade actually bids jobs.
You're not shopping for a writer. You're shopping for a system: research, trade-accurate drafting, on-page structure, internal linking into a content architecture, and a way of knowing whether any of it worked. That's a different evaluation than "can this person write a paragraph."
The Silo-and-Cluster Test: Ask to See the Architecture Before You Ask About Price
The single fastest way to separate a real content marketing company from a content mill is to ask them to describe, in specific terms, how they organize content around your services. A shop that understands the work will talk about pillar pages and cluster articles: one strong service page (say, "gutter guard installation") supported by a set of narrower articles that each answer one real question a homeowner has (cost by home size, mesh versus reverse-curve, does it work with pine needles, what voids the warranty), all of them linking back to the pillar and to each other.
This structure matters for two separate reasons. First, it's how search engines build confidence that your site is an authority on a topic instead of a single lucky page. A pillar surrounded by 8-12 accurate supporting articles typically outranks a lone page targeting the same keyword, because it signals depth. Second, this same structure is what gets pulled into AI-generated answers. Tools like ChatGPT, Gemini, and Perplexity favor sites that cover a topic from multiple angles with specific, checkable facts, because that's what makes a source safe to quote. An orphan blog post with no supporting cluster rarely gets cited; a well-linked topical cluster often does.
Ask a prospective vendor these questions directly:
- "Show me an example of a pillar page and its supporting cluster you built for a trade like mine."
- "How many cluster articles do you typically build around one core service?"
- "How do you decide what the cluster articles cover?"
- "Who does the internal linking, and is it planned in advance or added after the fact?"
A shop with a real answer will talk in ranges (a mature service cluster is often 15 to 40 pages deep across a full site, built out over months, not delivered as a batch). A shop without a real answer will pivot to talking about "content calendars" and "consistent posting," which sounds active but describes volume, not architecture. Volume without architecture is why so many contractor blogs have 60 posts and zero leads to show for it.
Who's Actually Writing It, and Can They Pass a Trade-Accuracy Test
This is the question most contractors skip, and it's the one that matters most. Ask directly: who writes the content, and what's their process for getting trade details right. There are three common answers, and only one of them should satisfy you.
The first answer is "a general content team" or "our writers," with no further detail. This usually means a rotating pool of freelance generalists working off a template, filling in your city and trade name into boilerplate. You'll get grammatically correct paragraphs that say almost nothing specific to your actual work.
The second answer involves AI-generated drafts with light human editing. This is common now and not automatically disqualifying, but it needs a real trade-accuracy check layered on top, not a spelling pass. Ask what that check looks like. If the answer is vague, assume it doesn't exist.
The third answer is a shop that either employs writers with real trade fluency or runs every draft through a review step specifically checking for trade accuracy: correct terminology, correct sequencing of a job, no invented pricing, no claims about services or warranties the company doesn't actually offer. This is the bar. Run your own test before signing: ask for one sample piece written for your specific trade, on a topic you know cold, and read it the way a competitor or a sharp customer would. Does it use the right terms? Does it describe the job the way it's actually done? Would a foreman on your crew wince at anything in it? A single sample tells you more than any pitch deck.
The stakes here are higher than embarrassment. Inaccurate trade content published under your business name is a liability if it misstates a code requirement, a warranty term, or a safety practice, and it actively damages trust with the exact skeptical, detail-oriented customers your best jobs come from.
This is also where trade-to-trade differences matter more than most vendors admit. A roofing crew and an HVAC crew don't just use different vocabulary, they have different buying cycles, different emergency-versus-planned-purchase splits, and different objections a homeowner needs answered before calling. A vendor writing across 20 unrelated trades with one template can't track that. Ask how many trades they actually write for regularly, and ask them to name the two or three questions homeowners in your specific trade ask most before hiring. If they can't answer that off the top of their head, they haven't done the research yet, and you'd be the one paying for them to learn it.
What Should It Cost, and What Should You Get for It
Content marketing pricing for contractors runs in a wide band because "content marketing" covers everything from a single freelancer emailing you 500 words a month to a full silo build with strategy, writing, on-page structure, and internal linking. Broadly, three tiers show up in the market.
| Tier | What you typically get | What it's missing |
|---|---|---|
| Freelance / marketplace | A handful of standalone posts per month, minimal research, no architecture | Trade accuracy checks, internal linking strategy, any tie to rankings or leads |
| Mid-market content retainer | Scheduled monthly posts, basic keyword targeting, some on-page SEO | Cluster architecture planned in advance, trade-specific editorial review, AI-citation structuring |
| Full silo-and-cluster build | Pillar pages plus supporting cluster articles built to a topical plan, trade-accuracy review, internal linking, structured for both search rankings and AI answer citation | Usually paired with SEO and AI Search services since content alone doesn't rank itself |
A realistic silo build for one core service typically runs 94+ cluster pages across a mature site over time, not delivered in month one. Ask any vendor to walk you through their delivery pace: how many pages in the first 90 days, how many by month six, and what the cadence looks like after that. If they promise everything at once, that's a red flag; a rushed cluster usually means shallow, generic pages that don't hold up under the trade-accuracy test above.
Also ask how a quote is actually built. A vendor pricing by word count alone is optimizing for the wrong variable, since a tight, specific 700-word page that answers one real question usually outperforms a padded 1,500-word page written to hit a length target. A better pricing conversation centers on pages per month, what each page is meant to accomplish inside the cluster, and how research and trade-accuracy review are baked into that number rather than billed as an afterthought.
Whatever tier you're evaluating, ask what happens to the content if you cancel. Do you own the pages outright? Are they hosted on your domain or a subdomain you don't control? A content marketing company that hedges on ownership is telling you something about how replaceable they think the relationship is.
Red Flags That Separate a Real Shop From a Content Mill
Some warning signs show up before you ever see a contract. Watch for these during the sales process itself.
- They can't name a single trade-specific detail unprompted. If a rep pitching you a content package never mentions anything specific to your trade beyond your business name and city, the pitch is templated, and the content will be too.
- They quote a flat monthly post count with no mention of architecture. "Four posts a month" is a volume promise, not a strategy. Ask what those four posts are supposed to accomplish together.
- They can't explain how content and rankings connect. A vendor should be able to say, in plain terms, how a cluster of pages is supposed to move a service page up in rankings or into an AI-generated answer. If the explanation is just "Google likes fresh content," that's outdated and incomplete.
- No sample writing in your trade, or a sample that reads generic. Insist on one. A vendor who resists sharing a trade-specific sample is telling you the samples don't hold up.
- Pricing with no page-count or deliverable clarity. "Custom quote after a call" is fine as a process; a contract with no specified number of pages, words, or deliverables per month is not.
- No mention of internal linking or site architecture. Content that isn't linked into a structure is an orphan page. Ask directly how new content connects to your existing pages.
- They own the content, not you. Confirm in writing that published pages are yours to keep if you leave.
None of these are hard to check. Most of them come out in the first sales call if you ask direct questions instead of letting the vendor run a generic pitch. A shop with a real process welcomes the scrutiny; a shop selling volume gets vague fast.
How Content Marketing Fits With SEO and AI Search (and Why You Usually Need Both)
Content marketing is the raw material. It is not, by itself, the whole job. Well-written, trade-accurate pages built into a silo still need the mechanics that make them rankable: technical SEO, internal and external linking strategy, and increasingly, the structured data and entity signals that determine whether an AI answer engine cites your business by name instead of a competitor's.
This matters when you're evaluating a content-only vendor. If a company only writes content and has no answer for how that content gets crawled, indexed, and ranked, or cited in an AI Overview or a ChatGPT answer, you're going to need a second vendor for that layer anyway. Some contractors run content and SEO as two separate relationships on purpose, because they want a specialist in each lane. That works, but it requires the two vendors to actually coordinate, which in practice often doesn't happen, and you end up as the go-between.
The other path is a single shop that treats content, SEO, and AI-search structuring as one connected system: content gets written to be trade-accurate and topically deep, then structured with the schema and internal linking that let search engines and AI tools actually use it. That's a fair thing to ask about directly: "If I hire you for content, who handles making sure it actually ranks and gets cited?" A company with a real answer will describe how their content process feeds directly into technical SEO work, not treat it as somebody else's problem.
Either way, go in with realistic timeline expectations. Content built into a proper cluster typically starts moving competitive terms in the 4 to 9 month range, not weeks. Anyone promising faster on competitive trade keywords is either overselling or targeting terms with no real competition, which isn't the same as winning the terms that matter.