How to vet a roofing marketing agency.
The questions to ask, the contract terms to refuse, and the pattern behind most bad engagements.
The complaint, as recorded.
The recurring line in this trade is that agencies do not know a ridge cap from a rain gutter. The same playbook gets run for roofers as for dentists and yoga studios. The fee is $3,000 to $5,000 a month, the contract is long, the leads are shared, and the reports do not move the needle. That is the documented pattern across the trade press and contractor forums, not a characterisation invented to sell against.
Why roofing punishes a generic playbook.
Most local service verticals tolerate a template because the purchase is small and routine. Roofing does not. A replacement is a five-figure decision, often made under time pressure, frequently after weather, sometimes with an insurer involved. The buyer arrives with a specific question and a deadline, and a page written to fill a monthly content calendar answers neither.
What a generalist plan leaves out.
The omissions are consistent enough to predict, and each is a revenue line the site never sees. Insurance and claims research, treated as a legal question rather than a brochure. Repair intent kept apart from replacement intent, because the urgency is different and so is the buyer. Commercial specification work, which is a materials conversation rather than a curb-appeal one. Suburb-level detail nobody outside the trade knows to ask about. And the seasonal calendar, which decides publication dates months before demand arrives.
Why the reporting hides the failure.
A generic engagement reports sessions, impressions, and keyword positions, all of which move for reasons unrelated to your revenue. None of them distinguish a hundred visits to a post about roof colours from ten visits to a claims page that produced two replacements. When nobody agreed what success meant before the engagement started, a year can pass before the failure can be proven, which is exactly as long as the contract runs.
This is structural, not a moral failing.
An agency serving forty clients across twelve trades cannot afford to learn roofing on any single account, so it standardises, because standardising is the only way that business model survives. The cost of learning your trade lands on your budget, and it lands again on the next roofer they sign. Specialisation is not a virtue claim. It is the only arrangement in which research compounds instead of resetting with every client.
What vetting actually looks like.
Competence first, accountability second, price last. Most vendors are rehearsed on price and unrehearsed on structure, so the questions that separate them are the ones about how a specific page would be built and who holds the keys after launch. The full eighteen-question sequence, and the two references worth insisting on, are in the first guide below.
Read the exit before the deliverables.
The terms that cost most are ownership and duration. If the agency holds the domain, the DNS, or the hosting account, leaving becomes a migration on their timetable. If the term is twelve months with no performance condition, the contract has been written so that a disappointing outcome costs them nothing. Both are negotiable, and a reasonable vendor will accept client ownership stated plainly and a performance release on any term.
How this practice answers its own checklist.
Roofing only, so the portfolio and the research sit in one trade. Page counts specified in writing in the audit blueprint. Query research delivered as a document you keep whoever builds the site. Prices published rather than quoted. No retainer, so there is no term to exit and no notice window to diarise. Domain, files, and content yours from day one. References on request, including a client who took the blueprint and built elsewhere.
The three failure modes, named.
Bad engagements in this trade fail in one of three ways, and each has a tell you can catch before signing.
One playbook across twelve trades, diagnosed by asking how they would structure a repair page against a replacement page
Deliverables described as hours and reports described as dashboards, diagnosed by asking which single metric defines success
Ownership of domain, DNS, or hosting retained by the agency, diagnosed by asking who holds the registrar login today
What a good portfolio looks like.
Depth beats breadth. A handful of roofing clients held for years, each with a substantial site, says more than a wall of logos spread across a dozen trades. The detail nobody volunteers is turnover: which of those clients have since rebuilt with someone else. It is checkable in an afternoon, and the portfolio guide below sets out how.
How to read reviews without being fooled.
Clustered dates, generic praise with no project specifics, no mention of the trade, single-review reviewer profiles, and no negative reviews at all. Real contractor reviews name the job, the season, and usually one thing that went wrong. Cross-check on a second platform, because purchased reviews are rarely bought twice.
The reference call is the whole vetting process.
Four questions, and the second one does most of the work. What did you expect that you did not get. How long did it actually take against what was promised. Who owns your domain and hosting today. If you were starting again, would you use them, and what would you ask for differently. A vendor who will not supply two references, including one client who left, has answered the question already.
What to negotiate rather than accept.
Five substitutions a reasonable vendor will agree to.
If you are already in a bad one.
You are less stuck than it feels, but sequence matters more than speed. Read the notice window and diarise the date. Establish who holds the registrar, DNS, and hosting. Take your own copies of analytics, Search Console, form submissions, and content. Request the site files and database export in writing before giving notice, while cooperation is still commercial. Then give notice in writing, citing the clause, and migrate before the final billing period ends rather than after.
Why specialisation is the argument, not the slogan.
Every agency claims to know your industry. The checkable version is whether their research compounds: whether the query set they built for the last roofing client makes the next one cheaper and better, or whether each engagement starts from nothing. One trade means one accumulating body of work, which is the only reason a small practice can outresearch a large agency, and it is the entire reason this one is roofing only.
Every guide in this pillar.
Eighteen vetting questions in the order that disqualifies fastest, and how to read the answers.
Long lock-ins, borrowed proof, vague reporting, and the ownership clause that keeps your own site hostage.
One template strategy applied across unrelated trades, billed monthly. The pattern is documented, not alleged.
Seven steps that separate real roofing work from a grid of screenshots, plus how to spot fabricated reviews.
An honest comparison, including the real argument in favour of a longer term.
Nine steps for a contractor stuck in a contract that is not working, in the order that protects the asset.
Questions about this pillar.
Are all generalist agencies bad?+
No. A good generalist beats a bad specialist. But a generalist has to learn your trade on your budget, and roofing carries enough specifics, claims work, seasonality, commercial systems, that the learning is expensive and usually incomplete.
What contract length is reasonable?+
For a build, a fixed scope with a defined end. For anything recurring, month to month. Twelve-month lock-ins with no performance condition exist to survive the period when results have not arrived.
Would Roof-Rank pass its own checklist?+
That is the point of publishing it. Roofing-only portfolio, references on request including a client who left, page counts in writing, prices published, no retainer, and the blueprint yours whoever builds it.
What if I am mid-contract with someone else?+
The audit is standalone and the blueprint is yours regardless of who builds from it. Plenty of contractors buy it to find out whether the current engagement is underperforming or whether their expectations were wrong.
Do you take on work you think will fail?+
No. If the revenue does not support the build, or the market is not winnable at your job value, we say so at the audit stage. That is cheaper for you and better for us than a case study that never happens.
Vet us the same way.
Ask us all eighteen. The audit is $2,500 and answers most of them.