Should you hire an SEO agency? A decision framework

We run a marketplace where businesses pay for a leaderboard position. We do not sell SEO services, we do not take referral fees from agencies, and we have nothing to gain from telling you to sign a retainer or telling you to avoid one. That makes this one of the few articles on the subject written by someone with no side in the argument.

So here is the honest position: good SEO agencies exist, they are worth real money, and most of the businesses that hire one are not ready for it. Both halves of that sentence matter.

Start with the question behind the question

Nobody actually wants an SEO agency. They want more enquiries, and SEO is one of the routes they have heard of. Before you evaluate any proposal, work out which of these you are really buying:

Capacity. You know roughly what needs doing — pages written, a site restructured, listings cleaned up — and you have nobody to do it. This is a resourcing problem. An agency solves it well, and so does a freelancer, and so does hiring someone.

Diagnosis. Traffic has fallen, or never arrived, and you do not know why. This is a one-off analytical job, not a twelve-month retainer. Work through a diagnostic checklist for a site that is not ranking before you pay anyone, because the answer is frequently something you can see yourself in an afternoon.

Judgement. You are in a competitive market, you have budget, and the decisions are genuinely hard: what to build, what to kill, where the ceiling is. This is what a strong agency is actually for, and it is the one thing you cannot easily buy cheaply.

Reassurance. You feel behind and want someone to be responsible. This is the worst reason to sign, and it is the most common one. Agencies can smell it, and the ones that price aggressively rely on it.

If your honest answer is "reassurance", stop here and go and do something you can measure instead.

The readiness gate

An agency multiplies whatever your business already is. If the underlying thing is not ready, the multiplier makes no difference. Run these five checks first.

1. Does the demand exist? People must already be searching for what you sell. If they are not, no amount of optimisation creates the searches. You can establish this for free — keyword research without a paid subscription covers how.
2. Can your site convert the traffic it already gets? If a hundred visitors a month produce zero enquiries, a thousand will produce zero enquiries more slowly. Fix the page before you buy the traffic.
3. Can you wait? Organic results are slow. If you need customers this quarter to survive, SEO is the wrong instrument — see how long SEO takes and what the evidence supports.
4. Can you supply what the agency needs? Access to the site, sign-off on content, someone who answers emails within a week. Retainers die from client-side silence more often than from bad work.
5. Can you afford twelve months? Three months of SEO is three months of cost and no result. If you can only fund a quarter, put the money somewhere with a quarter-length feedback loop.

Fail two or more and the answer is not "hire a better agency". It is "not yet".

What money actually buys

Agency pricing is opaque because the deliverable is opaque. The useful mental model is not "how much SEO do I get" but "how many senior hours am I buying, and who is doing the work".

The table below uses illustrative monthly bands to show how the shape of the engagement changes with spend. These are not surveyed market rates and not a price list — they are a worked illustration to help you sanity-check a proposal against what is physically possible at that price. Local currency and market will move the numbers considerably.

| Illustrative monthly band | Who realistically does the work | What is plausible at this level | What is not | |---|---|---|---| | Very low (a few hundred) | Largely templated, often offshore production, minimal senior input | A basic technical pass, a handful of pages, listings tidied, a monthly report | Strategy, original research, competitive markets, anything bespoke | | Low–mid | One junior or mid-level person part-time, senior oversight occasionally | Steady content production, on-page work, some outreach, quarterly review | Enough hours to move a genuinely competitive term | | Mid | A small team, a named strategist with real involvement | A researched plan, meaningful content volume, technical remediation, honest reporting | Guaranteed outcomes; still constrained by your own site and brand | | High | Senior-led, multi-discipline, embedded with your team | Deep technical work, digital PR, migration support, forecasting with stated assumptions | Miracles in a market where the incumbents outspend you by an order of magnitude |

The point of the table is arithmetic, not price-shaming. If a proposal promises a strategist, four long-form articles, technical remediation and link acquisition for a sum that could not cover forty hours of competent labour in their market, something in that list is not happening. Ask which.

Cheap is not automatically bad. A small, focused, genuinely cheap engagement — a technical audit, a one-off site restructure — can be excellent value precisely because it is bounded. It is open-ended cheap retainers that tend to produce a year of activity reports and nothing else.

Six questions that separate competence from confidence

Ask these in the first call. You are listening for texture, not for the right answer.

"What would make you tell us not to hire you?" Good agencies have a real answer — wrong market, too early, budget better spent elsewhere. An agency that has never turned work down is a sales operation.

"Show me a client where it did not work, and tell me why." Everyone has one. The quality of the post-mortem tells you how they think. Vagueness here is the loudest signal in the whole process.

"What are the first three things you would look at on our site, and why those?" You want specifics about *your* site, mentioned unprompted. Generic answers mean they have not looked.

"How will we know in month three whether this is working, given rankings may not have moved?" Competent answers involve leading indicators: pages indexed, crawl issues resolved, impressions on new queries, referring domains earned. Incompetent answers change the subject to traffic.

"Which of these results were driven by your work and which by the client's brand or budget?" Case studies attribute everything to the agency. Reality rarely does. Honesty here is rare and valuable.

"What would you need from us, and what happens if we do not provide it?" A named dependency list is a sign of an agency that has been burned by client silence and has learned to say so in advance.

Walk away if you hear any of these

These are not preferences. Each one indicates either dishonesty or a misunderstanding of how the search engine works.

• Guaranteed rankings. Nobody controls Google's results. Google's own documentation is explicit that no one can guarantee a position. A guarantee is either meaningless small print or a promise to do something against Google's spam policies.
• A fixed number of links per month. Links that arrive on a schedule were bought or placed on properties that exist to sell placements. Google's spam policies name link schemes directly, and the risk lands on your domain, not the agency's. If you want to understand why a purchased link behaves differently from an earned one, read how backlinks actually work.
• A secret proprietary method. There are proprietary tools and proprietary processes; there is no proprietary knowledge of the algorithm. Secrecy here is a device to prevent you evaluating the work.
• No direct access to your own analytics or Search Console. The data is yours. An agency that reports through a dashboard you cannot cross-check has removed your ability to audit them, which is usually the intention.
• Work performed on a domain they own. Some agencies build your content on their platform or their microsite. When you leave, your rankings leave with them.
• A refusal to name who does the work. Subcontracting is fine and common. Concealing it is not.
• Contact from a cold email claiming errors on your site. Unsolicited "we noticed problems with your website" mail is a volume business. Real agencies with capacity rarely need to prospect that way.

Contract terms worth more than the pitch

The proposal is marketing. The contract is what you actually bought. Four clauses decide whether a bad engagement costs you three months or three years.

Ownership and licence. Content, code, schema markup, and any assets should be yours on creation or on payment, not licensed to you while the retainer runs. Ask explicitly: *if we terminate, does the content stay live and does it stay ours?*

Notice and minimum term. Twelve-month minimums with ninety days' notice are common and not automatically predatory — SEO work genuinely takes time to pay back. But pair a long term with a stated break point: an agreed review at month four with a defined exit if specific deliverables were not produced. Deliverables, not outcomes; nobody can promise outcomes.

Accounts and access. Every account — Search Console, analytics, hosting, the CMS, any listings management, any tooling bought on your behalf — should be created under your ownership with the agency granted access, never the reverse. This single clause prevents the most common form of hostage-taking.

Exit handover. Specify what you receive on termination: raw data exports, documentation of changes made, credentials, a list of any third-party placements, and a redirect map if the site was restructured. Agree it while everyone is friendly.

When the answer is "do it yourself first"

For a large share of small businesses, the honest recommendation is not to hire anyone yet.

*Illustrative example.* A two-person plumbing firm serving one town has a five-page site and a partially filled Google Business Profile. The competitive set is other small local firms. The work that would move them — completing the profile properly, collecting reviews, writing a genuinely useful page per service and per area, getting listed in local directories that real people use — is a fortnight of unglamorous evenings, not a retainer. Paying a mid-band monthly fee for that is buying a project manager for a job with three tasks. Local search work rewards being present and specific far more than it rewards technique.

The general rule: if the gap between you and the businesses above you is *effort*, do it yourself. If the gap is *expertise* — a broken migration, a site with tens of thousands of URLs, a market where the incumbents have full-time teams — buy expertise. Technical problems of that kind are where an agency's hourly rate is most clearly worth paying.

And there is a third case worth naming, since we are on our own site: sometimes the honest answer is that you need visibility now and organic cannot supply it in time. Paid placement — search ads, a directory, a marketplace like ours — buys attention immediately and stops when you stop paying. We set out the trade-off in an honest cost comparison of paid versus organic, and the limits of what pay-to-rank can and cannot do is deliberately unflattering about our own model. Our public leaderboard and the spend behind every position exists so you can price that decision without asking us.

Short FAQ

Agency, freelancer, or hire? Freelancers are usually better value for a defined project and worse for breadth. Agencies buy you a bench. An in-house hire wins when SEO is a permanent, central channel — but a junior in-house with nobody to learn from is the worst of all options.

How long before I judge them? Judge the *work* from month one: is it specific, does it address your actual problems, does it arrive. Judge *results* no earlier than six months, and only against what was promised.

Should I pay for a paid audit before committing? Often yes. A paid, bounded audit from a good agency is a cheap way to test how they think before you commit to a year. Be wary of free audits generated by a tool — they are lead magnets and their findings are usually generic.

What if I have already signed and it feels wrong? Read the notice clause today, check who owns your accounts, and export your data. Then ask the agency directly for the month-by-month deliverable list against what was contracted. Most disputes are resolved by that email; the rest tell you what you need to know.