Should you hire an SEO agency or do it yourself

What an SEO agency actually does each month

The service name hides the work, and the work is more ordinary than the name suggests. Almost every retainer is a mixture of four things, and the mixture varies enormously between suppliers charging similar amounts.

• Deciding what to aim at. Which questions your buyers actually type, which of them you could plausibly win, and which page should own each one.
• Producing pages. Writing new ones and rewriting old ones so each answers a specific question completely. On most engagements this is the majority of the hours.
• Fixing what stops a page being read. Speed, broken links, pages nothing links to, duplicated titles, markup that describes something a visitor cannot see.
• Getting named elsewhere. Business listings, directories that matter in your category, and being mentioned by sites that are not yours. This is the slowest part and the one most often described rather than done.

Ask any prospective agency for the split across those four, in hours, for your account specifically. The answer is the single most informative thing you will get out of a first call, and it is also the fastest way to work out whether the fee is mostly buying you writing.

The do it yourself version that genuinely works

For a small site in a local market, a determined owner can do most of the first year's work. Not all of it, and not the technical rescue if something is badly wrong, but enough that hiring anybody would be premature.

The order below matters more than the effort. Almost everybody starts by writing more, which is fifth on this list for a reason.

• Claim and complete the business profile Google shows beside the map, with real photographs, correct hours and the categories your customers would choose.
• Give every page a title that names the thing somebody would type, rather than a title that names your company twice.
• Make sure every page can be reached from another page. Anything only findable from a menu that a script builds is close to invisible.
• Fix the two or three things making the site slow on a phone, which is usually images that were never resized.
• Then write. One page per question, answering it in the first paragraph rather than in the seventh.

Two hours a week for a year, spent in that order, beats twenty hours in January and nothing afterwards. That is the whole trick and it is the part that usually fails.

When doing it yourself stops working

There are four honest signals, and none of them is that you have run out of ideas.

The first is a technical problem you cannot diagnose. If pages are not being indexed at all, or traffic fell sharply in a week without you changing anything, that is a specific fault with a specific cause and hunting for it alone can burn a month.

The second is scale. A site with hundreds of pages, several locations or a product catalogue has classes of problem that do not appear on a ten page site, and the fixes are structural rather than editorial.

The third is that you have done the list above for six months and it worked, which sounds like a reason not to hire and is the opposite. Proof that the channel responds is the best moment to buy more of it, and it is also the moment you have evidence to judge a supplier against.

The fourth is simply that it stopped. If the last page you published was in March, the question is no longer whether you can do it. Buying somebody else's obligation to produce work every month is a legitimate reason to hire, as long as you say so out loud rather than pretending it was about expertise.

What hiring one costs, and what to check

You are buying a quantity of work per month rather than a position in the results. Nobody controls the results, and an agency that offers a guaranteed ranking is either selling a query nobody searches or selling something it cannot deliver.

Because the fee tracks output, the questions that decide value are about output. How many pages, written by whom, reviewed by whom. What technical work is included as opposed to quoted separately. What off-site work genuinely happens. And what the reporting will say in a month when nothing improved, which is the month that tells you most about a supplier.

The pricing models themselves work the same way as the rest of the category, and we set them out in what a marketing agency costs.

The trap on both sides

The do it yourself trap is the third month. The first month is interesting, the second is a habit and the third collides with a busy fortnight, after which the site quietly stops changing. Search rewards continuation more than it rewards brilliance, so the failure is almost never a bad decision. It is an absence of decisions.

The hiring trap is the retainer that becomes four articles a month and nothing else. Writing is the most visible deliverable and the easiest to keep producing, so an engagement with no pressure on it drifts towards writing regardless of whether writing is what the site needs. If the technical faults from month one are still on the list in month six, you are buying a content subscription and should price it as one.

The half that neither side wants to do every week

Look at the four activities again and notice which ones are the same every single week. Checking what changed. Watching for pages that stopped being indexed. Noticing a title that got duplicated. Assembling the report. Producing a first draft of the next page. That is the half that dies first when you do it yourself, and it is the half a retainer spends most of its hours on.

It is also the half that can be built once and left running. Not the judgement about what is worth writing, which is genuinely a person's job, but the checking, the drafting and the assembling. We run our own site that way and have described the arrangement openly, including which parts still need a human, in how a seven agent SEO system works each week and in the eleven jobs inside a search system.

There is a further reason this matters now, and it is not a sales point. The kind of visit search sends is changing, and some of what an SEO retainer has traditionally produced is aimed at a click that is disappearing. That argument is worth having before you sign anything, and it is in is SEO dead now with AI.

The weekly half, priced honestly, and who signs before anything is published

The last section names the half nobody wants to do every week, and it is also the half most often priced badly. No figure is published on this site. What moves the number is how many pages already exist and what state they are in, whether a technical fault is holding the site back, how many locations you serve, and who does the checking every week. A supplier quoting on article count has priced the writing and ignored the checking, and the checking is what decides whether any of the writing gets found.

Whichever way you go, ownership settles the third month problem in your favour. The pages sit on your own domain, the Google Business Profile stays in your account, and the research and the rankings history live in accounts under your logins, so you own the system whether an agency is engaged or not. Agree that in the first week, because an agency holding the analytics property or the profile is holding the exact thing you would need in order to take the work back.

Nothing is published in your name without human approval, and that covers a page draft, a schema change and a reply to a review carrying a customer's name. Where the weekly checking turns up something ambiguous, it hands it to a person with the finding attached rather than acting on it. That matters more in search than it sounds, because a page published about your business that is not quite true is a correction you end up making in public.

There are four co-founders here and no account layer between them and the work, which matters on a page about hiring, because the person taking the call is the person doing the job. Haad, who owns growth and client solutions, takes the first one. The company is based in Karachi, bills month to month rather than on a notice period that runs one way, and is answerable for what it operates across 25 engagements in six countries. Weigh one thing against all of that: the published search work is paid rather than organic, Zavcom at 89 conversions in ninety days at $1.54 a click, so an agency carrying an organic portfolio is answering a question this one cannot.

Common questions

Should I hire an SEO agency or do it myself?

Do it yourself when the site is small, the market is local and you can commit two focused hours a week for a year. Hire when a technical fault is holding the site back, when the site is large enough to have structural rather than editorial problems, or when the work has already stopped happening. Buying somebody else's obligation to produce work every month is a legitimate reason to hire.

What does an SEO agency do?

Four things in some mixture: decide which questions to aim at and which page owns each one, produce and rewrite pages, fix the technical faults that stop a page being read, and get the business named on sites it does not own. Ask for the split across those four in hours, because on most engagements page production is the majority and the off-site work is the part most often described rather than done.

How much does it cost to hire an SEO agency?

The fee tracks output rather than results, so the useful questions are about output: how many pages a month, written and reviewed by whom, what technical work is included rather than quoted separately, and what genuinely happens off-site. Nobody controls rankings, so a guaranteed position is either a query nobody searches or a promise the supplier cannot keep.

How long does SEO take to work?

Long enough that the honest answer is months rather than weeks, and the reason is mechanical rather than mystical. A new page has to be found, then judged against pages that have been answering the same question for years. Local and low competition queries move first. What you can check early is whether pages are being indexed at all, which is a yes or no rather than a wait.

Can a small business really do its own SEO?

For a small site in a local market, most of the first year, yes. Claim the business profile, give every page a title naming what somebody would type, make sure every page is linked from another, fix whatever makes the site slow on a phone, and only then start writing. The order matters more than the effort, and almost everybody starts at the writing.

What is the first thing to fix?

The business profile Google shows beside the map, if you have a location, because it is free, quick and often the largest single source of local enquiries. After that, page titles, because a title is the strongest signal on a page and a great many of them name the company twice and the subject not at all.