Choosing an agency · 12 min read
What an SEO agency actually does, month by month
If you have never bought SEO, the fee is abstract and the deliverables are vague. This sets out what the work consists of, what belongs in a report and what is padding, how long before anything is visible, and the conditions under which the honest answer is that your money should go somewhere else first.
Written by Zubair Afzal, FounderUpdated
The real anxiety
What people are actually worried about before they sign
Nobody hesitating over an SEO proposal is confused about what search engines are. They are trying to work out what arrives in exchange for the money, and most proposals go out of their way not to say.
- A monthly fee with no list of what turns up for it.
- You are quoted a number and given a paragraph about strategy. Nothing states what exists at the end of month one that did not exist at the start. This is fixable in a single question — ask what will have changed on the website by a named date — and the quality of the answer tells you most of what you need to know about the supplier.
- A proposal that lists activities rather than a plan.
- Keyword research, on-page optimisation, content creation, link building, monthly reporting. Every proposal contains these words and they carry no information, because they describe categories of effort rather than decisions about your business. A plan says what is wrong now, what will be done about it first, and why that order.
- Nobody will say what working looks like, or when you would know.
- Ask what should be different in three months and the answer arrives wrapped in caveats about how search takes time. Some caution is honest — nobody can commit to a position or a volume. But an unwillingness to name any observable change at any date is not caution. It is an arrangement designed to be impossible to fail.
- Reports that are long, colourful and answer nothing you asked.
- Forty pages of charts arrive each month. Traffic is up, impressions are up, a graph goes to the right. You still cannot say whether the money is producing customers, what was actually done, or what happens next. Length is doing the work that clarity should be doing.
The work itself
What the work consists of, in the order it usually happens
The proportions shift as the engagement matures. The sequence rarely does, and a supplier who starts at step four is guessing.
Establish why you are not being found
Before anything is produced, somebody has to find out what is actually wrong. Pages that cannot be indexed, a site structure that splits authority across near-identical addresses, content aimed at terms nobody uses commercially, a brand nobody links to. The diagnosis determines everything after it, and it is the step most often skipped in favour of publishing.
You get: A written diagnosis with faults ranked by cost of leaving them
Clear what stops pages being crawled, indexed and loaded
Broken canonicals, accidental blocks, redirect chains, duplicate addresses, slow loading on mobile connections, templates that bury the content. This is unglamorous and frequently the highest-value month of the engagement, because until it is done nothing else that gets built can perform.
You get: A log of shipped technical changes with before-and-after evidence
Decide what the business should be findable for
Not the highest-volume terms. The terms that produce customers who are profitable and who you can serve. This means understanding your margins and your best accounts, which is why the good version of this step involves your sales conversations and not only a research tool.
You get: A ranked query list justified by commercial value, not search volume
Build or rewrite the pages that answer those searches
Real pages, published, not a calendar of intentions. Each one has to be better than what currently occupies the results for its query — more useful, more specific, better evidenced — because parity with the incumbent is not a reason for anything to change.
You get: Published pages, each with a stated query and a reason it should win
Earn the authority that lets those pages compete
Coverage, mentions and links from places with their own standing, obtained by being worth referencing. This is the slowest and most expensive part of the work, and the part where the difference between a serious supplier and a cheap one is most obvious.
You get: A record of earned coverage with the reason each was obtained
Measure, then change the plan
Read what happened, decide what it means, and adjust. This is the step that distinguishes a programme from a subscription: if the plan in month eight is identical to the plan in month two, nobody has been reading the results.
You get: A short monthly note on what changed, what it means and what happens next
The alternatives
What an agency gives you that the alternatives cannot, and the reverse
An agency
Several specialisms on one account — technical, editorial, outreach, measurement — without hiring four people. Cover when somebody is away, and pattern recognition from working across many sites at once. You pay a margin for that, and you rarely get the most senior person on the day-to-day work.
A freelancer
The lowest cost per hour of real expertise, and often the sharpest single opinion available to you. The constraints are capacity and breadth: one person cannot be excellent at technical work, writing, outreach and reporting at the same time, and a holiday is a single point of failure.
An in-house hire
Context nobody external can match. They know the product, the sales team and why last year went the way it did. Slower to get moving, harder to replace, and usually deep in one discipline rather than spread across four. It works once there is enough continuous work to justify a full role.
Reporting
What belongs in a monthly report, and what is padding
A report exists so you can decide whether to keep paying. Judged against that purpose, most of what arrives in one is decoration: impressions charted without clicks beside them, tool exports nobody opens, and position screenshots for phrases no customer has ever typed.
The test is simple. If you finish reading and still cannot say what was done, whether it worked, or what happens next, the document was written to justify an invoice rather than to inform a decision.
- What shipped this month, in plain sentences, with links to the pages that changed.
- Enquiries, calls or bookings attributable to organic search — the business outcome, not session counts.
- Movement on the specific queries that were prioritised, with the ones that went backwards included rather than omitted.
- What was learnt that changed the plan, and what the plan is now as a result.
- Anything blocked, who has to unblock it, and what it is costing to stay blocked.
- Search Console figures read honestly: clicks and queries, rather than impressions used to imply progress.
- A brief note on competitors who moved and what they did, where it is genuinely relevant.
- Next month in three or four concrete items, present before you have to ask for them.
The money question
When organic search is the right first investment, and when it is not
Read down the right-hand column first. If two or more of those describe your business, spend the money elsewhere for now.
| Dimension | Conditions where organic search usually pays | Conditions where something else should come first |
|---|---|---|
| Margin per customer | Enough gross margin that one additional customer a month covers a meaningful share of the fee. | Thin margins on low-value orders, where the sum never closes at any volume the category could realistically produce. |
| How people find you today | People are already searching for what you sell, and finding somebody else. | Nobody is searching yet because the category is new. Demand has to be created, and search only captures it. |
| What the site does with a visitor | The website converts well enough that more of the right visitors is genuinely useful. | The website does not convert. Sending more people to it multiplies a loss rather than fixing one. |
| The sales process behind it | Enquiries get answered quickly and followed up properly by somebody accountable. | Enquiries sit for days before anyone replies. Fixing that costs nothing and works this week. |
| Urgency | You can fund two or three quarters before the work has to justify itself. | You need revenue this quarter. Buying demand does that; earning it does not. |
| Who occupies the results | Your competitors are ordinary at this and there is ground available to take. | National publishers, directories and marketplaces hold everything, backed by a decade of accumulated authority. |
| Budget available | Enough to resource one programme properly for a year rather than partially for two. | Barely enough for one, and most of it already committed to something you cannot stop. |
| How long you will own the business | You expect to still be running it in three years and to benefit from what compounds. | A sale, closure or pivot is planned inside eighteen months, so the payback lands after you have gone. |
The honest part
The answer nobody selling this wants to give
For a real share of businesses, the best first marketing investment is not search at all. It is replying to enquiries faster, fixing the page that loses people at the form, or buying demand directly while organic work would still be warming up.
None of that is an argument against SEO. It is an argument about order. A search programme multiplies whatever the business already does with a visitor, and multiplying something that does not work produces a larger version of the same problem at greater expense.
The most useful thing an agency can do in a first conversation is work out which situation you are in. If your enquiry-to-customer rate is poor, that is where the next pound belongs. If your website loses people at a specific step, fixing it raises the return on every channel you run, including the ones you already pay for.
When search is the right answer, judge it on a stated expectation rather than a feeling. Agree at the outset what should be observably different at three months and at six, write it down, and review against it. Nobody can commit to a position or a volume, and anyone who does is telling you something about themselves. But everybody can commit to what will have been built, tested and learnt by a given date — and that is a perfectly reasonable thing to hold a supplier to.
Questions
Questions from people buying this for the first time
What does an SEO agency do each month?
Four things, in varying proportions: fix what stops pages being found and loaded, decide what the business should be findable for, build or rewrite the pages that answer those searches, and earn the outside authority that lets those pages compete. Measurement runs underneath all four and changes what happens next month.
The proportions shift constantly. Early months are weighted towards diagnosis and technical repair. Later months tilt towards content and authority. A proposal promising the same fixed activity list every month for twelve months has not been thought about.
How long before I see anything?
Something should be observable within about eight to twelve weeks, but it will not be revenue. It will be pages getting indexed, the site appearing for queries it did not appear for before, and technical faults closing.
Commercial results take longer and depend heavily on your starting position, how competitive your category is and how quickly changes can actually be deployed to your website. A site where nobody can ship a change for six weeks will move at that pace whatever the plan says.
How do I know whether it is working?
Watch the leading indicators before the lagging ones. New queries the site appears for, click-through rate on impressions you are already earning, pages moving from unindexed to indexed, and the quality of the enquiries arriving — not the number of sessions.
Then hold the agency to a stated expectation. Ask at the start what should be different by month three and by month six. Any answer is fine as long as it is specific and written down, because it gives you something to check rather than a feeling to argue about.
Is SEO worth it for a small business?
It depends on three numbers: what you make from a customer, how many customers you could realistically add, and how long you can fund the work before it has to pay for itself. If one extra customer a month would cover a meaningful part of the fee, the arithmetic usually closes.
If your margin per customer is small and the category has limited search volume, it frequently does not close at any level of effort. That is worth working out on paper before signing anything, and a good agency will do the sum with you even when it argues against them.
What should be in the monthly report?
What shipped, in plain sentences, with links to the pages that changed. Movement on the specific queries that were prioritised, including the ones that went backwards. Enquiries or calls from organic search. What was learnt that changed the plan, and what the plan is now.
Padding looks like: impressions charted without clicks, a tool export nobody reads, screenshots of positions for terms nobody searches, and a list of hours spent. If you cannot tell from the report what happens next, it was written to justify an invoice.
When should I spend the money on something else instead?
When the website does not convert the visitors it already gets, when enquiries sit unanswered for days, when you need revenue this quarter, or when nobody is searching for what you sell because the category is genuinely new.
In the first two cases the fix is faster and cheaper than any search programme, and it improves the return on every other channel at the same time. In the last two, buying demand directly does something organic work cannot do at that speed.
Not sure whether this is the right spend yet
Tell us what you sell, what a customer is worth and where enquiries come from today. We will tell you whether search is the right first investment for you, and say so plainly when it is not.
Related
Where to go next
- the realistic timelineWhat changes in month two, month six and month twelve.
- what the fee is made of
- choosing between shortlisted agencies
- agency versus hire versus contractorThe staffing version of the same decision.
- the SEO work we run
- paid campaigns insteadOften the better first spend, and we will say so.
- conversion rate work
Last updated