What It's Actually Like to Work With an SEO Agency, Month by Month

Weeks 1–4: you will spend more time on logins than on SEO

The first month of an SEO engagement is administrative, and it disappoints almost everyone. You've just committed real money and the visible output is a shared folder, a spreadsheet and three requests for access you have to chase your developer for.

That's the correct shape. Everything after this depends on two things being right: complete data access, and a written starting number. Get either wrong and you'll spend the next eleven months arguing about whether anything worked.

Here's what a good agency asks for in the first fortnight, and why each one matters more than it looks.

Month 2: things ship, and they're all invisible

Month two is when work starts landing and none of it looks like anything. Redirect chains collapsed. Twelve pages de-indexed because they were thin duplicates. Canonical tags fixed. Forty title tags rewritten. A pillar page published that nobody has read yet.

This is the stage where founders start forwarding the report to their co-founder with a question mark. The honest framing: you're paying for plumbing, and plumbing is only noticeable when it's broken. On an older site — anything that's been through a redesign or a platform move — this month often recovers traffic that was quietly lost years ago, with no new content at all.

What you should actually feel in month two is a small, specific relief: someone finally knows why that section of your site never got indexed.

• What you'll see: a fix log, a crawl before-and-after, the first published pages, an internal-link plan.
• What you won't see: traffic. Google has to re-crawl, and re-crawl frequency on a small site is measured in weeks, not hours.
• What to check: indexed page count in Search Console versus what you published. If you shipped eight pages and two got indexed, say something now, not in month five.

Month 3: the trough, and why almost everyone quits here

This is the month the whole engagement turns on, so we'll be precise about it.

By month three, if the work is good, your Search Console impressions are climbing — sometimes steeply. Your average position is improving. Your traffic has barely moved and your leads haven't moved at all. The report looks like progress and the bank account looks like nothing happened.

The mechanic behind that is not mysterious. A page that moves from position 58 to position 19 has improved enormously, and it earns you almost no clicks, because click-through rate collapses below the first page. Impressions count every time Google shows you; clicks only happen when someone chooses you. You've gone from invisible to visible-if-you-scroll. That's real progress that pays nothing yet.

So month three feels like being sold to. The agency is genuinely pleased, you're genuinely not, and both of you are looking at the same chart. It's the single most common point at which a working engagement gets cancelled — and cancelling here means paying for the entire cost of SEO while collecting none of the return.

The correct response is not faith. It's two specific checks.

1. Are impressions rising for the queries in the plan, or for junk? Filter Search Console to your target queries. Rising impressions on "free", "salary" and "meaning" variants of your keywords is not progress, it's drift.
2. Is the position band moving in the right direction? Count pages sitting in positions 11–30 for commercial queries. That band is the waiting room. If it's filling up, month four to six pays you. If it's empty and impressions are up, you're gaining visibility on queries nobody buys from.

What a good monthly call sounds like

You will have roughly twelve of these calls a year and they are where the engagement is actually steered. Most of them are wasted on a recital of completed tasks, which is comfortable for everyone and useless.

The difference between the two versions is not tone or seniority. It's what fraction of the hour is spent on the past.

How much of your time this actually costs

Nobody quotes this in the proposal, so here it is: budget four to six hours a month of your own time, and slightly more in month one.

That's not the agency being needy. Everything that makes your content different from your competitors' — the objection your buyers raise, the reason your price is what it is, the customer who churned and why — exists only in your head or your sales team's. An agency can research your market. It cannot research your knowledge.

The engagements that go wrong on the client side almost always go wrong here. Not through malice — through a founder who's genuinely busy, so approvals sit for three weeks, the subject-matter call gets rescheduled twice, and the agency ships safe generic content because the alternative was shipping nothing.

• 60–90 minutes — the monthly call. Non-negotiable, camera on, both sides prepared.
• 2–3 hours — reviewing and approving content, and answering the awkward factual questions in the drafts. This is the part that pays.
• 30–60 minutes — unblocking your developer, or being the person who tells them a redirect map takes priority over the new landing page.
• Occasional — an intro to a customer for a case study, a quote for a PR pitch, a look at a competitor you know something about.

Months 4–6: the first leads, and a change in the conversation

Somewhere in this window the first organic lead arrives that clearly came from the work — usually from a page nobody expected, answering a question you thought was too basic to write about.

The texture of the engagement changes here. Month one to three you're asking "is this working?". From month four you start asking "why did *that* work?", which is a much more productive question and the one that shapes the second half of the year.

You'll also notice the reporting getting easier to read, because there's finally something to report. And you'll notice your own behaviour changing: you start noticing search results, sending the agency queries you typed yourself, forwarding a competitor's page you found annoying. That's the engagement working, incidentally — the client who pays attention gets better output, every time.

For a fuller version of this stretch, we've written up what actually happens in the first 90 days.

Month 9, when it works

The difference at month nine is not that the numbers are bigger. It's that the site behaves differently.

Pages you publish now get indexed in days rather than weeks, and reach a decent position in weeks rather than months, because the site has accumulated crawl attention and topical coverage. The pillar page from month two, which sat at position 30 for a quarter, is now on page one and pulling the cluster around it up. Queries you never targeted start showing up, because a site that covers a subject properly earns results for questions nobody wrote a page for.

The commercial version of that is straightforward: the leads are no longer occasional. They arrive weekly, they mention the article by name, and your sales team stops asking whether SEO is doing anything.

It doesn't always look like this. Some sectors are slower, a core update can flatten a quarter, and a brand-new domain competing with fifteen-year-old incumbents takes longer than nine months. But when it works, this is what it feels like — and it's why we measure against your own frozen baseline rather than promising a position. Positions are Google's to give. The baseline is yours, and it's the number your accountant cares about. That's the whole basis of how our SEO work is structured.

The three ways it goes wrong, and what each one feels like

Not every engagement works. These are the three failure modes we see most, and each has a distinct feel — which means you can catch it early if you know the smell.