What does an SEO agency actually do each month?

This question is almost always asked with a specific worry underneath it: am I paying for nothing?

Sometimes the answer is yes. But SEO also genuinely looks like nothing from the outside for the first several weeks even when it is being done properly, which makes the worry hard to resolve without a reference point.

So here is the reference point — what the hours actually go on, and how the mix should change over time.

Where the hours go

A local SEO engagement divides into five kinds of work. The proportions shift month to month; the categories do not.

Profile and review work

Google Business Profile audit and correction, category selection, service area, hours, services, photos, posts, duplicate listing cleanup, review requests, and responding to reviews.

For most Henderson service businesses this is the highest-return work in the engagement, because the map pack is where the phone calls come from. It is also the cheapest, which is why an agency that skips straight to blog posts should be asked why.

Content

Writing and publishing pages, rewriting existing ones, consolidating thin pages, improving titles and meta descriptions, internal linking.

Note that content is not automatically blog posts. For a local service business, a genuinely useful service page usually earns more than four articles.

Technical and on-page

Site speed, mobile rendering, crawl errors, broken links, schema markup, indexation problems, redirects after any URL change.

This is front-loaded. A lot of it is fixed once and then monitored.

Monitoring and measurement

Position tracking across a grid of locations, Search Console review, tracked call and form review, competitor movement.

Small in hours, large in consequence, because it is what makes the next month's decisions something other than guesswork.

Reporting and communication

Assembling the report, the work log, and actually talking to you.

This should be a modest fraction of the total. If reporting is consuming a large share of the hours, you are paying for the description of the work rather than the work.

What a real ninety days looks like

Month	Where most of the hours go	What you should be able to see
1	Baseline measurement, profile audit and correction, technical fixes, duplicate cleanup	A dated work log and a documented starting position — not results
2	First content, citation consistency, review process started	Early profile movement if the profile was broken; work log continuing
3	Content continuing, profile upkeep, first real measurement comparison	Map position beginning to respond, first attributable leads
4–6	Content and reviews compounding, adjustments based on the data	Lead volume trending; a fair point to judge the engagement

Month one contains work, not results. That is not an excuse, it is the shape of the thing. What month one should produce is evidence — a task-level log and a baseline you can compare against later. An agency that cannot produce those in month one is the problem; an agency that has not produced rankings in month one is not.

How to tell real work from a padded log

The whole advantage of a task-level work log is that it is checkable. Use it.

Named things are verifiable. "Rewrote the water heater page title and meta description on 14 March" names a page and a change. Open the page. Look at the title. That takes five minutes and settles the question.

Categories are not. "Content optimization," "technical improvements," "ongoing SEO" — none of these can be confirmed or refuted, which is precisely why they appear.

Some repetition is legitimate. Review management, monitoring and profile upkeep genuinely recur every month and should appear every month.

Some repetition is not. If the same three pages are being optimised in month two, month five and month eight, either the work is not finishing or the log is being inflated. Check whether the page actually changed between those dates.

The signal that matters most: does the plan respond to the data?

This is the thing that separates a competent engagement from an expensive subscription.

Measurement exists to inform decisions. If position and leads have not moved in ninety days, something should visibly change — different pages prioritised, a different diagnosis, a different approach to the areas where you are losing.

An agency running the same plan in month six that it ran in month one is not responding to evidence, regardless of how polished the reports are. That is true even if the plan was a good one, because a good plan that is not working is still not working.

What to ask, if you are worried right now

Three questions, one email:

1. What did you do last week, specifically — pages and changes?
2. How many tracked leads did I get last month, and from where?
3. What are you doing differently next month, and why?

The third question is the one that reveals most. An agency thinking about your account has an answer ready, and it references something they observed. An agency running a template has to invent one.

If the answers do not come, or come as a meeting request, that is covered in how do I know if my SEO is working, and what a monthly report should include sets out what you should have been receiving all along.

The reason we publish a dated task log with every report is that it is the only part of this industry a client can independently check. That is the whole idea behind The One Number Method.