SEO services: technical, content, authority, measurement
SEO services fall into four workstreams: technical, content, authority and measurement. What each one delivers, in what order, and how to verify it.
By Roozbeh Nazari · CEO
The work sold under the heading of SEO services splits into four workstreams that demand quite different skills: technical foundations, content, authority and measurement. One proposal may collapse all four into a single line; another may contain only one of them and still carry the same name. This article explains which task belongs in which box when you buy an SEO service, what you actually hold at the end of each one, and the order that makes sense. We covered how to read package tables line by line in the SEO packages article; the focus here is the anatomy of the work itself.
The practical value of the distinction is this: a site's problem is almost never "SEO". The problem is either that pages cannot be crawled and indexed, or that no page exists to match a real search intent, or that you are not treated as a credible source on a topic, or that nobody is measuring the result of the work being done. All four want the same budget, but they do not have to be solved at the same time.
Technical SEO: making the page visible and processable
The technical layer's job is to make sure the page you prepared can be found, crawled, indexed, and that the right version of it is selected. Its scope usually covers site architecture and internal linking, redirect and parameter rules that stop crawl budget being wasted, canonical tags, hreflang on multilingual sites, sitemap generation, structured data and page performance. Most of what we run on the technical SEO side comes out of that same list.
What distinguishes technical work is that its output is verifiable. You can see that an indexing problem was fixed through URL inspection and the page reports in Search Console; you know an hreflang error is resolved when the tag pairs become reciprocal. So on a technical line item, "done" is not enough — ask for the verification screen. If you are wondering where to start, the technical SEO audit checklist sets out the order.
The most time-consuming part of technical work is separating out which stage the problem sits at. A page may never have been discovered, or discovered but not crawled, or crawled but not indexed, or indexed and then dropped in favour of another version. All four produce the same symptom: the page is not in search. But the fixes are completely different, and a wrong diagnosis means weeks spent in the wrong place. On JavaScript-rendered pages this distinction matters even more, because what appears in the browser and what gets processed can diverge.
Content SEO: the page that matches search intent
The content layer connects the queries people actually type to the pages on your site. The visible part of the job is producing text; the real part is deciding which page serves which intent, merging or splitting existing pages, and stopping you from competing with yourself by covering the same topic twice. The output of content SEO work is not a list of articles but a page map: each row carries the target intent, the page that will serve it, that page's status and who owns updates.
Google's own starter guide holds the same line: the thing that most affects search visibility is content that is readable, original and written for people. In practice the gap we see most often is that existing pages are never revisited while new content is produced. On most sites the highest-yield work of the first three months is not adding new articles but rewriting pages that already rank and are underperforming.
The second content decision is page type. If a query seeks information, a guide page answers it; if it seeks comparison, a table and criteria page does; if it carries purchase intent, a service or product page does. Three separate intents on one topic need three separate pages, and the way they link to each other determines which one ranks. Content written without this map usually answers a single intent three times and goes deep on none of them.
Authority: links, brand mentions and reputation signals
The third layer is about whether you are treated as a source on a topic. Its classic component is external links, but that is not its only one. Brand mentions in the press and trade publications, organisation and author identity being visible on pages, the reviews written about you and how they are managed all belong to the same layer. On AI-assisted search surfaces the weight of these signals has increased: being cited as a source is not the same thing as ranking.
The most important thing to watch here is the shape of the promise that enters the contract. "This many links" is a numeric commitment and is easily filled with low-quality sources. The healthier approach is to write down which kinds of publications are targeted, which content will earn the link, and how the approval process works. Google's own advice to people hiring an SEO points the same way: stay away from proposals that guarantee rankings and will not explain their method.
Measurement: which decision rests on which data
The fourth layer is the one that tells you whether the other three worked, and it is the most frequently skipped. Its scope is the correctness of the analytics setup, conversion definitions, search console and analytics data complementing each other, language and market breakdowns being reportable, and a regular record of decisions. In projects started without measurement, the question "did it work?" has no answer in the third month; everyone looks at their own chart.
The method that works in practice is asking three questions in the same order at a monthly meeting: what changed last month, which metric moved on which pages as a result, and what decision is being taken for next month. The article on turning Search Console reports into decisions sets up the data side of that meeting.
On multilingual, multi-market sites the measurement layer has one more job: making the same report readable by language and market. A case where total traffic rises while one market declines is only visible in a segmented report. Conversion definitions also have to mean the same thing in every market; if a form submission is the primary touchpoint in one market while another runs through a messaging app, putting the two markets' numbers side by side misleads.
Which service to start with
On most sites the order goes like this: first remove whatever is stopping pages from being visible in the technical layer, set up measurement at the same time, then draw the content map; authority work is meaningful once there is something worth pointing at. The exception is sites whose technical side is already sound; there the first month goes straight to content and intent matching. To see which situation your own site is in, you can follow the order in the 30-minute site analysis.
One last check when you receive a proposal: for each of the four layers separately, ask "what is this month's output and how will it be verified?" A proposal that gives the same answer for all four is usually a proposal that works in one layer and covers the other three with a report. The reason we also separate our services under these four headings is so that you can see from the start which layer the scope goes deep in.