SEO & GEO

SEO analysis, monitoring and optimization

SEO and GEO are ongoing processes, not one-off projects. We monitor what is working, explain what changed when results move, and turn that into the next set of decisions.

Reporting exists to change decisions

A monthly report that gets read and filed has failed, however well designed it is. The only reason to measure is to decide something differently next month, so we build reporting backwards from the decisions you actually make: what to publish, what to fix, where to stop spending.

That is why the first question in any engagement is not which metrics to display. It is which choices are open to you and what evidence would change them.

Fix the measurement before interpreting it

Almost every setup we inherit has something quietly wrong with it. Two analytics tags firing on the same page and doubling sessions. Internal and agency traffic never filtered. A consent banner configured so that a large share of visits are never recorded, with nobody accounting for the gap. Conversions counted on page load rather than on submission.

None of this is exotic. All of it produces reports that look authoritative and describe something other than reality, which is worse than having no reporting at all, because people act on it. So the first phase is auditing the collection layer and correcting it. Only then does the historical data become worth arguing about.

The distinctions that make a report honest

A single organic traffic line hides most of what you need to know.

Branded versus non-branded. People searching your company name were already coming. Growth there usually reflects a campaign, a mention or word of mouth. Non-branded queries are the part SEO is genuinely responsible for, and reporting that merges them flatters everyone involved.

New pages versus existing ones. A site can publish steadily, show flat totals, and be losing older pages at the same rate it gains new ones. That is a very different situation from a plateau, and it needs a different response.

Impressions versus clicks. Rising impressions with flat clicks means you are appearing and being passed over, which points at titles, snippets and intent mismatch rather than at rankings.

Assistant referrals. Traffic from ChatGPT, Perplexity and similar sources is small for most sites today, but it is measurable and its direction is informative.

Explaining movement requires a change log

The most valuable artefact in ongoing SEO work is unglamorous: a dated list of what changed. Deployments, redirects, content published, structural edits, algorithm updates, campaigns launched, seasonality.

With it, a traffic drop is usually traceable within an hour. Without it, everyone involved constructs a plausible story, and plausible stories in this field are cheap. We keep the log alongside the dashboard and annotate the timeline, so the conversation starts from evidence rather than from whoever speaks most confidently.

It also protects you in the other direction. When results improve, the log is what distinguishes work that caused it from work that happened to coincide with it.

What the data cannot tell you

There is no control group in SEO. You cannot run the same quarter twice, once with the changes and once without, so causal claims are always partly inference. Anyone presenting them as proof is overstating what the method supports.

There are narrower limits too, worth knowing before they surprise you. Search Console holds sixteen months of data, so year-on-year comparisons need exporting before the window closes. Rare queries are omitted from Search Console reports for privacy reasons, which is why query clicks never sum to the total. AI Overview impressions are not broken out as their own category. Assistant visibility can only be sampled, not tracked.

We say all of this in the reporting rather than in a footnote, and we label which conclusions are well supported and which are our best reading. It makes for less impressive slides. It also means that when we tell you something is working, you have grounds to believe it.

What you get

A measurement setup that can be trusted

Analytics, Search Console and consent handling checked and corrected first. Reporting on a broken setup produces confident conclusions from data that was never right.

A dashboard you can read in two minutes

Traffic, visibility, conversions and AI referrals in one place, split into branded and non-branded search, so a rise in people typing your company name does not get counted as SEO progress.

Query and page level analysis

We look at which pages and which queries moved, not just at a total. An aggregate that stays flat regularly hides one section collapsing and another growing.

A change log

Deployments, content publications, algorithm updates and campaign starts recorded on the same timeline as the metrics. Without it, every explanation of a drop is a guess.

Monthly commentary with recommendations

A short written interpretation of what happened and what we propose next, prioritised. Numbers without a reading are just a chart nobody acts on.

Alerting on the things that hurt

Indexing drops, crawl errors, tracking failures and sudden traffic loss surfaced quickly, because these are cheap to fix in the first week and expensive in the third month.

How we work

  1. 01

    Audit the data before trusting it

    Duplicate tags, missing consent categories, filtered internal traffic, conversions that fire twice. We fix the collection layer first, since every later conclusion depends on it.

  2. 02

    Agree what actually counts as success

    Usually inquiries, sales or qualified sessions rather than raw traffic. We define the metrics with you up front so the reporting cannot be quietly redefined when a number goes the wrong way.

  3. 03

    Build the reporting

    Search Console, analytics, rank data and AI referral traffic pulled into one dashboard, with the definitions written down so the numbers mean the same thing in six months.

  4. 04

    Review on a fixed rhythm

    A monthly call over the same view, with the change log open. The output is a short list of decisions, not a walkthrough of every chart in the document.

  5. 05

    Act on the findings

    Recommendations feed straight back into technical, content and GEO work. Monitoring that never changes what anyone does is a cost, not an investment.

Tools and technology

Where a solid open-source tool exists, we choose it over a closed one. No lock-in to a single vendor, and costs you can actually predict.

  • Matomo
  • Umami
  • PostHog
  • Grafana
  • Metabase
  • Lighthouse
  • Playwright
  • PostgreSQL
  • Google Search Console
  • Google Analytics 4
  • Looker Studio
  • Ahrefs

Frequently asked questions

More services in this category

Read testimonials from companies that trusted us

They're always a few steps ahead.

Mikołaj

CEO & Founder, GBS®

View on Clutch
GBS® logo

Delivered well ahead of the deadline.

Yasniel

CEO, IMEGA Sp z o.o.

View on Clutch

ZanReal's individual approach is impressive.

Adam

Executive, w-studio.pl

View on Clutch

Knowledge and business intuition make them a valuable partner.

Magda

Designer, DIGITALUNI

View on Clutch

Quick solutions that reduced costs by 99%.

Andrei Kapytau

Team Lead, busel.uk

View on Clutch

+20% deliverability for our email campaigns.

Joan Calabria

Sales Director, 36NORTH

View on Clutch
36NORTH logo

Reports should end in decisions, not charts

Message us

Show us how you report SEO today, and we will point out what the setup misses, from broken tracking to branded traffic counted as progress.

Zanek

Can't keep up with changes in AI world?

Let us do the heavy lifting. Every week we distill the most important AI developments into a focused 5-minute briefing — so you stay ahead without the noise.

Find out more
Weekly AIonline