Services › Impact and evolution
What happens after launch is part of the project too.
Once a site is in use, data, searches, friction points, and needs appear that didn't exist before. We work from those signals to improve visibility, performance, and operations with increasingly informed decisions.
Why impact and evolution?
A site in use starts answering questions that didn't exist before.
Now it's possible to see which products people search for, which documents they download, which pages fall outside the paths they take, which queries they enter in site search, and which maintenance tasks take the most time.
Other problems only appear with real traffic: filters that respond slowly, legacy URLs, content Google keeps finding in a PDF, or internal processes that repeat every week.
That material makes it possible to work with concrete evidence instead of imagining how the platform should behave.
Going live opens a new source of information: real use.
Going live changes the kind of questions.
Improvements can be seen before and after.
Learn moreThe approach
Measurement is useful when it leads to a decision.
Ask a question
Before opening a dashboard, it helps to know what needs to be understood.
It might be a download, a search, a product page nobody finds, or a task that repeats too often.
Find the signal
Analytics, Search Console, site search, technical logs, performance, and operations each show a different part of the same site.
Prioritize
Not every finding needs to become a task.
Each finding is ranked by how much it weighs on the decision at hand.
Intervene and verify
The change is tested on the real product and measured again to see whether it had the expected effect.
Ways to work on a live site
SEO
We work on structure, pages, and migration so Google can find the content the company wants to make visible.
Learn about the serviceVisibility in AI search (GEO and AEO)
We review which of the company's own sources an assistant can interpret, how the information is structured, and what's missing so products and content have a better chance of being represented correctly.
Learn about the serviceMeasurement and web analytics
We define which events can answer useful questions and how to interpret them.
Learn about the serviceWeb Intelligence
We use signals from the site itself to spot patterns: what people search for, what they don't find, which content connects different paths, and where new opportunities appear.
Learn about the serviceWeb performance and accessibility
We measure before intervening and check afterward what changed.
Learn about the serviceAI process automation
We look for repetitive tasks that take hours of work and analyze which ones can be turned into repeatable processes.
Learn about the serviceOngoing maintenance and optimization
Updates, issues, and improvements go into a prioritized backlog.
Learn about the service
SEOWe work on structure, pages, and migration so Google can find the content the company wants to make visible.
Visibility in AI search (GEO and AEO)We review which of the company's own sources an assistant can interpret, how the information is structured, and what's missing so products and content have a better chance of being represented correctly.
Measurement and web analyticsWe define which events can answer useful questions and how to interpret them.
Web IntelligenceWe use signals from the site itself to spot patterns: what people search for, what they don't find, which content connects different paths, and where new opportunities appear.
Web performance and accessibilityWe measure before intervening and check afterward what changed.
AI process automationWe look for repetitive tasks that take hours of work and analyze which ones can be turned into repeatable processes.
Team trainingThe site also has users on the admin side.
Ongoing maintenance and optimizationUpdates, issues, and improvements go into a prioritized backlog.How it works
Real use starts to show what's worth improving.
Observe
We gather signals from the site, search engines, performance, and operations.
We also listen to the team that maintains it: many friction points show up there first.
Measure
We set a baseline before touching anything.
When there's a measurable condition, the change needs to be compared against it.
Prioritize
We separate what blocks, what matters, and what can wait.
A technical finding doesn't go into production automatically just because it exists.
Intervene
The improvement can involve content, search, SEO, performance, accessibility, automation, or structure.
The aim is to work with changes that can be controlled and verified.
Verify
We test the real experience, without relying only on the admin panel or an automated test.
Whatever changed gets measured again.
Document and continue
The intervention is documented and informs the next backlog.
Each cycle adds more information for making better decisions.
Ideas
What happens after going live deserves to be understood too.
What an AI doesn't read
Why a scanned PDF is invisible to an assistant, and what to give it instead.
What it takes for ChatGPT and Gemini to cite a medical device company
An AI assistant builds its answer from what it finds written about the company, on its own site and on other sites. Which part of that depends on content, and how to check whether the company shows up.
How to audit a medical device website while it keeps running
A medical device website can't be switched off to review it. Physicians, distributors, and printed labels keep sending people to it while it's being fixed.
Which website maintenance tasks to automate in a medical catalog
Before a maintenance task is automated, its actual steps need to be examined. The name in the spreadsheet may describe something else.
What to measure on the website of a company with inquiry-driven sales
A medical device website doesn't have a shopping cart: the sales process may involve a sales rep, a tender, or a visit to the operating room. What to track on the site to know whether it's doing its part.
Five views per user: how to interpret usage in a medical catalog
On a corporate page, repeat visits are unusual. In an implant catalog, they're a normal part of how people use it, and the views-per-user metric shows that.