In the brief
Practically free. One sentence gets reworded and no code has been written yet.
Quality is not something a meeting can declare; it has to be measured. Give us your application, your shop or your portal and we travel through them the way your users will, then return an ordered list of findings, each one reproducible and costed in repair effort. Five services cover the five questions worth asking before a launch. Everything happens remotely, on your own test environment.
Does the product do what the brief says? Will it hold up on the busiest day? Can it be checked without anybody clicking through the same screens? Do people follow what they are looking at? Could a stranger reach data they have no right to? Each of the five takes one of them.
The journeys that matter get travelled end to end: registration, basket, promotional codes, card payment, picking a collection point, returns and refunds. Odd input receives a pass of its own, since the costly defects like to shelter there.
Hundreds and then thousands of simultaneous visitors get simulated, pinning down the point where pages slow and the point where they start returning errors. Far better to meet that limit in October than on the biggest day of your year.
Your essential scenarios run on their own inside the pipeline whenever something is deployed. The same screens no longer need clicking through every fortnight, and shipping on a Friday stops feeling like a wager.
Genuine users carry out tasks on their own equipment with the screen shared. Within minutes it is clear which wording misleads, which buttons nobody notices, and at what moment the tab gets closed for good.
Run purely on a written authorisation from whoever owns the system, inside the scope and the hours named in the agreement. Authentication, permissions, incoming data and server configuration all come under the probe.
A decimal point in the wrong place inside a discount rule stays harmless until a customer meets it. Repair grows dearer with every stage it survives, which is why steady checking while work is under way beats a single grand review the night before sign-off.
Practically free. One sentence gets reworded and no code has been written yet.
Inexpensive. Whoever wrote it still holds the logic in their head and reworks it in minutes.
Distinctly dearer. The ticket travels backwards, and code review plus a staging deployment happen a second time.
Expensive. An unplanned repair, generally after hours, trailed by an anxious look at everything nearby.
Dearest of all. Alongside the repair sit abandoned baskets, public reviews and a support inbox buried for two days.
A script only earns its keep by being replayed. Automation therefore goes to settled journeys, the ones obliged to work in every release. Screens still redrawn each iteration remain in manual hands; otherwise your budget disappears into script upkeep rather than into the product.
No overall grade along the lines of “quality is fair” comes your way. What arrives is a numbered, ranked set of findings with the steps for reproducing each one. All remote: over secure access to your test environment, or through test accounts you open for us.
A video call covering who depends on the system, what it talks to, and which failure hits revenue or the service you owe people. Attention follows importance, because not every screen deserves the same amount.
Scenarios, browsers and devices picked from your audience figures. You approve the plan, the hour count and the price before any work starts.
Everything lands in the tracker you keep already, whether GitHub, GitLab, Azure DevOps or Jira, carrying steps, data, the result expected, a screenshot and a severity.
With your repairs in place we confirm the fixed items and poke at the features beside them, since a repair often shifts the problem one step along.
A developer confirms the code delivers what was intended. A tester goes looking for the moments it delivers something else: a form submitted empty, the back button pressed mid-payment, two tabs open on one basket. Those are opposite frames of mind, and hardly anybody holds both towards work of their own making.
Yes, and the request comes up often before an acceptance certificate is signed, or before a local authority formally receives software ordered through a public contract. None of that code is ours, so nothing needs protecting. The moment of greatest value is before acceptance, while corrections can still be settled with whoever built it.
It is. Where a legal obligation applies, to local authority sites, schools and similar bodies, the RGAA 4.1 framework governs the review: keyboard navigation, contrast, text alternatives, the way a screen reader announces a page. Elsewhere WCAG 2.1 AA does the job. Findings feed your accessibility statement and your multi-year plan, and since an online shop may also fall under the European Accessibility Act, that point deserves a word with your legal adviser.
One-off engagements go by time spent, EUR 95 an hour excluding VAT, under a ceiling approved in the plan. Recurring checks, per release or per iteration, belong inside a monthly Start, Business or Premium plan, which keeps the figure predictable. Invoices leave in electronic form, as the French e-invoicing rules require.
Best of all is for none to sit there: fabricated records or an anonymised database copy do the job. Where real records cannot be avoided, a GDPR processing agreement gets signed beforehand, access is trimmed to the bare minimum, logged throughout, and withdrawn when the engagement closes.
Tell us what needs checking and when the launch is due. You get a test plan and an estimate built on the number of scenarios.
Message received
An answer follows inside one working day. Report an outage that is stopping people working and it moves ahead of everything else.
Nothing here under that name. Check the spelling, or simply choose the nearest large city instead. Since every engagement runs remotely, whichever you pick changes nothing about what we do for you.