Website Change Alert vs Evidence: What Holds Up in a Dispute

The short answer

A change-detection tool answers one question: did this page change? A dispute asks different ones: what exactly was on the page on a given day, since when, and how can anyone be sure the record was not altered afterwards? An alert is excellent at the first job and was never built for the second.

That does not make change detection useless. Keep it, because it tells you quickly that something moved. But if a page could end up in front of a judge, a regulator or a platform's appeals desk, you need a record built as evidence: dated, complete, and checkable by someone who does not have to trust you or your vendor.

What a change alert gives you, and where it stops

A typical change-monitoring tool fetches a page on a schedule, compares it with the previous version and sends a notification with the difference highlighted. That is useful, and it often runs more frequently than an evidence tool. The limits appear when you try to use the output later:

The date comes from the tool's own records or from an e-mail. Nothing outside the vendor's system fixes it.
What is kept is often a snapshot or a text difference, not the page's HTML, network traffic and certificates together.
Many tools store a version only when something changed, so the days between alerts are a gap in the record.
Access usually depends on the account, so the history can end with the subscription.

Check what your own tool does, because products differ. The point is not that alerts are bad. It is that none of this was designed to answer an objection from the other side.

The four questions a dispute asks

Whatever the forum, the facts tend to come down to the same four questions.

What: the full content of the page, not only the words that changed.
When: a date and time that does not rest on the word of one party.
How long: whether the content was there throughout a period, or only on the days someone looked.
How it was obtained: which method and tool were used, what was done to the page during capture, and whether the record changed later.

The days when nothing changed are evidence too

Alerts stay silent precisely when a page stays the same. In many disputes, that is the very fact you have to prove.

A clone shop that kept trading after you reported it, where the length of the operation matters for damages or for what the host knew after a notice under Article 16 of the Digital Services Act.
A reference price that has to be the lowest price of the prior 30 days under Article 6a of Directive 98/6/EC.
A promotion, claim or set of terms that the other side says was corrected in March and you say ran until June.

A series in which every check is sealed answers this directly: each scheduled check is its own dated record, whether or not anything moved. A record of changes alone leaves you inferring the middle. If a vendor seals captures only when a change is detected, that is the gap to ask about.

What recent decisions actually show

Courts do not insist on one particular tool, and it would be wrong to say they do. These four decisions show what does go wrong.

General Court, T-580/25 Doors Bulgaria (29 April 2026): a Wayback Machine screenshot can serve as evidence of disclosure. The abstract possibility of manipulation was not enough to reject it, and the design regulations prescribe no mandatory form or certification for such evidence.
UK IPO, O/0721/26 Savant (11 August 2026): Wayback printouts from 2017 to 2023 did not help for a use period of 2003 to 2008, because nothing showed what the website looked like at that time.
UK IPO, O/0748/26 (18 August 2026): website screenshots were either undated or dated outside the relevant periods, so they could not show use within them.
Tribunal judiciaire de Paris, 23 January 2025, n° 22/03006, as reported by the French APP: the court was not assured of a bailiff's web report because it could not follow the path through the pages visited, and the screenshots carried no date or time.

Read together: an archive or a screenshot can be perfectly acceptable, and formal certification is not a general requirement. What fails is a record with no reliable date, one that covers the wrong period, or one whose method nobody can retrace. And no tool can add a check to a day that has already passed.

What a sealed series consists of

In GetProofAnchor Website Monitoring, each scheduled check of a public page produces a complete proof, the same as a manual capture:

Full-page screenshot, raw HTML, text, MHTML snapshot, video, HAR with cookie values and authorisation headers masked, TLS certificate chain, DNS from three public resolvers and RDAP/WHOIS.
A SHA-256 hash of every artifact, all bound into a qualified eIDAS timestamp from an EU-listed qualified trust service provider.
Sealed capture metadata recording the method, including consent handling and any visual interventions in the screenshot.
Offline verification with the open-source gpa-verify, so any change to a sealed artifact after the timestamp can be detected without relying on us.

Under Article 41(2) eIDAS a qualified timestamp enjoys the presumption of the accuracy of its date and time and of the integrity of the data it is bound to, and Article 41(3) makes it recognised in all member states. It does not prove that what the page said was true, and it says nothing about what another visitor saw on another device. The weight of the record remains a matter for the court; the Council of Europe guidelines ask courts to be aware of the value of trust services when assessing reliability.

Side by side

Purpose: an alert tells you to look; a sealed series is what you show to someone else.
Frequency: alert tools can run every few minutes; our checks run daily or weekly, so this is not a real-time service.
Unchanged days: an alert says nothing; a sealed check still records the day.
Integrity: an alert rests on the vendor's records; a sealed check rests on a qualified timestamp and hashes anyone can recheck.
Cost: alert tools are generally cheaper and suit a long casual watch list; sealed checks use one proof each and are for the pages that could matter.

How to run both without duplicating work

Most teams should not choose. Use alerts to notice and sealed checks to prove.

Keep your alert tool on the broad list of pages you merely want to watch.
Move every page that could reach a dispute, a regulator or a takedown onto a sealed schedule, before the dispute starts.
When an alert fires on an important page, make a manual proof straight away, because the next scheduled check may come too late.
Group the proofs by client or matter in a case folder, and export the whole case as one ZIP with an index CSV when the lawyer asks.

Choosing a schedule and a plan

A daily check uses about 30 proofs a month per page, a weekly check about 4 to 5. Daily suits volatile pages and campaigns; weekly suits terms, claims and slower pages.

Business, 99 EUR per month: 10 monitors, 400 proofs per 30 days, 10 users.
Monitoring 50, 399 EUR per month: 50 monitors, 1,800 proofs per 30 days.
Monitoring 100, 599 EUR per month: 100 monitors, 3,500 proofs per 30 days.
All prices exclude VAT; checks and manual proofs share the quota, and yearly billing is available.

Monitors cover public pages. Pages behind a login go through the browser extension, captured on your device and sealed on receipt, or through Forensic Browser from 290 EUR per year excl. VAT, which also handles multi-page journeys and video. If a page must be secured within the hour, Assisted Capture Rush costs 249 EUR excl. VAT with a 60-minute guarantee. Server captures are public by default and can be unpublished in the dashboard. Nothing is recorded between two checks, so set the interval to match the risk.

Six questions to ask any monitoring vendor

Is every check sealed, or only the ones where a change is detected?
Does the timestamp come from a qualified provider on an EU trusted list?
Can I verify a record offline, without your account or your help?
Which artifacts are kept for each check, beyond a screenshot?
Is the capture method, including anything done to the page, recorded and sealed?
What happens to my records if I cancel?

FAQ

Can a change alert be used as evidence?

It can be submitted in many procedures, and its electronic form alone is no reason to refuse it. Its weight depends on the forum and on what it can show about date, content and integrity. We do not promise how any court will weigh any record, ours included.

Do we have to replace our current change monitoring tool?

No. Keep it for fast notice on a broad list of pages, and put the pages that could reach a dispute on a sealed schedule.

Is the Wayback Machine not enough?

Sometimes it is, as T-580/25 shows. But it holds only what a crawler or a visitor happened to save, so it may have nothing for the dates you need.

Does scheduled monitoring catch every change?

No. Checks run daily or weekly and nothing is recorded between two runs. For a known event, such as a launch or a price change, add a manual capture at that moment.

Can we see a sample before buying?

Yes. Send one public URL to sales@getproofanchor.com and we capture it as a free sample, so you can inspect and verify the evidence package before choosing a plan.

Sources

For compliance and audit teams

Prove what was published — and when it changed

Scheduled forensic captures of policies, prices, disclosures and competitor claims. Every run is a sealed proof with an eIDAS qualified timestamp, building an unbroken evidence series.

Solutions for compliance Set up website monitoring

See plans and pricing →

Related guides

Put the pages that matter on a sealed schedule

Website Monitoring checks your pages daily or weekly and seals every check with a qualified eIDAS timestamp, including the days when nothing changed. You still get an e-mail when a page changes.

Create a proof Verify a proof

Business from 99 EUR/month, Monitoring 50 399 EUR, Monitoring 100 599 EUR, Forensic Browser from 290 EUR/year, Assisted Capture from 49 EUR, all excl. VAT. Not legal advice; no guarantee of how a court or regulator weighs the evidence.

A change alert is not proof: website monitoring versus a sealed evidence series