Evidence of what was published, and when it was published

Compliance teams rarely need a screenshot in the moment. They need one eleven months later, when a customer disputes what a fee page said, a regulator asks about a promotion that ran last spring, or a partner claims your terms permitted something they did not. By then the page has been edited, the person who wrote it has left, and the only honest answer is that nobody knows.

Nothing here is legal advice, and what satisfies your obligations is a question for your own advisers. What follows is what tends to separate a record that helps from a folder of images that does not.

Systematic beats comprehensive

The instinct after a scare is to archive everything. That produces enormous volume and no confidence, because nobody can say what the collection policy was. A smaller set of pages captured on a stated cadence, continuously, is far stronger: the policy is visible in the record itself.

Decide the list deliberately, write down why each page is on it, and let the schedule run. When someone asks why a page was captured daily, the answer should be a policy, not a hunch.

Timestamps are only useful if they are unambiguous

"14 March" means different things either side of a dateline, and a daily job that drifts by an hour twice a year raises questions nobody wants to spend time on. Schedules run in a stated timezone, so a 09:00 capture is 09:00 where the business is, through daylight-saving changes. State the timezone in your records policy and the ambiguity disappears.

A time written on a capture is how a person reads the exhibit; it is not what proves the file is unchanged. That is the job of the SHA-256 recorded as each capture is stored, and of the checksum file that travels with a month's download. Where timing itself may be contested, Enterprise adds an RFC 3161 token from an independent Time-Stamp Authority to each kept capture, so the date does not rest on our word.

Capture what the customer saw, not what you meant to publish

A disclosure that exists in the page source but sits behind a consent overlay, below six screens of marketing, or off the bottom of a phone layout, is not the same as a disclosure a customer read. This is precisely where disputes concentrate.

• Capture mobile separately. A second schedule with a real device profile records the layout most of your customers actually got.
• Think carefully about banner removal. Stripping cookie banners produces a cleaner image of the page — but if the question is what was obscured, the banner is the evidence. Consider capturing both ways.
• Record by market. Where obligations differ by country, capture from each location rather than assuming one rendering served everyone.

Where the files live is part of the control

Evidence held in a product you might stop paying for is fragile. Deliver captures into your own Google Drive, Dropbox or S3-compatible bucket and they inherit the retention schedule, access controls, audit logging and backups your organisation already applies to records. Our console then becomes the convenient way to browse and compare them, not the place your evidence depends on.

Use tags and a consistent naming structure from the start. Producing "every capture of the fee schedule between March and September" should be a folder listing, not a project.

The adjacent uses

Once the mechanism is running, teams almost always extend it. The same schedules become a general web archive of pages the business depends on, and marketing discovers the same captures answer competitor and brand questions. Set it up for the obligation; the other uses follow at no extra effort.