Removing Sensitive Data

Our feature guides are written by walking through the real site with a real account, which is what makes them trustworthy — every screenshot shows what you would actually see. It also means the raw material picks up things that should never be published: customer names, addresses, account numbers, email addresses, phone numbers and exact pricing.

We used to take those screenshots as they came and cover the sensitive parts afterwards. That worked, but it left the guides studded with black boxes — exactly where the reader most wanted to look. We've changed the approach: the fictional data now goes in before the picture is taken.

What it does

• Swaps in stand-ins, not boxes. Immediately before each screenshot, real names, addresses, emails, phone numbers and prices are replaced on the page with fictional ones, in the real value's own font, size and alignment. The screenshot is clean at the moment it is taken, and the page still looks like a working page.
• Keeps the page coherent. The same real value always becomes the same stand-in, and a price keeps its rough size. A subtotal doesn't change between two shots of the same journey, and a price table still reads like a price table rather than a jumble.
• Never tells the application. The substitution is display-only. The site itself never receives the fictional value, so nothing we do for a screenshot can affect an order, a quote or an account.
• Runs a final safety scan. Every finished page and screenshot is still checked before publishing, images included — they're read for text, so a name sitting inside a table or behind a dialog is caught just as one typed into a sentence. Anything flagged is reported loudly and the step is re-captured.

A note about the prices you'll see

Every price in a guide screenshot is fictional, and there's a deliberate tell: they all end in 55 cents. $10.55, $98.55, $1,204.55.

That's not a coincidence, and it's there for you as much as for us. It means anyone can confirm at a glance that a screenshot has been through the process — without taking our word for it, and without running anything. It also means a figure that doesn't end in .55 stands out immediately as something to check.

The practical consequence: the numbers in the screenshots won't add up. A line total won't be the unit price times the quantity, and a subtotal won't be the sum of the lines above it. That's expected. Follow the steps rather than the arithmetic — on the live site, with your own pricing, everything reconciles.

One exception you may notice: on cross-reference results, the savings figure is shown as it really is. It's a difference between two products, not anyone's price, so it tells you something useful and reveals nothing.

What stays visible

This is targeted, not blanket. Quote and order numbers, part and catalog numbers, warehouse codes, dates, city and state, and job titles are all left alone — they're what make the guides useful to follow, and none of them identify a customer. Hiding more than we need to would be its own kind of failure.

Why it matters

It lets us keep building guides against the real, live site — the only way to be sure a journey actually works — while the published version stays safe to share with a wider audience. And because the fictional data is in place before the shutter, there's no window in which a sensitive screenshot exists waiting to be cleaned up.

A few older guides are still being rebuilt and their screenshots show an earlier style of stand-in data. They're safe to read; they just won't show the .55 pattern until their turn comes round.

If you spot something in a guide that looks like it shouldn't be there, let the documentation team know and we'll re-capture that step and republish.