That is the whole answer, and I would have said it was a small point until it bit us.
Almost everything we collect can be fetched again tomorrow. This cannot.
Meta's image and video URLs expire. A campaign landing page gets deleted the week the campaign ends. A price gets quietly corrected and nobody writes down what it used to say. The evidence of the most interesting thing your competitor did this quarter has a shelf life measured in weeks.
So we do not measure the landing page. We record it. Raw HTML, kept, dated.
The value of an archive comes entirely from when you started it, which is an unsatisfying thing to sell and a very satisfying thing to own eighteen months later when somebody in a meeting says "did they always position it that way" and you can answer with a file instead of a feeling.
Then there is the analysis I did not expect to be so revealing.
We already know where each ad sends people. We had never asked whether the page says the same thing as the ad that sent them there.
An ad promises one thing, the page leads with another, and the gap between them is money leaving the building.
It happens constantly, and it happens for a boring organisational reason: the person who wrote the ad and the person who owns the page are different people, and they last spoke in a kickoff meeting. The ad gets iterated weekly by a media buyer. The page gets touched twice a year by whoever has access.
You can see it in the archive. Ad says "sleep". Page leads with "concerts". Both are true about the product. Only one of them is what the person who just clicked was thinking about.
Which is worth knowing about a competitor. It is worth considerably more when you point it at yourself, and I would do that first, because there is a decent chance you are the one with the mismatch.
What this is not: a screenshot tool. We keep the markup and a full snapshot, so you can read what it said, not admire what it looked like. If you need pixel-perfect visual history, that is a different product.
One thing to do next: click your own best-performing ad, as a stranger would.
Read the first line of the page it lands on. Ask whether it answers the sentence you just clicked.
Pick one competitor. See what we find.
Your first benchmark is free. One domain, measured against you, no card and no call.