What Google's, Meta's and TikTok's public ad archives actually contain — measured from live pulls, including the fields people assume are there and are not. Plus 22 ready-made extractions built on top of them.
Ad copy, landing URLs and CTA on every ad — and no spend or reach figure anywhere. Fill rates from 100 live ads, plus why searching a brand name returns its resellers.
The EU DSA archive gives dates and audience bands; Creative Center gives CTR and video length. Almost nothing overlaps. Which fields we would not build on, and why.
Google stores search ads as images, so there is no headline field to query. What that means, and what comes back reliably instead.
We build and maintain three scrapers for these archives, so we end up measuring their output more often than most people need to. The fill rates and gotchas above are a by-product; publishing them saves someone else the pull.
Every number here comes from a run we executed, on the date shown. Where a field is unreliable we say so rather than shipping it as if it were solid.
If you're tracking competitor ads across a client roster and re-pulling the same accounts by hand, we'll run a two-week watch for a handful of advertisers at no cost — you get the daily diffs, we get to see if the fields hold up under your workload. No account or call required. Email us the advertisers and we'll set it up.