Ad research across Facebook, Google, TikTok and YouTube

Good ads are rarely invented from nothing. They come from knowing what a category already says, how it says it, and where the gaps are. Ad research is the habit of looking at that systematically. The big platforms publish their ads in public libraries, so the raw material is there for anyone. What is missing is a way to collect it as rows you can sort and compare.

The four sources

Facebook and Instagram. Facebook Ads Scraper searches the public ad library by keyword and optional country, finds advertiser pages by name, and lists all ads of an advertiser. Each ad comes with its text, call to action, start and end dates, an active flag and its media. For video ads, an Ad transcript mode returns the spoken words. For many consumer categories, Facebook ad research is the natural place to start.

Google. works with the Ads Transparency Center. It finds advertisers by keyword, collects their public ads, and returns the details of a single ad, including its format, regions and landing page.

TikTok. can search the TikTok ad library by keyword and look up an ad by its id. It sits in the same tool as TikTok hashtags, comments and shop data, so one account covers both the ads and the organic videos around them.

YouTube. YouTube Ads Scraper takes video links and returns the likely sponsor of each video, with the description lines it based that on and a confidence level. It answers a different question: which brands pay creators in your category for a mention.

What to compare

Collecting ads is the easy part. The value comes from comparing them on a few dimensions:

• Hook. The first line of the text, or the first sentence of a video transcript. List them side by side and you will see the two or three angles the whole category uses.
• Offer. Discount, free trial, free shipping, bundle. Count how many ads lead with each.
• Call to action. "Shop now" versus "Learn more" shows whether the advertiser wants the sale now or later.
• Run time. End date minus start date. Sort by it. The long-running ads are the ones worth studying.
• Format. Image or video, from the media kind. Some categories have moved almost entirely to video.
• Landing page. Where the click goes: the home page, a product page, or a dedicated campaign page.

A one-afternoon project

1. Search the Facebook ad library for three category keywords in your main country. Export the ads.
2. Search Google's transparency data for the same keywords. Export the advertisers and their ads.
3. Pick the five advertisers that show up most. List all their live Facebook ads with the Companies and Company ads modes.
4. Run Ad transcript on their longest-running video ads.
5. Build one sheet: advertiser, platform, hook, offer, call to action, run time, format.
6. Write down three angles nobody in the category uses. Those are your test ideas.

Repeat it

Ad libraries change every week. Save each search as an Apify task and attach them to a weekly schedule. Combining the exports by ad id shows you, over time, which new ads an advertiser launches, which they stop after a few days, and which keep running for months. That is as close as public data gets to knowing what works.

Notes

The tools read public ad libraries and public video pages only. Sponsor detection on YouTube is an inference from what the creator wrote, so treat it as a strong hint rather than a fact. Spend, reach and targeting are not part of the output. Use what you learn to make your own ads better, not to copy someone's creative. Everything runs on Apify, payment goes through your Apify account, and each Store page lists the current price per 1,000 results.