Competitor ad analysis: a six-step workflow for agencies

Competitor ad research starts with the page, not the brand name. Search the brand in the Meta Ad Library and in an ad database, and pick the page that carries the volume. For Magic Spoon, the database listed five similar pages on 2 November 2026; the one with by far the most ads, Magic Spoon Cereal (page ID 1978160732305312), is the one to analyse. Write down the page ID, because names repeat and IDs do not.

In the Meta Ad Library, set the country (United States here) and the status to Active, and open the page. On 2 November 2026 it showed about 57 results for Magic Spoon Cereal. The ad database counted 53 active ads for the same page on the same day. The library's figure is marked approximate and the two sources count differently, so record both and name the source next to each number.

Open the first 12 cards Meta returns and note, for each, the start date, the format and the first line of text. This is Meta ad research by eye, and it is quick. Magic Spoon's 12: 9 started in October and 3 in September; 3 push a custom bundle; 2 announce a new flavor, both started 29 October; and every call-to-action the library showed was Shop now (10 of 12 cards showed one; the other two are grouped cards).

The card below is one of the bundle ads. It started on 11 September, it offers up to 26% off a custom bundle when you subscribe, and the library says 2 ads use this creative and text, a reuse signal you can read without any tool.

Days running is today minus the start date, and it is the nearest public proxy for an ad that works. The ad database returns it per live ad. Of Magic Spoon's 53 live ads, 27 (50.9%) had run 14 days or more and 6 (11.3%) had run 40 days or more; the oldest ran 52 days, from 11 September, the same start date as the card above (11 September to 2 November is 52 days). The second-oldest ran 49 days, from 14 September.

The free way is to do the sum by hand: every live card in the Meta Ad Library shows “Started running on”, so list the dates in a sheet and subtract. The limit is that the library shows live ads only, so you cannot see how long a stopped ad ran or which creative was cut.

Treat age as a shortlist, not a verdict. A long-running ad may simply be forgotten, so open the oldest few and ask what they share: the two oldest both sell a custom bundle with up to 26% off, as the cards show.

Competitor ad tracking is mostly pace. Count ads launched in a fixed window and how many are still live. Magic Spoon launched 90 ads (versions, not unique creatives) between 3 October and 1 November 2026, and 41 of them (45.6%) were still active on 2 November; those 41 make up 77.4% of its 53 live ads. So most of what the page runs today is under a month old. The advertiser card gives the scale: 874 ads in the page's history on 2 November 2026, all in the United States. “Ads” here means ad versions, not unique creatives: one creative can run as several ads, as the “2 ads use this creative and text” label shows. Without a database you can approximate pace by counting, among the live cards, how many show a start date in the last 30 days.

The newest cards confirm the pace. This one launched on 29 October with a new flavor.

End with a row in a shared sheet: date, competitor, source, one number and one decision. Ours: “2 November 2026, Magic Spoon Cereal, Meta Ad Library plus ad database, 77% of live ads launched in the last 30 days, the two oldest ads both sell a custom bundle with up to 26% off, and the 11 September card ties the discount to subscribing: test a subscribe-and-save discount against a one-time purchase.” Keep one row per competitor per check, so the sheet shows change over time and any colleague can repeat the check on the same page ID.

On cost: steps 1 to 3 need nothing but the free Meta Ad Library. Steps 4 and 5 are faster in any ad database with live data; the database figures above are from Spytrend, and its free tier covers only ads older than 90 days.