How do I see my competitors' Facebook ads?
Use Meta's own Ad Library at facebook.com/ads/library, which is free, public and shows every ad an advertiser is currently running across Facebook and Instagram. Search by the advertiser's Page name, set the country and set the status filter to All to include paused ads. It shows creative, copy, start date and platform, but only shows spend and reach figures for political and issue ads, never for commercial ones.
Last updated 2026-08-11
Searching it well
Search by Page rather than keyword to get a brand's complete set, and confirm you have the real Page, since large brands often run regional Pages and lookalike-named pages exist. Set country deliberately, because results differ per market, and note that one advertiser often runs many near-identical variants: the count in the header is ads, not distinct concepts. The useful filters beyond country are platform, media type and date range, which let you isolate, say, video ads launched in the last month. Expect to click into individual ads to see all placements and variations of a dynamic ad, because the tile view compresses them.
What it will not tell you
For commercial advertisers there is no spend, no reach, no CTR and no results data. Anyone selling you a competitor's exact ad spend for a normal ecommerce brand is estimating, not reporting, usually by modelling from ad counts and observed impressions, with error bars wide enough to be decorative. Longevity is the honest proxy: an ad running unchanged for months is usually working, because advertisers kill losers quickly. Ad count trends carry signal too, since a brand suddenly tripling active ads is testing or scaling into something. Read the library for direction and patterns, and distrust any third-party number it cannot itself show you.
Reading it like an analyst
The value is in patterns across many ads, not any single ad. Look at which angles a competitor keeps returning to, because repetition across months means the angle survives their internal testing. Look at the ratio of formats, UGC versus polished studio versus static, and at how hooks open in the first line and first frame. Cluster their active ads by concept and count the variants per concept; the concept with the most variants is the one they believe in enough to iterate. Then check start dates: the old survivors are the proven core, and the newest cluster is this quarter's bet, which together sketch their whole creative strategy.
Keeping a record
The Ad Library only shows what is live now, so an ad you found today can disappear tomorrow with no history, and the disappearance itself is information you lose: you cannot later reconstruct when a competitor killed a concept or how long their winners actually ran. Saving ads into your own library preserves the creative, the copy and the dates permanently, and turns one-off browsing into a longitudinal record you can mine. Volume Creatives does this with a Chrome extension that captures ads as you browse and rehosts the media so links do not die. Whatever tool you use, capture at the moment of finding, because there is no going back for it later.
From swipe file to test plan
A folder of screenshots is not research until it changes what you launch. The working loop is: collect broadly across your category, tag what you save by angle and format, notice which patterns recur across several competitors, and then translate the pattern, not the ad, into your own concept and put it into your next test batch. Translating means keeping the structural insight (open on the problem, demo in the first three seconds, price anchor in the copy) while replacing every asset and claim with your own. Run it through the same kill-and-scale rules as any other concept; a pattern that works for a competitor's offer and price point is a hypothesis for yours, not a result.
Copying a competitor's ad directly is both a legal risk and usually a strategic mistake, since you inherit their positioning without their context. Use the library to read patterns (angles, formats, hooks, offer structure) rather than to duplicate assets.