Every marketing team has a swipe file. Usually it is a shared folder called "inspo" with a few hundred screenshots, no dates and no names. It grows for a month, then nobody opens it again, because a screenshot without context cannot answer the only question that matters in a creative meeting: did this work for them?
A swipe file is useful when each entry carries enough context to be compared. This guide is about building that kind of file from competitor ads, and doing it quickly enough that you keep doing it.
What a usable entry contains
Before collecting anything, decide what every entry must record. We suggest the fields below because they can all be read from the ad itself, without guessing:
| Field | Why it matters |
|---|---|
| Brand and domain | So you can group entries by competitor |
| Button | Tells you which end of the funnel the ad is working |
| Days running | The closest thing to a performance signal the library offers |
| Landing page | Shows what the creative is actually selling |
| Date you added it | Lets you see what changed the next time you look |
Everything else — the hook, the format, the proof — is your own note on top.
Step 1: narrow the format before you watch anything
Creative research gets slow when you watch everything. Decide first what you are looking for.
The Meta search in the side panel can already limit the media type before the page loads: video, plain images, viral images (text-heavy memes), all images or ads with no media. Once cards are on screen, the extension's card filters narrow it further:
- Media Type — video, all images, or no media.
- Video Duration (s) — keep only short videos, or only long ones.
- Creative Count (Carousel) — separate single creatives from carousels. We look at carousels on their own in our carousel ad research article.
- Search card text — keep only ads whose copy contains a word.
There are 14 card filters in total, and they work on the cards already loaded without sending a new request to Meta. If you use the same combination every week, save it as a preset chip and apply it with one click next time.
Step 2: watch fast, in place
The slowest part of creative research is opening each ad to watch it. Hover a card and the creative grows to the size you picked while the video plays by itself. You can raise the playback speed up to 4×, which is enough to get the structure of a long video in a few seconds.
Two settings make this faster:
- Focus mode dims the other cards while you hover one, so your eye stays on the creative.
- When enlarged lets you choose between Show whole and Fill frame.
Step 3: let the extension remember what you have seen
The second time you open a competitor's ads, the problem is no longer finding creatives. It is remembering which ones you already reviewed.
The extension marks the ads you hover and look at. The Reviewed filter (in the Pro plan) then lets you show only the ads you have not reviewed yet. On a weekly check, that turns a list of hundreds into the handful that are new since last time.
Step 4: save the keepers, with their context
To save an ad, double-click its creative or right-click the card and choose Save. The Saved state filter shows only the ads you saved, so you can review your shortlist without the rest of the grid.
The Double-click a creative setting decides what a double-click does: Save (the default), Send to spreadsheet or Off. In the Starter plan, saved ads go to a spreadsheet the extension creates in your own Google account, one row per ad: brand, domain, button, days running, landing URL, Shopify type and the date it was sent. That is the table from the start of this guide, filled in for you.
Add your own columns next to those: hook, format, offer, proof. Keep them short; you are building an index, not an essay.
Step 5: tag the pattern, not the ad
A swipe file becomes valuable when you can read across it. After a few weeks, sort the sheet by your own columns and ask:
- Which hooks survive? Filter your sheet to entries with long days running and read the first line of each. The ones that repeat across different brands are category patterns.
- Which offers keep coming back? The Offer language block in the Analytics tab shows how often scanned ads mention free shipping, a discount, cash on delivery, limited stock, a money-back promise or buy-one-get-one.
- Which words dominate? Top Keywords and Top CTAs in the Analytics tab summarise the scanned copy. Click a word or a button name to filter the cards by it.
The goal is sentences such as "in this category, long-running videos open with a problem, not with the product". That is something you can brief a designer with.
Step 6: read creatives by age, not by taste
It is easy to save the creatives you personally like. It is more useful to save the ones your competitors keep paying for.
Every card shows days running and a stage badge. Winner means the ad has been running 45 days or more and is still active; an ad of the same age that was stopped counts as Ended. Sort by Longest Running and look at the top of the list first. The Winning videos preset chip combines video, duration and the Winner stage in one click.
You can also group cards by advertiser, so a single brand with dozens of near-identical ads does not bury everyone else.
Why copying fails
A swipe file is for patterns. Copying a single winning ad rarely works: you cannot see the audience, the offer or the landing page behind it, and you would be the second brand saying the same thing to the same people. We explain this in more detail in our guide to competitor ad analysis.
A weekly routine
- Open the search link you keep for each competitor.
- Run Auto Scan to load the full list.
- Filter to Not reviewed and watch only what is new.
- Save the keepers; let them flow into the sheet.
- Once a month, read the sheet across brands and write down three patterns.
The free plan scans 2,000 ads a day. The reviewed filter and the sheet export are paid features; the plans are compared on the pricing page.
To do this with the extension: Competitor creative research · Export to Google Sheets