Shopify ad spy: a worked Shopify study

What you would do

You paste a store that is clearly advertising. The spy opens three public libraries on the brand it found. You check that the advertiser name is the store you meant. In Meta’s library you read the first line, the offer, and the start date. In TikTok’s library you watch the opening of a video that has stayed up. In Google’s transparency center you see whether Search ads exist beside the social ads.

You then take the landing URL from the ad into the CRO Audit. A long-running ad is a reason to study that page. It is not a measurement of profit, and it is not a product recommendation. If the brand name is ambiguous, you stop. The wrong advertiser makes every later conclusion wrong.

Before you change the store

These are examples. Run the tool on your own URL or your own costs. That result is the one to act on. Do not put an example in a deck as if it were your store.

Example: a cookware storefront in three libraries

Say you paste the URL of a small cookware store whose homepage sells pans, lids, and a weekday set. You want a public read on what that store shows in ads, so you can describe it to a teammate without logging into anything you do not own. The finder matches the brand on that URL and opens the Meta Ad Library, the TikTok ad library, and the Google Ads Transparency Center. You set the country to one you ship to, and you write that country at the top of a blank note before you read any card.

Say the Meta library returns cards for that brand name. You do not count them, and you do not write a total. You choose the first card whose landing domain is the cookware store and you ignore the urge to rank it. The pass from here is a reading of that card, plus a reading of whatever the other two libraries do, plus a check for any card that only shares the name. The example stays inside one store and one sitting.

You keep your own Shopify admin closed during the sitting. Your products, your costs, and your orders would answer different questions. This sitting answers a narrower one: what do the public libraries show for this storefront today, in this country. You will finish with a note a teammate can read tomorrow morning.

What the Meta card shows on that sitting

Say the kept Meta card is a short video. The opening frame is a skillet on a home stove, with a hand setting down a lid. A later moment shows chopped vegetables in the same pan. There is a line of on-screen text about weeknight cooking. You write those three visuals in that order. You name the format as video. You leave out any guess about who edited it or how much it cost to film.

Say the copy beside the video speaks to people who cook after work, names the skillet, and invites the viewer to shop the set. You quote the invitation only if you need the exact words, and you paraphrase the rest as a weeknight-cooking offer that points at a set. You record the start date exactly as Meta prints it. Say the date is from a previous season. You add one sentence: a long start date means the advertiser left the ad on. You do not add a sentence about profit, efficiency, or a winner.

Say the landing URL is a product page on the cookware domain for the weekday set. You open it, confirm the domain, and write "product page" as the page type. Say the page shows the set the ad named. You write that the page and the ad are talking about the same set. You do not write a conversion story. Spend, results, and audience stay as blank lines under the card.

What TikTok, Google, and a colliding card do

Say the TikTok ad library shows nothing for that brand in the country you selected. You write a TikTok block that says nothing was found for that brand in that library. You leave the creative lines empty there. You do not borrow the Meta video and pretend it was a TikTok ad. Coverage depends on country, so you also write the country in the TikTok block. If you later repeat the lookup in another country, that will be a new block with its own label.

Say the Google Ads Transparency Center shows a text ad for the same cookware brand. The words mention free shipping over a threshold and the landing URL is the store homepage. You describe it as a text ad, you paraphrase the shipping line, you record the start date Google prints, and you note the homepage as the page type after you confirm the domain. You keep this block separate from the Meta video. Two formats can both be public on the same day. Neither one tells you which the advertiser prefers, because preference is an account fact.

Say another card in the Meta result belongs to a restaurant that shares a word with the cookware brand. The advertiser name is the restaurant, and the landing URL is a reservation page on a different domain. You move that card to a collision line: restaurant, reservation domain, set aside. You do not describe the restaurant creative in the cookware brief. Brand-name collisions happen, and this sitting includes one so the note shows the check in action.

The note you file from that sitting

The finished note has four parts. The header carries the cookware store URL, the day, and the country. The Meta block carries the skillet video, the weeknight copy, the start date, the product-page URL, and three blanks. The TikTok block carries the empty result. The Google block carries the text ad, the shipping phrase, the homepage URL, and three blanks. A final line names the restaurant card and the reason you excluded it.

Say a teammate reads the old Meta date and calls the video a proven ad. You correct the reading in writing: the date means the advertiser left the ad on. You do not have spend, results, or audience, so the note cannot carry a proof of profit. The teammate can still use the visual description and the landing URL. They can open the product page and talk about whether your own store has a page that clear. That conversation is the useful remainder.

You stop the example there. You do not invent a test result, a budget, or a count of ads for the cookware shop. You do not claim the empty TikTok library means the shop avoids TikTok. The sitting produced a transcript of public fields. Any next step, such as a page review or a test in your own account, starts from your own store and names its own evidence.

What the cookware note still withholds the next morning

Say you reread the note before anyone else comments. The skillet video is still a skillet, a lid, and vegetables, with weeknight wording and a product-page URL. The start date still means the advertiser left the ad on. The TikTok block is still an empty result for the country you wrote down. The Google block is still a text ad toward the homepage. The restaurant card is still excluded. A night of delay does not create spend, results, or audience, so those lines stay blank.

Say a colleague asks for a count of the cookware shop ads, based on how far you scrolled. You decline. You saw the cards that library showed during the sitting, and you kept a description of the matching ones. You did not take a census, and the libraries do not owe you one through this finder. A later sitting, if you need it, becomes a new note with a new date. The old note keeps the transcript it had.

Say someone treats the long-running skillet video as a reason to add a similar set to your own Shopify catalog. You move the talk to questions you can answer from your store and from the saved URL. Do you already sell a comparable set? Can you ship it to the country on the note? What does the landing page offer once it is open? The example ends as a transcript with those questions still open, which is what makes the transcript safe to keep.

Shopify questions