Mike KarrA useful Facebook ad example gives you a specific production decision. This guide starts with three live Meta Ad Library cards from the lululemon advertiser, then turns the observations into original video briefs. The emphasis is on matching the message, product evidence and destination before adapting a concept for Feed, Stories or Reels.
Research snapshot: October 19, 2026; United States, All ads, Active ads, keyword “lululemon”. Only cards linking to the lululemon Facebook advertiser were selected. These are Meta library entries, not proof of delivery to a particular Facebook placement. We reviewed visible card copy, media type and destination links; we did not establish results or analyse the complete moving footage.
Three current ad-library examples
| Source and observed format | What the card shows | Original adaptation to test |
|---|---|---|
| Seasonal collection · 15-second video ID 1120295040943844 | Started October 6, 2026. The copy connects lightweight apparel with focus; the link leads to a seasonal collection. | Show one specific use situation, then send the viewer to a matching collection. A collection message needs a destination that helps people choose. |
| Outerwear · 15-second video ID 1128934312906085 | Started October 6, 2026. The card foregrounds insulation and material benefits and links to an outerwear collection. | Choose one supported material property. Prepare real footage or specifications that substantiate it; an avatar can explain the evidence. |
| EasyFive trousers · image ID 29257757217142684 | Started October 1, 2026. The card introduces new arrivals and links to a named wide-leg trouser product with fit details. | Turn a product-selection question into a short fit explainer. Show the actual garment and direct people to its own sizing information. |
The creative lesson is an editorial interpretation of the visible copy and link relationship. An active badge or an early start date does not establish profitability, spend or conversion rate. For a deeper brand study, read our lululemon research guide.
Choose the right adaptation
Keep the underlying product claim constant while adjusting the composition and pacing. The following are production suggestions, not a list of current upload limits. Check the selected placement in Ads Manager before exporting.
| Placement | Adapt the example | Preview check |
|---|---|---|
| Feed | Open with the product and a short caption naming the buying question. Make the demonstrated detail large enough to inspect on a phone. | Watch without sound. Can you identify the product and understand the answer? |
| Stories | Use a vertical composition with a clear sequence: question, evidence, next step. Keep the product away from interface overlays. | Read every caption in the placement preview; check that the final action remains visible. |
| Reels | Begin with a visible action rather than a long introduction. Let a presenter explain the detail while real footage supplies proof. | Watch with sound, then muted. Check voice clarity, captions and any cropped product detail. |
Three original video examples
These three travel-pouch examples are hypothetical original scripts, not transcripts of lululemon ads. The durations are editing plans. Confirm the product details before using the copy.
1. Collection-led Feed example · 20 seconds
0–4 s: show two pouch sizes beside a day bag. “Which pouch fits your daily bag?” 4–14 s: put the actual everyday items into each size and label the dimensions. 14–20 s: show both options. “Compare sizes in the collection.”
Use this when the destination offers several relevant choices. Keep the first frame and collection thumbnails recognisably related.
2. Feature-led Stories example · 15 seconds
0–3 s: open the pouch. “Where do the small items go?” 3–11 s: show the actual internal pocket and place a cable inside it. 11–15 s: show the pocket and product name. “See the pocket layout.”
Use a continuous product shot for the demonstration. Do not infer waterproofing or durability from appearance; neither is shown by a pocket demonstration.
3. Selection-led Reels example · 25 seconds
0–4 s: place the pouch next to a ruler. “Will this fit your bag?” 4–18 s: show the external dimensions and opening, then insert it into the specific bag shown. 18–25 s: presenter: “Check your bag measurements against the product page.”
Name the tested bag or show its dimensions. One successful fit does not establish universal compatibility. An AI presenter should describe the test without inventing personal ownership.
Build the variation in Kreopatra
- Write a brief containing the product URL, intended buyer, one supported feature, real product footage and the destination. Choose whether the next step is collection browsing or selection of one product.
- Use a ready UGC/avatar template in Kreopatra to develop the presenter script and video variation. Keep actual product footage for claims that depend on fit, movement or materials; review generated product details before use.
- Prepare versions for your selected placements. Review pronunciation, captions, product appearance and the destination promise in each preview. Keep an unchanged version so that a new opening has a meaningful reference.
Keep a useful research and test record
First record the observation: date, country, advertiser link, library ID, media type, visible promise and destination. Separate that observation from your proposed explanation of why the ad might work. If the card disappears, keep the dated note rather than claiming it is still active.
- Choose one question: does a fit question or a pocket demonstration attract more qualified product visits? For a narrower hook test, keep the body, offer, audience and landing page comparable and change only the opening.
- Set the spend limit, evaluation window and primary business outcome before delivery. Note placement and delivery differences. Clicks or viewing time can diagnose attention; they do not establish profitable sales.
- Record spend, qualified actions and purchases where measured. Inspect the promise-to-page match if attention rises without the intended action. Small samples can remain inconclusive; do not label a variation a winner by default.
Copy this worksheet: source ID | observation date | country | visible message | destination | evidence needed | original concept | placement | changed variable | comparison version | spend limit | evaluation window | primary outcome | result | next decision.
Frequently asked questions
Are these the best-performing Facebook ads?
No. The library cards were active when checked, but their performance was not verified. The guide supplies research observations and original testing ideas.
Can I reuse the brand’s video or script?
Build your own execution using your own product, assets and supported claims. The examples above adapt a message-to-destination relationship, not the brand’s footage or exact dialogue.
Do I need a different message for every placement?
Start with one supported product promise. Adjust framing, timing and captions for each preview; change the message deliberately when testing a separate hypothesis.
Where does Kreopatra fit?
Kreopatra helps generate video ads from ready UGC/avatar templates using a product URL or brief. Research informs the brief; authentic product evidence and review remain part of production.