Separate the reference from the result
An attractive ad can be a useful creative reference even when you know nothing about its performance. The mistake is labeling it a winner because it has many views or appears repeatedly in a library.
The Meta Ad Library, Google Ads Transparency Center and TikTok Commercial Content Library expose different fields and coverage. Inspect the actual entry and record its country, date and source. Do not assume the same information is available across platforms.
Keep an evidence sheet
Download the ad-library research sheet.
| Observed | Interpretation to investigate | Still unknown |
|---|---|---|
| The ad opens with a product demonstration | The demonstration may clarify the use case | Whether it improved conversion |
| A public view count is visible | The creative received distribution | Spend, profitability and incremental effect |
| The entry shows dates or active status | The advertiser used it during that period | Why it remained active |
| Several ads use a similar message | The message is common in the sample | Whether it is effective or distinctive |
Attach the source URL and capture date to each observation. A library can change after your review, and public counts can update.
Look for patterns worth adapting
Group references by the creative question: how the product is introduced, how proof appears, how price is explained or how an objection is handled. A category folder containing dozens of ads is less useful than three examples that answer one problem in your brief.
Check whether the examples represent a narrow selection, such as a top-ads programme. A curated sample is useful inspiration but cannot describe the distribution of all ads in the market.
Turn a pattern into an original concept
Write the mechanism in neutral terms. “Shows the result before the process” can inspire a different execution. Copying the scene, script or distinctive identity of another advertiser is not necessary to use that learning.
Then connect it to your own product truth. What can you actually demonstrate? What proof can you substantiate? What offer and audience are relevant? A pattern detached from those facts becomes a fashionable imitation.
A fictional research example
Suppose three reference ads show a refill action early. The observation is the repeated demonstration. The hypothesis is that a visible refill makes the product easier to explain. An original concept could demonstrate your own refill mechanism in a different setting.
The reference ads do not prove that your version will work. Review the new asset and decide what evidence is needed before investing. Keep the inspiration source separate from any performance claim about your campaign.
Use Market as a starting point
SaliencyLab’s market-intelligence documentation explains the reference workflow. Inspect which public fields are available in each record and do not fill missing metrics with assumptions.
Bring the strongest references into a creative brief, then use the decision matrix to compare your own candidates. The outcome is an original creative direction supported by explicit observations, not a list of supposed winners with unknown economics.