Features
Ad Creative Intelligence: Research Ads and Design a Test
Inspect public TikTok ad examples, separate creative observations from performance claims, and build an original ad test with a practical research worksheet.
Ad creative intelligence means studying advertising examples to form a testable creative hypothesis. Record the message, demonstration, offer and destination, then check which performance fields the source actually provides. Seeing an ad, repeated variations or an impressive public counter does not establish profitability. Your output should explain what to test in your own campaign and what evidence would change the decision.
Choose the campaign decision first
Start with one uncertainty: whether buyers understand the product, believe its demonstration or recognize the offer. A broad request for winning ads makes it easy to collect entertaining examples that solve a different problem. Specify the product, audience situation, country and campaign objective before selecting references.
Separate competitor research from your own account analysis. Public examples can reveal creative choices; your authorized Ads Manager reports are needed to connect spend, clicks and attributed outcomes. An ad that has run for a long time could be profitable, underperforming or a brand exercise. Its duration alone cannot settle the question.
Understand what the source exposes
TikTok Top Ads contains advertiser-authorized examples with region, industry and objective filters. This is a selected inspiration collection rather than an exhaustive ledger of every advertiser. Open the source and preserve the filter settings and capture date.
If a source supplies a performance graph or percentile, retain its metric label, population and scale. An indexed curve is not necessarily the percentage of viewers remaining; a click-related comparison is not return on ad spend. Missing spend, attribution windows and conversion values remain unknown. Do not manufacture those fields from engagement.
Ad reference and test worksheet
| Observed creative choice | Test or evidence still needed | |
|---|---|---|
| Opening promise | Copy the claim in your own notes and identify the audience problem | Can your demonstration support the same type of promise without copying the script? |
| Proof sequence | Describe shots, evidence and visible conditions | Which original footage or substantiation must your team produce? |
| Offer and destination | Record displayed price, qualification, CTA and source URL | Does the live destination match the ad, and are terms current? |
| Metric provenance | Record the source metric, scale, country and capture date | Are spend, conversions and attribution actually supplied? |
| Experiment | Specify one creative variable and the owned outcome | Define a comparison period, allocation method and decision rule before launch. |
Read a creative without claiming its cause
Watch the complete ad once for the user problem, then again for the sequence of proof. Note when the product appears, how it is used and whether the CTA follows a demonstrated answer. TikTok Creative Codes offers an ad structure of hook, body and close. Treat that as production guidance, not proof that a particular opening caused sales.
Follow the available landing-page link and compare the promises. Record whether an important condition disappears between the creative and the page. A mismatch is a usability question for your own test, not a diagnosis of the competitor’s conversion rate. Keep rights and original execution in view when preparing references for a producer.
Worked example: a foldable travel bottle
A fictional bottle brand wants to explain why a collapsible container is useful after a flight. Three inspected ad references might use a luggage reveal, a spill demonstration and a discount announcement. The research question is whether showing the empty bottle folded beside a full one makes the benefit clearer than naming it in a headline. No real ads or performance measurements are represented here.
Draft two original treatments with the same product, offer, destination and length: one begins with the folded state; the other begins with the packing problem. Film your own usage and specify any conditions behind leak-related claims. Record the proposed primary outcome in the authorized campaign account and a comprehension check before paid release. If viewers click but misunderstand the capacity, a higher click rate alone is not a successful answer. The reference file records the reasoning for this test, rather than declaring one competitor a winner.
Retain the learning after execution
After your test, record the actual allocation, dates, spend and outcome definitions alongside the creative versions. Label observational campaign comparisons as such when delivery and audiences differ. Do not call a change a controlled experiment merely because two videos were published. Keep inconclusive outcomes and weak examples in the file to reduce winner-only selection.
Use Ad Creative Intelligence to inspect available examples by niche and country, then verify the supporting source. The tool does not independently establish another advertiser’s private account economics. Continue with content pattern analysis, adapting without copying and idea validation when turning references into production.
Sources and research notes
Reviewed on 11 October 2026 against the linked official documentation and current Superlurk tool descriptions. The procedure and ad reference-to-experiment worksheet are editorial guidance. Worked examples, sample counts and decisions are illustrative, not a measured customer study. Public-source coverage and native account eligibility vary; inspect the actual source and record missing evidence. How we approach social research.
- TikTok: About Top Ads
Advertiser-authorized creative examples and available performance context.
- TikTok: Creative Codes
Ad creative structure; recommendations are not organic guarantees.
Frequently asked questions
Do public ad libraries reveal profitable ads?
Not by visibility alone. Profitability requires appropriate spend, revenue and attribution evidence.
Is a retention-looking graph always viewer retention?
No. Check the metric and scale on the actual source before interpreting the curve.
What should an ad research deliverable contain?
Source context, observed creative choices, unknown performance fields and an original test with a decision rule.
Tools used in this guide
Written by
The Superlurk Team
We build Superlurk — a cited social search engine across 24 platforms. We write about social media search, insights, and marketing.
Keep reading
Winning Content Patterns: Compare Structures and Test an Original Idea
Research content patterns with a coding worksheet, comparable examples and counterexamples. Turn observed structures into an original test without claiming causation.
How to Learn from Creator Content Without Copying It
Keep the useful question, add your own contribution and produce assets you can use. This worksheet helps turn references into a distinct piece of work.
How to Validate a Content Idea Before Scaling It
A content opportunity is a hypothesis. Validate what you can actually test, distinguish comprehension from distribution and keep inconclusive results honest.
How to Track Competitors on Social Media: A Manual Review Routine
Set up a repeatable competitor review with named queries, a baseline, verified changes and decision owners. Use sampled comparisons without promising automatic alerts.