Competitor Ad Research: How to Find the Ads Already Winning in Your Market

Competitor ad research is the practice of pulling your rivals' live ads from public ad libraries — Meta, Google, and LinkedIn all publish them — and reading them for one signal above all others: how long each ad has been running. An ad that has run continuously for months is an ad someone keeps paying for, and nobody keeps paying for a loser. Run length is the closest thing to a public performance metric these libraries give you.

Done well, this takes about 20 minutes and hands you the hooks, structures, and offers your market has already validated — before you spend a dollar testing your own. Done lazily, it turns into copying, which is both an ethics problem and a strategy problem (a copied ad competes with the original in the same feed).

This page covers what each library actually shows you, a practical 20-minute workflow, and how Hermoso automates the whole loop — from finding competitors to remixing their longest-running ad into a finished ad of your own.

Why is run length the signal that matters?

Because ad libraries don't publish performance data, run length is the honest proxy. Public libraries show you the creative, the copy, and the dates an ad ran — but not its ROAS, CTR, or spend. What you can infer: an advertiser who launched an ad in January and is still running it in July has seen enough return to keep funding it through six months of budget reviews. Meanwhile, the ads that appeared for two weeks and vanished were tests that failed.

So the single most useful move in competitor ad research is sorting by longest-running. It filters out the noise of everything a competitor is trying and leaves you with what's working. It's not perfect — brand campaigns and retargeting ads can run long for reasons other than raw performance — but across a set of competitors, the pattern of what survives is the best free signal available.

What can you actually see in each ad library?

Each major platform publishes a searchable library of ads, with slightly different depth:

All of this is public by design — ad transparency rules exist so anyone can see who is advertising what. You're not scraping anything private; you're reading a published record.

A practical 20-minute workflow

1. List 5–8 rivals (5 min). Direct competitors first, then similar brands one category over — a supplement brand can learn as much from a skincare brand's hooks as from another supplement.
2. Pull each one's ads, sorted by longest-running (5 min). Search each brand in the Meta Ad Library (and Google/LinkedIn if they're relevant channels) and order by start date. Ignore everything under ~30 days old.
3. Read the survivors for structure, not surface (5 min). For each long-runner, note: the hook (first line or first two seconds), the format (UGC testimonial-style, demo, static with big claim, founder-to-camera), the offer, and the CTA. Patterns across brands matter more than any single ad.
4. Save a swipefile (2 min). Keep the best 10–15 in one place with a note on why each earned its spot. A swipefile you annotate is a strategy document; one you don't is a screenshot folder.
5. Remix, never copy (3 min to brief it). Take the structure — "problem-agitate hook, product demo at second 3, price anchor, CTA" — and rebuild it with your product, your proof, your voice. The skeleton is what was validated; the skin has to be yours.

How Hermoso automates competitor ad research

Hermoso's Ad Spy — an ad spy tool that ends in a finished ad, not a screenshot — runs this entire workflow from a single input: your domain. It auto-finds your competitors and similar brands, pulls their ads from the Meta, Google, and LinkedIn ad libraries (plus organic TikTok), and sorts by longest-running by default — so the first thing you see is each rival's most battle-tested creative, not their newest experiment.

From there the loop closes in a way manual research can't: any ad you find can be remixed into a finished ad for your brand. Hermoso is an AI ad studio, not just a research tool — it generates the image or video with your real product photo composited in, your palette and logo applied, and your copy and CTA included. It routes each job across current-generation models (Seedance, Veo, Kling for video; Google's latest imaging models, Flux, and GPT Image for stills) and picks the right one per job, so you never have to. The remix keeps the validated structure of the reference ad and rebuilds every surface detail around your product.

Ads you like along the way save to a swipefile inside the app, so the research compounds instead of evaporating into browser tabs. There's a free plan with 250+ earnable free credits — no card required — and an image ad costs a few credits, with video from roughly 25.

Is competitor ad research ethical?

Researching competitor ads is ethical and expected — the data is public precisely so markets stay transparent, and every serious media buyer does it. The line is between learning from structure and stealing expression. Studying that a rival's six-month ad opens with a cost-comparison hook and lands on a demo? That's market research. Lifting their script, their claims, or their footage? That's plagiarism, and potentially a legal problem.

The practical rule: remix the skeleton, never the skin. Validated structures — hook types, pacing, offer framing — are ideas, and ideas aren't ownable. Specific copy, footage, and brand assets are. Hermoso's remix flow is built around this distinction: it takes the structural pattern of a reference ad and generates entirely new creative grounded in your product photo, brand, and copy.

Frequently asked

Is it legal to look at competitors' ads?

Yes. Meta, Google, and LinkedIn publish their ad libraries deliberately, driven by ad-transparency requirements. Viewing them is exactly what they exist for. What's not okay is reproducing a competitor's specific copy, footage, or trademarks in your own ads.

How do I know which competitor ads are actually performing?

You can't see performance metrics directly, so use run length as the proxy: sort by longest-running and treat ads live for 60+ days as validated, since advertisers don't keep paying for ads that don't return. LinkedIn additionally shows impression ranges, which adds a rough volume signal.

Does TikTok have a public ad library like Meta's?

Not with the same open access — which is why researching organic TikTok content in your niche is the practical substitute. Top organic videos reveal the hooks and formats that later become paid ads, and Hermoso includes organic TikTok in its research alongside the Meta, Google, and LinkedIn libraries.

What's the difference between remixing an ad and copying it?

Copying reuses a competitor's expression — their script, visuals, or claims. Remixing reuses the validated structure (hook type, format, pacing, offer framing) while rebuilding all the actual content around your own product, proof, and voice. The structure is what the market validated; the content must be yours.

How much does competitor ad research cost with Hermoso?

Hermoso has a free plan with 250+ earnable free credits and no card required, so you can run competitor research and generate your first ads without paying. Paid plans are $19, $49, and $149 per month; an image ad costs a few credits and video starts around 25.

How is this different from just browsing the Meta Ad Library myself?

The library shows you one brand at a time and stops at research. Hermoso auto-discovers your competitor set from your domain, pulls and ranks their ads across Meta, Google, LinkedIn, and organic TikTok in one view, and then closes the loop: it can remix any reference ad into a finished ad with your product, logo, and copy.